Home

Implementation of a wireless OFDM system using USRP 2

1. sampling_freq 48000 ampl 0 1 fg gr flow_graph srcO gr sig_source_f sampling_freq gr GR_SIN_WAVE 350 ampl src1 gr sig_source_f sampling_freq gr GR_SIN_WAVE 440 ampl dst audio sink sampling_freq fg connect srcO 0 dst 0 fg connect src1 0 dst 1 return fg if name__ _ main fg build_graph fg start raw_input Press Enter to quit fg stop We start by creating a flow graph to hold the blocks and connections between them The two sine waves are generated by the gr sig source _f calls The fsu_x indicates that the source produces floats One sine wave is at 350 Hz andthe other is at 440 Hz Together they sound like the US dial tone audio sink is a sink that writes its input to File Edit Build Options Help Ya m4 fa Z Signal Source ID gr_sig_source_x0 Output Type Float Sample Rate 32000 Waveform Sine Frequency 350 Options ID options Title untitled Author unknown Description gnuradio flow graph Window Size 1280 1024 Generate Options No GUI Amplitude 1 Audio Sink Offset 0 ID audio_sink Sample Rate 32KHz Variable Signal Source Device Name ID samp_rate ID gr_sig_source_x 1 OK to Block Yes Value 32000 Output Type Float Num Inputs 2 Sample Rate 32000 Waveform Sine Frequency 400 Amplitude 1 Offset 0 x gt gt gt Warning connection can on y be Laa SEEST a source and an ee sink gt gt gt
2. 50 60 im 10 80 90 15 10 Figure 4 18 Transmitted OFDM signal Bandwidth of ofdm Frequency 6 25564 kHz piv Amplitude 37 5439 h FFT 38 1958 a en 5 0 5 Frequency kHz 10 15 Now can you get number of symbols per packet and number of samples per packet Left for the reader Channel filter is designed to cover the whole B W with an extra value 0 08 B W Figure 4 19 shows the os filter taps in the time domain based on fft_length 512 occupied_tones 200 Figure 4 19 Time domain representation for filter taps channel filter taps 57 180 4 7 3 PN based synchronization This part is discussed through figures The following three figures 4 20 4 21 and 4 22 show the output of the timing metric the matched filter and the peaks respectively timing Matched filter output DAAA Ahhh MA lA DANA el ANIMAN rl o AN A Nan Anions AA be A AEN 10000 20000 30000 40000 0 10000 20000 30000 40000 50000 Figure 4 20 Output of the timing metric Figure 4 21 Output of the matched filter peaks signals Figure 4 22 The peaks Ve 10000 20000 30000 40000 50000 58 4 7 4 Peak detection Figure 4 23 shows simulation for the timing metric matched filter and the peak signal for only 1200 points This figure shows also how the peak is detected based on fft_length 512 cyclic_prefix 128 Peak signal
3. Tx Rx Ok Gigabit Ethernet awayu peog 4a y3neg pazipyepuers Reference and System Clock SMA 1PPS Generation m SMA Ext Ref JI Int GPSDO SMA GPS mm Reference Optional zl PPS ee Figure 1 7 Internal construction of USRP 2 14 Daughter Board 1 6 4 Different sections in USRP We will discuss shortly some important parts inside USRP and USRP2 such as ADC DAC FPGA and daughter boards 1 ADC Section There are 4 high speed 12 bit ADC converters The sampling rate is 64M samples per second In principle it could digitize a band as wide as 32MHz For USRP2 there are two high speed 14 bit ADC of type LTC2284 used at 100 MS s There is two other auxiliary ADCS of type AD7922 used at 100 MS s for each daughter board connector Giga Ethernet can sustain gigabit s So this is max 800 MB s given integer decimation and a 100 MHz DSP clock 2 DAC Section At the transmitting path there are also 4 high speed 14 bit DA converters The DAC clock frequency is 128 MS s so Nyquist frequency is 64MHz For USRP 2 The DAC clock frequency is 400 MS s so Nyquist frequency is 200 MHz USRP 2 has main DAC Dual of type AD9777 used at 400Ms s and two auxiliary DACs of type AD5623 3 FPGA Now let s have a look inside the FPGA in order to understand the functionality of each of its building blocks Probably understanding what goes on the USRP USRP 2 FPGA is the most importa
4. After the unpacking the mapping process on the chunks is executed In this sub block a binary sequence is mapped to symbols As we deal with BPSK it maps the 0 and 1 to 0 to 1 respectively But its benefit will appear in the case of QPSK with gray code chosen the mapping to the 2 bit chunks of 11 and 10 corresponds to 2 and 3 respectively while 00 and 01 are mapped to 0 and 1 It exists in file map_bb cc 3 Differential Encoder A stream of chunks is encoded differentially by Yau Yn bai In a differential modulation the different binary chunks add a certain change in phase to the current phase in the carrier signal That means that the previous phase is taken into account It exists in file diff_encoder_bb cc Example When the input chunks 1 bit to this sub block is 10110100 the output will be as shown below in figure 3 4 10110100 01101100 Differential encoder Figure 33 4 Example showing input and output of differential encoder sub block 4 Chunks to Symbols It maps a stream of chunks unpacked bytes to stream of float or complex constellation points The combination of packed to unpacked sub block followed by this sub block handles the general case of mapping from a stream of bytes into arbitrary float or complex symbols It exists in file chunks_to_symbols_bc cc 25 5 Root Raised Cosine Filter Once the modulated signal is obtained a raised cosine filtering is carried out This filter is required to minimize
5. OFDM stands for Orthogonal Frequency Division Multiplexing and it is the modulation used for the data transmission in the system developed for this project As its name reveals OFDM is a multiplexing method which means that different data channels share the bandwidth available In the particular case of OFDM the signal is made of independent channels Each of them will use a fraction of the available bandwidth Each of these independent channels is called subcarrier and transports data that is modulated in the amplitude and phase of the signal All subcarriers together form the OFDM carrier OFDM is called orthogonal because all subcarriers are orthogonal to each other This orthogonality can be achieved by multiplexing the subcarriers by Frequency Division Multiplexing FDM which is called multi carrier transmission or by using Code Division multiplexing CDM which is called multi code transmission The OFDM transmission system that has been implemented in this project uses multi carrier transmission Therefore it uses the same concept as FDM which assign different frequencies to different signals Each sub carrier will use a range of the frequencies available to transmit data that will travel in the phase of the signal 4 2 Advantages and disadvantages of OFDM compared with SC Advantages Disadvantages 1 Efficient way to transmit at very high 1 More sensitive to frequency offset data rates over multipath fading 2 OFDM has
6. Warning A connection can only be created between a source and an unconnected sink Generating home alex helloworld py Random Source GLFSR Source Null Source File Source UDP Source Audio Source Wav File Source Pad Source Y Sinks Vector Sink Null Sink File Sink UDP Sink Audio Sink Wav File Sink Pad Sink gt Graphical Sinks D Operators op Add Figure 1 2 Example of a Python code the sound card It takes one or more streams of oats in the range 1 to 1 as its input We connect the three blocks together using the connect method of the flow graph connect takes two parameters the source endpoint and the destination endpoint and creates a connection from the source to the destination An end point has two components a signal processing block and a port number The port number specifies which input or output port of the specified block is to be connected In the most general form an endpoint is represented as a python tuple like this block port number When port number is zero the block may be used alone These two expressions are equivalent fg connect srcl 0 dst 1 and fg connect erc dst 1 Once the graph is built we start it Calling start forks one or more threads to run the computation described by the graph and returns control immediately to the caller In this case we simply wait for any keystroke 10 1 5 4 SWIG It is the wrapper for the C modules and ge
7. bank alone contains insufficient i information to determine if the timing should be time advanced held or retarded As shown in figure 3 10 a sample set formed prior to the peak will generate positive slope if the correlation values are positive and a negative slope if the correlation values are negative Negative d Negative OG a a Derivative ss Correlation Figure 3 10 Taking the derivative output only is not enough Polyphase Matched Filter Detector The solution for this problem is to get the product of the derivative matched filter output with the matched filter output or the sign of it The whole system is described in figure 3 11 Polyphase Derivative Matched Filter Select Path Figure 3 11 Whole system block diagram Moving to the algorithm the implementation of the timing recovery system is implemented in python and C The number of filters in the filter bank is chosen to be 32 filters with 22 taps for each one It is represented by two blocks are root raised cosine and pfb clock sync ccf we will discuss them now shortly e Root raised cosine filter taps Root raised cosine filter taps is generated using the function root_raised_cosine The implementation of this function is found in the file gr_firdes cc 31 This function returns a vector of float numbers representing the filter taps e Poly phase filter bank and timing recovery proc
8. peak is detected after 512 128 sample Matched filter output Timing metric max corr after 512 sa Figure 4 23 Peak detection process 59 4 8 Simulation codes Previous discussions results and figures were obtained from writing some codes through Python programming language Codes are listed below with same sequence of previous subsection 4 8 1 Preample from gnuradio import digital import pylab import math from gnuradio import blks2 from gnuradio import eng_notation from gnuradio import gr from gnuradio import window from gnuradio eng_option import eng option from gnuradio gr import firdes from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser Import wx class top_block grc_wxgui top_block_gui def init self grc_wxgui top_block_gui mt self title Top Block H PATITE Variables H PATITE ARRETE self samp_ rate samp_ rate 32000 _fft_length 512 _occupied_tones 200 zeros_on_left int math ceil _fft_ length _occupied_tones 2 0 ksfreq digital known_symbols_4512_3 0 _occupied_tones 60 for i in range len ksfreq if zeros_on_left i amp 1 ksfreq i 0 hard coded known symbols preambles ksfreq padded_preambles list for pre in preambles be padded _fft_length 0 Hit padded zeros_on_left zeros_on_left _occupied_tones pre Hit padded_preambles append padded padded_preambles _fft_length 0 padded_preambles zeros
9. pylab show The previous code is for the peaks file only The codes of the matched filter and the timing metric are the same except the location of the file and the size of the throttle and the file source output Then replace self gr_throttle_O gr throttle gr sizeof_char 1 samp_rate self gr_file_source_0 gr file_source gr sizeof_char 1 current directory ofdm_sync_pn peaks_b dat False By self gr_throttle_O gr throttle gr sizeof_float 1 samp_rate self gr_file_source_0 gr file_source gr sizeof_float 1 current directory sync_pn_analysis your_file dat False Peak detection First code generate files of data from gnuradio import digital from gnuradio import eng_notation from gnuradio import gr from gnuradio eng_option import eng_option from gnuradio gr import firdes from grc_gnuradio import blks2 as grc_blks2 from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import numpy import wx class top_block grc_wxgui top_block_gui 70 def mt self grc_wxgui top_block_gui mt self title Top Block ANN Variables AM self samp_rate samp_rate 32000 EH H E H H E EH HE EH HE H HH EH HE EHH H H EH HHHH Blocks EH H E HEH E EH HE HEH H EEH HH EH HE HEH HE H HHEH HEHHEHE self random_source_x_0 gr vector_source_b map int numpy random randint 0 256 1000 True self gr_vector_sink_x_2 gr vector_sink_b 1 self gr_vector_sink_x_1 gr
10. self gr_vector_sink_x_0_0 0 def get_samp_rate self return self samp_rate 73 def set_samp_rate self samp_rate self samp_rate samp_rate A if name main parser OptionParser option_class eng_option usage prog options options args parser parse_args tb top_block tb Run True fig pylab figure 1 sp fig add_subplot 1 1 1 p2 sp plot tb gr_vector_sink_x_0 data 1200 r linewidth 2 label Peak signal sp hold True p2 sp plot tb gr_vector_sink_x_0_0 dataQ 1200 b linewidth 2 label Matched filter output sp hold True p2 sp plot tb gr_vector_sink_x_0_0_0 dataQ 1200 g linewidth 2 label Timing metric sp annotate max corr after 512 sample xy 512 1 xycoords data xytext 150 30 textcoords offset points arrowprops dict arrowstyle gt connectionstyle arc3 rad 0 2 sp annotate peak is detected after 512 128 sample xy 640 1 xycoords data xytext 50 30 textcoords offset points arrowprops dict arrow style gt connectionstyle arc3 rad 0 2 sp grid sp legend pylab show 74 References 1 Venkat vinod Patcha EXPERIMENTAL STUDY OF COGNITIVE RADIO TEST BED USING USRP Thesis by B TECH in Electronics and Communication Engineering Padmasri Dr B V Raju Institute of Technology A_liated to JNTU India 2008 August 2011 2 T M Schmidl and D C Cox Robust frequency and ti
11. 1 0 10 10 10 10 10 1 0 10 1 0 1 0 1 0 10 10 10 10 10 10 10 10 10 A0 10 0 16 010 41 010 10 19 10 10 10 40 10 10 16 10 10 10 1 0 20 160 160 10 10 10 4 0 1 0 10 1010 10 1010 40 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 e Header contains a constant data represented in two parts one is payload length _with_crc and the other is whiten_offset which is Ax021x041x021x04 ant it is generated from make header function From the above output we can see that the header is 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 55 4 7 2 Preample As said before after adding zeros i
12. 4 2 Delay spread and cyclic prefix In wireless communication systems the received signal will always be received many times due to the multipath propagation This effect gives as a result in the receiver a number of signals with different amounts of delay respect the first multipath signal which usually corresponds with the line of sight path The difference of delay between the first of the multipath components and the last one is called delay spread The effect of delay spread is specially present in urban environments in which the number of 36 multipath components is higher than in rural environments but also in environments where sender or receiver are moving at high speeds This scenario could cause the multipath echoes to have important delays in time as the target might be moving close to some of the multipath components and far from others The problems that delay spread can cause are basically two First of all the different echoes that arrive at different times can come with a different phase in respect to the main signal and they can cause some distortion in the main component The second problem is the effect that the delayed echoes can cause in the next transmitted symbol This is called inter symbol interference ISI Figure 4 1 shows the effect that multipath echoes can cause on a transmitted signal To eliminate ISI completely a guard time is introduced for each OFDM signal as shown in figure 4 2 The guard time is chosen large
13. OFDM symbol 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 10 10 10 10 1 0 1 0 10 10 1 0 16 10 10 1 0 1 0 10 10 1 0 LO 10 1 0 10 1 0 10 10 LO 10 LO 10 1 0 1 0 10 1 0 1 0 1 0 10 1 0 10 10 1 0 10 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 0 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 54 1 0 1 0 10 4 6 10 1 6 1 6 1 0 1 0 10 10 10 10 1 0 10 10 10 10 10 16 16 10 10 10 1 0 1 02 10
14. between the two blocks Data flows in one direction from a signal source to one or more signal sinks This construction of software radio is similar to development of hardware radios but with an additional restriction that the signal flow in a flow graph cannot form a feedback cycle so implementation of any feedback mechanisms must be contained within one signal processing block In GNU Radio the signal processing blocks are defined in C for performance while the connections between the blocks for a given application are declared in Python Using a high level language like Python allows users to quickly create different applications by constructing a signal flow graph simply by making connections between smaller building blocks This approach meant that the agility of software development in a high level language can be maximized while at the same time sidestepping its drawback of slow performance by acting only as glue code and offloading the heavy lifting to C compiled code Interoperation and data marshalling between Python and C is done by employing the Simple Wrapper Interface Generator SWIG GNU Radio uses a number of data types to represent the signal at the interfaces of each of the signal processing blocks The data type used by a particular block can usually be identified through the naming convention that each block should be suffixed with a code to represent its interface For example the block gr_rms_cf has a suffix of _cf w
15. ee EE edd aes 54 Acid PLEASE a Geen A attend nn ne ns 56 Ada BW and channel filt r ls 56 4 1 4 PN based synehronizaO Msi 58 4545 Peak detection eene 59 4 8 Simulation COLES sisi eiii 60 SON A e en cals teas cae cat vata Sages 60 4 8 2 BW and channel Mera 62 4 8 3 ARTE nn need sn lisent 66 ROLETENCES gare ne N DU EE 75 F t re WOTES SE RSR Aleit alee A ee AAA 78 List of Tables Table 1 1 Comparison between USRP1 USRP2 and USRP N210 Table 1 2 USRP daughterboards available from Ettus research Table 4 1 Advantages and disadvantages of OFDM comapred with SC vi List of Figures Figure 1 1 Module diagram of a SDR sender and receiver c oconoccnocononnconnnonnnonnncnoncnnnnos 3 Figure 1 2 Example of a Ereegnes oa 10 Figure 1 3 USRP Motherboard cuiuncininin iia iia aia ps 12 Risure 1 lat level view of USRP oscila 12 Figure 1 5 RTE 13 Figure 1 6 USRP 2 and its motherboard e ne ei 13 Figure 1 7 Internal construction of USRP 2 224 slim aglia terne 14 Figure 2 1 Single tone transmission Path cvrmi ii ia 19 Figure 2 2 Single tone reception A ee ee Rene re a 20 Figure 3 1 DBPSK transmission path ii A 23 Figure 3 2 Data flow through the modulation block 24 Figure 3 3 Example showing input and output of packed to unpacked sub block 25 Figure 3 4 Example showing input and output of differential encoder sub block 25 Figure 3 5 DBPSK reception path vicio ii tinte 26 Figure 3 6 Data flow th
16. filter The first module of the receiver it is a simple fourier low pass filter for the input signal coming from the antenna that takes the bandwidth corresponding to the number of carriers that contain actual data which is usually less than the size of the FFT and in our implementation it was 200 used subcarriers for an FFT of 512 subcarriers BW Occupied tones FFT size 2 sampling frequency It is defined in gr_fft filter cec cc The time domain signal of it simulated under GNU Radio which is roughly a Sinc function is shown in figure 4 9 channel filter taps Figure 4 9 Time domain signal of channel filter 2 OFDM symbol synchronization It is defined in the python file ofdm_sync_pn py Once the signal has been filtered it is fed into the synchronization block This module makes a very important task it is responsible for making timing synchronization and frequency error correction based on Schmidl and Cox s theory 2 Synchronization is based on PN sequence correlation This theory is used for the correlation of received signal r n in many 46 OFDM timing synchronization methods The preambles which are used for timing synchronization are designed for having good correlation property In order to achieve the well correlation property Pseudo noise PN sequence is used for preambles of timing synchronization They construct the first training symbol with 2 identical halves in time by transmitting a pseudo noise PN
17. function with the payload as a parameter The callback function will be executed in the uppermost level of the blocks hierarchy and the output this data after remove CRC32 and the header sent to output port in ofdm demodulator block Again and in short this block performs Synchronization FFT demodulation of incoming OFDM symbols and passing the packets to the higher layer 4 6 2 Blocks of OFDM demodulator A OFDM receiver This block is defined in a python file ofdm_receiver py It includes four main modules defined in C files as shown in figure 4 8 Channel Filter OFDM Symbol Synchronization FFT Frame Acquisition The detailed construction of this block is shown in figure 4 8 where the blue arrow shows the path the signal follows while the black arrows carry other kinds of data that are not the actual OFDM signal channel frequency filter modulator ofdm sampler Figure 4 8 Block diagram OFDM receiver block Its input is the complex modulated signal at baseband And the output is synchronized packets are sent back to the demodulator It performs receiver synchronization on OFDM symbols It performs channel filtering as well as symbol frequency and phase synchronization The synchronization routines are available in three flavors l Preamble correlator Schmid and Cox 2 Modified correlator with autocorrelation 3 Cyclic prefix correlator Van de Beeks Deeply into the sub blocks 45 1 Channel
18. generator generates different types of signals such as sine and cosine waves square wave and triangle wave Here in our case we generated cosine signal with frequency 10 KHz 2 2 2 Wide band FM transmitter It is defined in python file wfm_tx py It has one input audio samples and one output modulated signal It takes a single float input stream of audio samples in the range 1 19 1 and produces a single FM modulated complex baseband output based on some parameters such as audio rate and quadrature rate 2 2 3 USRP sink The sink used here is taking the baseband complex sampled signal at the modulator block output in order to transmit it through the GE to the USRP 2 motherboard Basically the parameters set the transmitting frequency to which the baseband carrier is going to be up converted 2 3 Reception Path To evaluate the USRPs performance when receiving we set six main blocks as it can be seen in figure 2 2 WX GUI FFT Sink Band Pass Filter Decimation 1 Gain 10 Sample Rate 64k Low Cutoff Freq 9 9 High Cutoff Freq 10 1k Transition Width 100 Window Hamming Beta 6 16 UHD USRP Source Samp Rate Sps 640k Ref Level dB 0 Ref Scale p2p 2 FFT Size 1 024k Refresh Rate 15 Audio Sink Sample Rate 44 1KHz D Center Freq Hz 220M Ch0 Gain dB 0 Figure 2 2 Single tone reception path 2 3 1 USRP source This block as the beginning of the chain provides us the recei
19. in figure 4 13 where the first is the fine frequency correction value and the other is the timing metric signal due to correlation Timing metric signal Frequency error correction value Figure 4 13 Block diagram of OFDM symbol synchronization module 48 Also the block diagram of internal structure of the synchronization block is shown in figure 4 14 lt Moving average filter lt Matched Filter Regenerate Peaks Moving average filter Poronccrrnnnnncnnnnncsnn nsrr aa Si EN DT Ee Figure 4 14 Internal structure of OFDM synchronization block It is clear that there are five steps for synchronization process to get the start of frame and frequency error correction value 1 The first step is calculation of signal power 2 Then apply correlation between the first half of delayed signal and the conjugate of second half of the received signal as shown below in figure 4 15 Correlation Figure 4 15 Correlation between of halves 49 3 Thus we can calculate the timing metric according to the Schmid and Cox equation as mentioned before 4 To solve the problem of uncertainty due to the plateau outputted from timing metric the timing metric is inputted to a matched filter its taps are ones with CP length to find the end of the plateau of the metric and get the smooth peak as we will show in the simulation section 5 Calculating the angle between the first half and
20. port As the output are bytes it is necessary to perform a unpacking from bytes to a 1 bit vector 1 e groups of 1 bit chunks as the number of bits per symbols is 1 to treat the signal bits properly in the demodulation process However In order to understand the behavior of the modulation block the random source is replaced by a vector source so that we can introduce a known binary data 23 3 2 2 DBPSK modulator Basically this block takes care of the root raised cosine filtering and performs a phase shift modulation The input data consists of a byte stream coming from the vector source and the output is a complex modulated signal at baseband The complete functionality of the modulator block can be easily divided in different five sub blocks functions defined in C code and connected with Python code The parameters that are necessary here are the number of samples per symbol and the excess in bandwidth that refers to the roll off factor of the root raised cosine filter These parameters will be explained on greater detail in this section The signal is flowing in a chain of sub blocks as shown in figure 3 2 Differential Encoder Modulated Data Chunks to Symbols Root Raised Cosine Figure 3 2 Data flow through the modulation block Note that the vector source generates a binary sequence from a list of byte type numbers with no predefined timing The timing is set at the USRP sink and the demodulator block by means of
21. sequence only on even subcarriers and transmitting nothing on odd subcarriers i e X 2i 1 0 fori 0 N 2 1 as shown in figure 4 10 E 0 110 Till FFT size Figure 4 10 Distribution of preamble in frequency domain Also The form of preamble proposed in time by Schmid and Cox is shown in figure 4 11 vomseno Pes Figure 4 11 Preamble symmetry in time The Schmidl and Cox s timing estimator finds the start of the symbol at the maximum IPC R d point of the timing metric given by M N N 4 1 Where P 3 d W r d k 4 And R d gt ro k D k 0 k 0 47 Here d is the time of first signal sample in the window of length N for symbol timing estimation P d is the conjugate multiplicative result and R d is energy of received signal The symbol timing offset is estimated by finding d which maximizes the timing metric M d The timing metric of Schmidl and Cox s method has a plateau as shown in figure 4 12 due to the CP which leads to some uncertainty as to the start of OFDM symbol Peak signal peak is detected after 5124128 sample Matched filter output Timing metric max corr after 512 sa 0 8 0 6 0 4 0 2 9 200 400 600 800 1000 1200 Figure 4 12 Timing metric of Schmidl and Cox method Back to OFDM symbol synchronization it has a single input which is the filtered signal from channel filter and two outputs as shown
22. the ISI which causes a smearing into adjacent time slots due to time spreading in a real system At the output we have the complex filtered samples of a baseband carrier ready to be transmitted It exists in file root_raised_cosine cc 3 2 3 Multiply const This block as its name reveals is responsible for multiplying the signal by a constant It simply increases the amplitude of the digital data or any data generally sent 3 2 4 USRP Sink The sink used here is taking the baseband complex sampled signal at the modulator block output in order to transmit it through the GE to the USRP 2 motherboard Basically the parameters set the transmitting frequency to which the baseband carrier is going to be up converted 3 3 Reception Path To evaluate the USRPs performance when receiving we set three main blocks as it can be seen in figure 3 5 DPSK Demod Type DEPSK Samples Symbol 16 Excess BW 600m FLL Bandwidth 1757 Phase Loop Bandwidth 175m Timing Bandwidth 500m Omega Relative Limit 5m UHD USAP Source Samp Rate 5ps 250 Cho Center Freg Hz 1836 Cf Gain dB 0 Figure 3 5 DBPSK reception path 26 3 3 1 USRP source This block as the beginning of the chain provides us the received signal coming through the GE link from the USRP 2 motherboard This signal is a complex digitalized signal with a sample rate of 250 kHz down converted to baseband 3 3 2 DPSK demodulator This block takes care of t
23. the gain when a loud signal appears Decay rate How fast the AGC increases the gain when the loud signal is gone Reference This is the level the AGC will try to maintain Gain The overall gain of the AGC Max gain The maximum gain the AGC can have 27 Conclusion in short from C algorithm we can say that AGC takes an input complex sample and multiplies it by a gain parameter Then it calculates the output sample power It makes a control loop to check signal power reached the desired power represented in Reference parameter or not till reaching it In addition it exists in the file gr_agc2_cc cc 2 FLL It stands for Frequency Locked Loop Its main aim is carrier recovery using band edge filter But how this is done It is based on the theory of carrier recovery implemented in C algorithm We use a maximum likelihood frequency estimator to compensate frequency offset in the received signal We construct a filter that is the derivative of the matched filter in the frequency domain Note that the derivative is zero everywhere except in the transition band controlled by roll off factor of the filter seeing that the non zero spectral response of the derivative matched filter spans the band edges of the matched filter Then product the input spectrum of the received data that contain some frequency offsets with band edge filter and monitor the two spectral segments in the output of the band edge filter then comparing the
24. the sample rate and the factor of samples symbol That is to say for a sample rate of 250 kHz at the USRP sink and a factor of 2 samples symbol for the interpolation filter at the modulation block the symbol rate of 250 Ksamples s _ 2 samples symbol ES 125 ksymbols s The filter considered for any possible demodulation is a root raised cosine with a roll off factor of 0 6 For that reason the input parameters are DPSK and an excess bandwidth of 0 6 to match it the signal filter The number of samples symbol is set to 2 to provide a signal with symbol rate of 125 ksymbols s the source is defined and its value is 125 ksymbols s 1 Packed to unpacked Since the signal at the output from the vector source are packed bytes it is necessary to convert it into a 1 bit chunk stream It is for BPSK what about another modulation schemes Left for the reader in order to perform a proper identification and treatment of the symbols Data type is still being byte type at the output It exists in file packed_to_unpacked_bb cc This sub block converts the data based on some parameters such as bits per chunk or Endianness LSB or MSB e g If the data of vector source is 101 and is inputted to this sub block with 1 bit per chunk and LSB parameters the output must be as shown below in figure 3 3 24 101 Pocito unpacked TT 00000000 00000001 Figure 3 3 Example showing input and output of packed to unpacked sub block 2 Mapper
25. 2 2 3 2 3 2 3 1 2 3 2 2 3 3 2 3 4 2 3 5 2 3 6 Chapter 3 3 1 3 2 3 21 3 2 2 3 2 3 3 2 4 3 3 3 3 1 3 3 2 3 3 3 Chapter 4 4 1 4 2 Analog Environment 19 Motivar cates 19 BREED ed EE 19 SIM Al OU rd dis 19 Wide band FM rasa eke eden 19 USRP Sikes lili aed chiens tee ide EE 20 Reception Pad eh Miia i ate del Ga Men ne en 20 USRP SOU tae aitor 21 Wide Band FM TEE ot xs cates Socata ee deele 21 Band pass llena ii 22 ue EE 22 A cated tesa Se ee AS al haute E se Ne 22 EE SIREN A ER A A 22 Digital Environment int deaiaisiaactcuacdeia tee dee eae annie 23 Motivation St ta nine sn ER Te As 23 Transmission EG 23 Kalen Vector SOUrCE issiran iasi a aae e aai a 23 DPSK modulator e A SER ERa 24 Multiply CONS EE 26 USRE Sinki aea cy e She eed sateen A 26 el EE He USRP EE SH DPSK demodulator iii id 27 KT 34 OFDM NEE 35 TOCA RS AS E 35 Advantages and disadvantages of OFDM compared with SC nino 35 iv 4 3 EE 35 4 4 Iimportantissues im OFDM nie nee nn ln 36 BAN O 36 442 Delay spread and cyclic prod 36 4 5 OFDM modulator A ee aioe 38 AST Mou VatiOn o suicida 38 4 5 2 Packet structure and construction eegen io 39 45 3 Blocks of OFDM modulator eins 39 4 6 OFDM demod latot ci a ne 44 46 1 Mota zeegt detec Ge 44 4 6 2 Blocks of OFDM demodulator AAA 45 Ofdm symbol synchroniZatl OM id tennis 48 4 7 A point by point simulation Riad 54 ATA A siadsiiatiscien
26. 275 00 100 RFX2400 Transceiver 2300 to 2900 275 00 50 Table 1 2 USRP daughterboards available from Ettus research 1 6 5 General Notes These notes are collected from our experience with USRP boards and may be helpful for you Sometimes while using USRPs O U u a characters appear on the screen when you run gnuradio program It appears conceptually when data flows from USRP to PC is stopped or something near that And now we will show the meaning of these characters u USRP a audio sound card O overrun PC not keeping up with received data from USRP or audio card U underrun PC not providing data quickly enough aUaU audio underrun not enough samples ready to send to sound card sink uUuU USRP underrun not enough sample ready to send to USRP sink uOuO USRP overrun USRP samples dropped because they weren t read in time A faster machine will generally cure this problem This assumes that you are not asking the USB GE for USRP USRP 2 respectively to do something that it can t RISC processor used in USRP 2 is ZPU Minimum interpolation and decimation rates are 4 and maximum is 512 The sampling rate is very important factor in implementations through the USRP 2 The DC component is blocked by USRP 2 to remove LO leakage The best way to work around this is to use the advanced tuning parameters to offset the 17 LO aw
27. 56 1000 True self gr_throttle_O gr throttle gr sizeof_gr_complex 1 samp_rate self digital_ofdm_mod_0 grc_blks2 packet_mod_b digital ofdm_mod options grc_blks2 options modulation bpsk fft_length 512 occupied_tones 200 cp_length 128 pad_for_usrp False log None verbose None payload_length 0 EH H E H HE EH HE HEH HEE HHE EH HE HEH HE H HE EH H HHHH Connections FEH H EE HEHH EH HEHEH HE HHE EH HE HEH HE H HEH H HEHH self connect self random_source_x_0 0 self digital_ofdm_mod_0 0 self connect self digital_ofdm_mod_0 0 self gr_throttle_0 0 self connect self gr_throttle_0 0 self wxgui_fftsink2_0 0 def get_samp_rate self return self samp_rate def set_samp_rate self samp_rate self samp_rate samp_rate self wxgui_fftsink2_0 set_sample_rate self samp_rate A if name main parser OptionParser option_class eng_option usage prog options options args parser parse_args 64 tb top_block tb Run True Channel filter import pylab from gnuradio import eng_notation from gnuradio import gr from gnuradio eng_option import eng_option from gnuradio gr import firdes from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import wx class top_block grc_wxgui top_block_gui def mt self grc_wxgui top_block_gui mt self title Top Block H PATITE Variables HR RER RER RER fft_length 512 occup
28. Implementation of a wireless OFDM system using USRP 2 and USRP N210 kits By Amr Youssef Fathy Youssef Karim Mokhtar Hamdna Allah Hassan Mohamed Gamal Mostafa Mohamed Taha Saad Under the Supervision of Dr Mohamed Khairy A Graduation Project Report Submitted to the Faculty of Engineering at Cairo University in Partial Fulfillment of the Requirements for the Degree of Bachelor of Science in Electronics and Communications Engineering Faculty of Engineering Cairo University Giza Egypt July 2012 Do not go where the path may lead go instead where there is no path and leave a trail Ralph Waldo Emerson Table of Contents Eistof Lables sca leer otras yea ao vi List OF PIU ta vil List of Symbols and Abbreviations aos ix eet e xi beta ascii illa iia xii Chapter le nod Wenn seniii n A 1 1 1 e EE 1 1 2 Kee EE 1 1 3 Software Defined Radio musicians 2 LS DEPTO a Gaede 2 1 3 2 Basic principle and difference to analog radios 3 1 3 3 Defining software radio using tiers 4 1 4 GNU Radial 5 Lag Concepts and architecture 5 1 4 2 GNU Radio asta lO cda 6 1 5 Li R E a N 8 o e eee ee eae 8 Lk Canaan 9 A eene Ee ebe 9 E SWIG re RE Re tn en 11 1 6 US e sais 11 LE Motivar a as 11 1 6 2 Motherboard and internal construction 11 1 6 3 Af USRPLtOUSRP2 ne den ne ne nn i 13 1 6 4 Different sections in USRP ss eegene 15 1 6 57 General Notes us ini iaa 17 111 Chapter 2 2 1 22 2 201 222
29. _on_left zeros_on_left _occupied_tones preambles AM Blocks ANN self gr_vector_to_stream_0 gr vector_to_stream gr sizeof_gr_complex 1 512 self gr_vector_source_x_0 gr vector_source_c padded_preambles False 1 self gr_vector_sink_x_0 gr vector_sink_f 1 self gr_vector_sink_x_1 gr vector_sink_f 1 self gr_fft_vxx_0 gr fft_vcc 512 False True self gr_stream_to_vector_0 gr stream_to_vector gr sizeof_gr_complex 1 512 self gr_complex_to_real_0 gr complex_to_real 1 self gr_complex_to_imag_0 gr complex_to_imag 1 AM Connections ANN self connect self gr_vector_source_x_0 0 self gr_stream_to_vector_0 0 self connect self gr_stream_to_vector_0 0 self gr_fft_vxx_0 0 self connect self gr_fft_vxx_0 0 self gr_vector_to_stream_0 0 self connect self gr_vector_to_stream_0 0 self gr_complex_to_real_0 0 self connect self gr_vector_to_stream_0 0 self gr_complex_to_imag_0 0 61 self connect self gr_complex_to_real_0 0 self gr_vector_sink_x_0 0 self connect self gr_complex_to_imag_0 0 self gr_vector_sink_x_1 0 def get_samp_rate self return self samp_rate def set_samp_rate self samp_rate self samp_rate samp_rate self blks2_stream_to_vector_decimator_0 set_sample_rate self samp_rate A if name main parser OptionParser option_class eng_option usage prog options opt
30. _option from gnuradio gr import firdes from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import wx class top_block grc_wxgui top_block_gui 12 def mt self grc_wxgui top_block_gui mt self title Top Block ANN Variables AM self samp_rate samp_rate 32000 RETREAT TERETE TEEPE EE TEE ETE ETE ER ER Blocks H TEEPE EE TEE ETE ETE RER EEE self gr_vector_sink_x_0_0_0 gr vector_sink_f 1 self gr_vector_sink_x_0_0 gr vector_sink_f 1 self gr_vector_sink_x_0 gr vector_sink_b 1 self gr_throttle_O gr throttle gr sizeof_char 1 samp_rate self gr_file_source_0_0_0 gr file_source gr sizeof_float 1 home mohamed Desktop Simulations sync_pn_analysis ofdm_sync_pn theta_f dat False self gr_file_source_0_0 gr file_source gr sizeof_float 1 home mohamed Desktop Simulations sync_pn_analysis ofdm_sync_pn mf_f dat False self gr_file_source_0 gr file_source gr sizeof_char 1 home mohamed Desktop Simulations sync_pn_analysis ofdm_sync_pn peaks_b dat False EH A H H E H HE EH HE HHE EH HE EH HE H H EH H HEHHEHE Connections EH H EE HHE EH HE HEH HEE H HE EH HE HEH HE H HH HHHH self connect self gr_file_source_0 0 self gr_throttle_0 0 self connect self gr_throttle_0 0 self gr_vector_sink_x_0 0 self connect self gr_file_source_0_0_0 0 self gr_vector_sink_x_0_0_0 0 self connect self gr_file_source_0_0 0
31. a mixed signal 11 USRP Hew 4 1 12 27 2005 u 1 H es cars i r Figure 1 3 USRP Motherboard processor that takes care of the conversion between analog and digital signals digital up conversion in the transmit path and interpolation decimation of the signals The motherboard can have up to 4 daughterboards two for receive and two for transmit to achieve wireless communication at different frequencies They consist of the RF front end where the signal is up converted from the intermediate frequency to the carrier frequency or vice versa for the received signal And The following figure is a high level view of the USRP Figure 1 4 High level view of USRP 12 1 6 3 From USRP1 to USRP2 USRP1 and USRP2 provide the hardware platform for SDR in order to receive and transmit the signal Figure 1 5 shows a typical USRP2 board Figure 1 5 USRP 2 The USRP motherboard was discussed before For USRP 2 it is built on the success of USRPI but it is not meant to replace USRP1 Motherboard of USRP2 is shown in figure 1 6 The following new features are added to USRP2 Gigabit Ethernet interface 25 MHz of instantaneous RF bandwidth Xilinx Spartan 3 2000 FPGA Dual 100 MHz 14 bit ADCs Dual 400 MHz 16 bit DACs 1 MByte of high speed SRAM Locking to an external 10 MHz reference PPS pulse per second input Configuration stored on standard SD cards Standalone operation The ability
32. absolute value squared of the two segments which providing error signal used to adjust the spectrum of the modulated data The carrier recovery process is also shown in figure 3 7 for more clarification Band edge Filter Spectrum Centered Input Spectrum Output of Band edge Filter Band edge Filter Spectrum Offset Input Offset e Smaller Spectrum Output of Band edge Filter Figure 3 7 Carrier recovery process 28 Its important parameters are e Samples per symbol Sets the number of samples per symbol the system should use This value is used to calculate the filter taps e Filter Roll off Factor This parameter sets the roll off factor that is used in the pulse shaping filter and is used to calculate the filter taps e Prototype Filter Size This sets the number of taps in the band edge filters This should be about the same number of taps used in the transmitter s shaping filter and also not very large A large number of taps will result in a large delay between input and frequency estimation and so will not be as accurate Between 30 and 70 taps is usual In this implementation use 55 e Loop Bandwidth Sets the loop filter bandwidth in control process This 2 2m oi q should be between 05 and A in rads sample It must also be a positive number Now what about tasting the flavor of the communication algorithm implemented on C in the file gr_fll_band_edge_cc cc e First We form ban
33. airo University for his kind supervision and valuable suggestions Dedicating some of his valuable time and encouraging guidance were the reasons for enriching this work It seems very difficult to select the suitable words expressing our respect and appreciation to Eng Hazem Nasef for his valuable assistance encouragement and helpful instruction throughout the work Also we would like to extend our deep thanks to our families for their permanent help and support through every step not only in this work but also through every step in our life Lastly some words which must be said the GNU Radio mailing list is a treasure and some mailing list answers should be gold weighted Please send feedback corrections for any technical mistakes GNU Radio is great and impressive project and the USRP is an amazing device We always ask ourselves how they build this great project We will never know the answer Authors Xi Abstract As technologies quickly evolve and computers and devices become more powerful and economical paths of research appear allowing a new mass of researchers the chance to work on technologies that were only available to few This is the case for wireless communications technologies The practical research was very costly in terms of time and also money sometimes even being necessary to build prototype circuit boards for testing a possible model Actual commodity computers have become powerful enough to be able to undertak
34. als www csun edu skatz katzpage sdr_project sdr grc_tutoriall pdf www csun edu skatz katzpage sdr_project sdr gre_tutorial3 pdf www csun edu skatz katzpage sdr_project sdr gre_tutorial4 pdf www csun edu skatz katzpage sdr_project sdr grc_tutorial2 pdf 30 loading_sd_card_image for usrp2 www csun edu skatz katzpage sdr loading_sd_card_image pdf 31 GNU radio mailing list 76 http lists gnu org archive html discussgnuradio 32 http www trondeau com blog 2011 8 13 control loop gain values html 33 http www gnu org software gnuradio doc howto write a block html 34 http gnuradio org redmine projects gnuradio wiki TutorialsWritePythonA pplications 97 Future Work There is a great scope for improving the current project of OFDM The effective area of research and development work includes 1 Extending the project to increase the spectral efficiency and decrease BER using methods of channel coding and other communication concepts 2 Implementation of communication standards such as WI FI and LTE 3 Add multiple users We would like to have a group with three kits representing two users and a base station We would also like to have multiple antennas 78
35. amp_rate samp_rate 32000 FE EEE HE EH HEHEH HEE H HE EH HE HEH HEH HE EH HEHHEHE Blocks EH HEE HEHE EH HEHEH HEHHE EH HEHEHHE H HEH H HHEH self random_source_x_0 gr vector_source_b map int numpy random randint 0 256 1000 True self gr_vector_sink_x_2 gr vector_sink_b 1 self gr_vector_sink_x_1 gr vector_sink_b 1 self gr_vector_sink_x_0 gr vector_sink_f 1 self gr_throttle_O gr throttle gr sizeof_float 1 samp_rate self digital_ofdm_sync_pn_0 digital ofdm_sync_pn 512 128 True self digital_ofdm_mod_0 grc_blks2 packet_mod_b digital ofdm_mod options grc_blks2 options modulation bpsk fft_length 512 occupied_tones 200 cp_length 128 pad_for_usrp False log None verbose None payload_length 0 67 AM Connections FREER a self connect self digital_ofdm_mod_0 0 self digital_ofdm_sync_pn_0 0 self connect self gr_throttle_0 0 self gr_vector_sink_x_0 0 self connect self digital_ofdm_sync_pn_0 0 self gr_throttle_0 Ou self connect self digital_ofdm_sync_pn_0 1 self gr_vector_sink_x_1 0 self connect self random_source_x_0 0 self gr_vector_sink_x_2 0 self connect self random_source_x_0 0 self digital_ofdm_mod_0 0 def get_samp_rate self return self samp_rate def set_samp_rate self samp_rate self samp_rate samp_rate 1 if name main 1 parser OptionParser option_class eng_o
36. and using certain input parameters the input signal is stored in a vector file or sent to a binded TCP lor UDP2 socket 3 Operation blocks These blocks use a configurable number of input signals with configurable data types to produce a certain number of output signals with specific data types using the input parameters to perform a certain operation on the samples at the input These operations can be modulations or demodulations coding operations filters synchronizations type or stream conversions etc Among the different parameters needed to perform the operation the sampling rate stands always out so that a correct treatment of the signal can be done 4 Visualization blocks These blocks can be classified as a type of sink block which generates a graphical output from the input signals In this group of blocks we can mention scopes to provide a time domain representation FFT sink for a frequency domain screening constellation plots etc The different blocks are connected in a proper way so that the signal data can flow along a chain taking into account data 8 types vector lengths etc The core functionality of each block is defined by Python or C code 1 5 2 C Signal processing blocks process streams of data from their input port to their output port The input and output ports of a signal process block are variable So a block can have multiple outputs and multiple inputs The signal processing blocks are written i
37. at freedom at a very moderate price Some years ago the leap between the theoretical part of a technology and its practical implementation was very big both in terms of requirements and specially costs The proliferation of SDR projects specially based on open source software and the interest from the academic and industrial community for its utilization have created a relatively simple and comparatively very affordable solution for the implementation and testing of a very large number of wireless technologies One of the objectives of this project is to gain knowledge about actual SDR projects its state of the art and the possibilities it offers Another important point for the development of this project more specifically its practical part is our interest for the GNU Radio project the chosen implementation platform and one of the most developed and active toolkits for the development of SDR applications as of today This report will allow us to gain experience and get acquainted with GNU Radio its structure its programming its advantages and its limitations The implementation of the communication system in the GNU Radio environment is the perfect way of getting to understand this tooklit In order to finish shaping this project the content of the implementation should be decided The most attractive technology for us was OFDM It has a very good behaviour in terms of spectral efficiency and it is a very hot technology that has found its way in
38. ay from your signal this will move the LO and thus the DC offset correction outside the band of interest Sometimes the gain setting in the USRP 2 sources sinks is important It controls the gain of the RF frontend itself both settings are necessary for proper operation in your application In all our practical communication systems done in this project we used daughter boards Basic TX and Basic RX RFX 1800 and RFX 2400 18 Chapter 2 Analog Environment 2 1 Motivation We are going in this phase of the project to show practical analog implemented on USRP boards under GRC Successful transmission and reception of single tone is successfully accomplished Connecting received tone to audio sink enabled us to distinguish between frequencies according to intensity of pitch We will discuss transmission path as well as reception path in details in the coming sub sections 2 2 Transmission path For transmission path the scenario set up is basically constituted of three main different blocks signal source WBFM transmitter and USRP sink as shown in figure 2 1 Now we will discuss the operation of each block Options ID top block Generate Options WX GU ea CIS Rs a UHD USAP Sink Rabe tt Samp Rate Sps 64k Spee a Ch0 Center Freq Hz 2204 Variable Max Deviation 75k D Gain dB 0 ID samp rate Value 4k Figure 2 1 Single tone transmission path 2 2 1 Signal source Simply it is just a
39. back is that even if the power needed could be reduced the size of the hardware needed to process the signal is also much bigger than the dedicated hardware of the traditional radios This is why the project of having a portable device based on SDR that can be used as many different radios has not evolved very much and the use of SDR has been reduced to research or applications in the base station instead of the mobile terminal where the power requirements are not so critical at the moment 1 3 3 Defining software radio using tiers The SDR Forum has defined the following tiers describing evolving capabilities in terms of flexibility e Tier 0 The Hardware Radio The radio is implemented using hardware components only and cannot be modified except through physical intervention e Tier 1 Software Controlled Radio SCR Only the control functions of an SCR are implemented in software thus only limited functions are changeable using software Typically this extends to inter connects power levels etc but not to frequency bands and or modulation types etc e Tier 2 Software Defined Radio SDR SDRs provide software control of a variety of modulation techniques wide band or narrow band operation communications security functions such as hopping and waveform requirements of current and evolving standards over 4 a broad frequency range The frequency bands covered may still be constrained at the front end requiring a switch in the a
40. board is for transmission while the other is for reception Different scenarios and issues will be handled and discussed After that conclusions will be deduced 4 4 Important issues in OFDM 4 4 1 Orthogonality The OFDM modulation method relies on the orthogonality between its subcarriers to achieve a good spectral performance In the case of multi carrier transmission the chosen frequencies must be orthogonal between each other This means all frequencies must be multiples of the inverses of the symbol duration The orthogonality of the frequencies used reduces greatly the crosstalk interference between sub carriers and increases the spectrum utilization Each of the OFDM subcarriers has a range of frequencies assigned to it and all of them together fill the spectrum used for the OFDM carrier that is the bandwidth available The data in bits will be split among the subcarriers by using a serial to parallel converter and for each subcarrier it is independently modulated using in most cases a Quadrature amplitude modulation QAM or a Phase shift keying PSK modulation Then the OFDM carrier is created with all the modulated subcarriers by using an inverse Fast Fourier Transform iFFT module that calculates the time domain signal with all subcarriers to create a single broad band complex signal containing all data belonging to all subcarriers the OFDM carrier signal This signal will be used to modulate a Radio Frequency RF carrier 4
41. boards Similar to the TVRX board this is also a receive only It is a complete receiver system for 800 MHz to 2 4 GHz with a 3 5 dB noise figure The DBSRX features a software controllable channel filter which can be made as narrow as 1 MHz or as wide as 60 MHz The DBSRX is MIMO capable and can power an active antenna via the SMA E RFX Daughterboards The RFX family of daughterboards is a complete RF transceiver system They have Independent local oscillators RF synthesizers for both TX and RX which enables a split frequency operation Also it has a built in T R switching and signal TX and RX can be on same RF port connector or in case of RX only we can use auxiliary RX port Most boards have built in analog RSSI measurement All boards are fully synchronous design and MIMO capable Table 1 2 shows USRP daughter boards currently available from Ettus research Name Functionality Frequency range MHz Cost Minimum power From To mw Basic Rx Receiver 2 to 300 75 00 Basic Tx Transmitter 2 to 200 75 00 LFRX Receiver 0 to 30 COO ll 5 LFTX Transmitter 0 to 30 75 00 li Re 16 TVRX Receiver 50 to 70 100 00 DBSRX Receiver 800 to 2400 150 00 RFX400 Transceiver 400 to 500 250 00 RFX900 Transceiver 800 to 1000 275 00 200 RFX1200 Transceiver 1150 to 1450 275 00 200 RFX1800 Transceiver 1500 to 2100
42. ce 4 12 bits y 12 bits stream of data bits hits Figure 4 5 OFDM packet structure Now moving to construction of the packet the following steps show how packet is constructed Calculate CRC32 of a payload from cre function and append it to payload Make header function that take a parameter length of payload with CRC and offset start of window of size 512 which take a known pseudo random number from generation vector Append a character U to a payload with CRC which separate between packets Make whitening Scrambling for payload with CRC and character U Append scrambled data with the header and send it to queue of messages The end of Send_PKT function will put the data that needs to be outputted in the end of these queue of messages and the first module in the modulator chain will access to this data automatically 4 5 3 Blocks of OFDM modulator 1 OFDM Mapper Itis defined in digital_ofdm_mapper_bcv cc Main aim Mapping the incoming data messages into OFDM symbols 39 Inputs and Outputs It has one indirect input which is the pushed messages And has two outputs the first is a stream of mapped OFDM symbols of type gr_complex while the other output is an array of characters flags each indicates the beginning of the first OFDM symbol to ease the process of adding the Pre amble Steps of operation 1 Checking and Preparation e Check and compare between occupie
43. ceived data complex symbol by symbol calculates the minimum Euclidean distance and returns this index of nearest symbol Demapper Takes the coming data and produces the output demapped data The demodulation of OFDM symbols are performed using PLL The demapper block not only produces demapped data but it returns the number of bytes produced as well It makes that by passing through each symbol bit in our case demaps it and stuff it into a byte each byte will contain eight symbols After that it pushes this byte a byte by byte into the output vector As explained before in Mapper Oe same step in Mapper is done here too e Checking and Preparation Checking that occupied carriers lt FFT length Otherwise the process couldn t be completed Then filling out carriers to occupied carriers length by starting with the string FE7F with equivalent binary 1111111001111111 you know Then a loop is considered to pad F 1111 from both sides till reaching a maximum of 7 bits is remaindered At this case we pad ones too from both sides with the remaindered length and it will be in equivalent for odd number of bits e g 7 bits will result in padding four bits to the left and three bits to the right After that padding zeros for the remainder length of the FFT size from both sides too is done Finally passing on each occupied bit a bit represents ONE and push it in the map with a known index Of course it was done after moving t
44. d carriers length and FFT length e Filling out carriers to occupied carriers length e Padding zeros to the remainder of FFT length e Store the used subcarriers in the subcarrier table with known indexes 2 M ary Modulation Scheme and Constellation Table e Identifying the used modulation scheme with the M ary concept e Build the constellation table by generating random symbols for the pre defined modulation scheme and insert them in a table to choose from them later 3 The Mapping Process itself and building the symbols e Checking on EOF packet is finished or not gt If the packet is still in consuming process and not finished yet then gt Pass on the messages message by message The flag indicates the first OFDM Symbol is raised at the beginning of consuming the packet Pass on the message byte by byte Pass on the byte bit by bit Then building the symbols is done by mapping each subcarrier from the known indexes to one of the pre defined symbols from the constellation map pre defined too gt Different cases were handled to be able to fit the number of M ary bits gt Still loop till the packet goes finishing gt If the packet is finished VVV WV gt EOF is raised gt Then go and execute the work function again as long as more packets exist e Check if there is more messages or not Deeply into previous tree Actually the most effective thing that should be kept in mind always is the proce
45. d edge filter which the derivative of the matched filter in frequency by Take a quarter sine wave at the point of the matched filter s roll off if it s a raised cosine the derivative of a cosine is sine Then extend this sine wave by another quarter wave to make a half wave around the band edges is equivalent in time to the sum of two sinc functions as shown in figure above After creating this baseband filter which its bandwidth is determined by the excess bandwidth e g roll off factor of the modulated signal we form band edge filters by not only spinning the baseband filter up and down to the right places in frequency but also by normalizing the power in the two filters e Performing the dot product of the output modulated samples with the two filters e The error signal is the difference between energies of the upper and down output from the result of the dot product e Then adjust the spectrum of modulated data in the right frequency by entering the error signal in the control loop to correct the frequency offset e Maximum and minimum frequencies that adjust by error signal in the control loop are 2 pi 2 samps_per_sym and 2 pi 2 samps_per_sym respectively Oh the C flavor with some communications algorithms looks different We expect your smile right now 29 3 Poly phase filter bank Its main aim is timing recovery It exists in the file gr_pfb_clock_sync_ccf cc It is expected that samples of the received s
46. d one will receive the character that marks the beginning of the frame If it is the case it will buffer the OFDM data symbol output a preamble and then output the buffered data symbol Flow chart is shown in figure 4 6 Idle State Op Symbol Preamble WES First Payload State l YESI Payload State Figure 4 6 Flowchart of OFDM insert preample algorithm There are 4 states to insert preamble for each data packet coming from previous block 42 gt Idle state checks the LSB of each char equal to 1 or not if not eat on input symbol gt Preamble state outputs a specific number of OFDM symbols known symbols as a preamble gt First payload state runs after adding the preamble symbols which copies the first OFDM symbol of the payload from the input to the output gt Payload state copies an OFDM symbol from input to output then checks the LSB corresponding to the next symbol if 1 or not if 1 repeat the process starting from preamble state 3 FFT Reverse or iF FT It is defined in gr_fft_vec cc Main aim Computing 1FFT of the produced symbols with pre ample Inputs and Outputs It has one input and one output It takes a vector of complex values and computes the iFFT and it represents the output Operation gt Orthogonality is realized with FFT iFFT whose output is a sum of orthogonal signals Recall the basic functions for an IFFT are N orthogonal subcarriers each have a dif
47. dev libcppunit dev libboost all dev libusb dev fort77 sdec sdcc libraries libsd11 2 dev python wxgtk2 8 git guile 1 8 dev libqt4 dev python numpy ccache python opengl libgsl0 dev A python cheetah python lxml doxygen qt4 dev tools libqwt5 qt4 dev libqwtplot3d qt4 dev pyqt4 dev tools python qwt5 qt4 sudo apt get install git core cmak 2 Download and install UHD from git git clone git code ettus com ettus uhd git cd uhd host mkdir build cd build cmake make make test sudo make install cd sudo gedit bashrc write this before append to the history file don t overwrite it LD_LIBRARY_PATH for uhd export LD_LIBRARY_PATH LD_LIBRAR Y_PATH usr local lib LD_LIBRARY_PATH sudo ldconfig 3 Download and install GNURADIO from git git clone git gnuradio org gnuradio git cd gnuradio mkdir build cd build cmake make make test sudo make install sudo Idconfig cd sudo gedit bashrc put it after UHD path python path for gnuradio export PYTHONPATH P YTHONPATH usr local lib python2 6 dist packages python path for gnuradio sudo ldconfig sudo apt get install gnuradio companion then reboot Ubuntu 4 Configure SD card by fpga amp firmware e format SD card in windows then open ubuntu again and insert it by reader or in laptop e cd home uhd host utils e open gui burner by sudo usrp2_card_burner_gui py e download the lateast release of uhd binary image and select imag
48. e only then downlaod tar then extract it in home uhd host utils e after open burner select usrp2_fpga bin amp usrp2_fw bin and select dev sdb and burn it e unmount SD card and put it in usrp2 5 Configuration usrp2 and Open GRC e configure ethO address 192 168 10 1 subnet 255 255 255 0 gateway 192 168 10 2 e sudo ping 192 168 10 2 to check usrp2 send echo response e sudo uhd_find_devices Reading usrp2 and print some information serial number version e sudo uhd_usrp_probe Reading name of Rf_daughterboard and range of frequencies that operate it e sudo gnuradio companion 1 5 Tools 1 5 1 GRC GRC is a graphical tool which provides a user interface that lets us create signal flow graphs and activates its source code This graphical interface by means of graphical blocks allows us to set the input parameters which are taken by the source code of each block in order to generate a signal flow and to visualize the signal at every step of the block chain using graphical sinks There are mainly four kinds of blocks 1 Source blocks Their main functionality is to generate an output signal by means of some input parameters For this reason these blocks have no input signal There are many types of sources depending on the number of output ports data type vector lengths etc 2 Sink blocks In this case there is no output signal Sink blocks receive an input signal with a specific data type and length
49. e reader not feeling afraid of GNU Radio The implementation of digital systems which in fact more sophisticated than analog comes in chapter 3 Actually in chapter 4 we reach our aim of this project which is implementing OFDM complete systems you can consider that chapter 2 and chapter 3 as motivation for the reader to be able for understanding OFDM Successful transmission and reception of OFDM packets have been accomplished with reasonable BER Also we discussed in this chapter different issues affecting the whole process In the end chapter 5 summarizes the work done and provides some suggestions for the future development of this project 1 3 Software Defined Radio 1 3 1 Definition SDR is a concept that has been used since the early nineties Its original purpose was the creation of a device radio capable of emulating many radios working at different frequencies In addition it can tune to any frequency band transmit and receive different modulations and different physical parameters across a large frequency spectrum by using a programmable hardware and powerful software An alternative definition for is a collection of hardware and software technologies that enable reconfigurable system architectures for wireless networks and user terminals SDR provides an efficient and comparatively inexpensive solution to the problem of building multi mode multi band and multi functional wireless devices that can be enhanced using software up
50. e the signal processing tasks that have always been done by dedicated devices Cheap computers like the ones we use at home are now able to do the necessary computation that these dedicated devices are doing This is what Software Defined Radio SDR is all about The translation of the signal processing into software run by a regular computer opens up a huge number of possibilities at an affordable price Now we can access to all the parameters that were embedded and invariable before Thanks to SDR now we can analyze and change every value of the system This project will try to get a grip of the state of the art in both wireless communication technology and SDR projects With that objective in mind in this project an implementation of a wireless communications system will be made For the physical layer of this system Orthogonal Frequency Division Multiplexing OFDM was chosen as the transmission multiplexing method This choice has been made because of the advantages that OFDM has shown in terms of channel capacity It has proven its importance by gaining relevance in two of the most important technologies for the ge generation of wireless communications WiMAX and LTE The software toolkit that has been used for the implementation of the prototype has been GNU Radio an open source project that is being used by many researchers and manufacturers all over the world and that is growing steadily in source code available and in active members and project
51. erwise the loop is still considered to produce OFDM symbols Oh we are done GOD I cannot believe it I have been writing this stuff for hours 2 OFDM Insert Preamble Itis defined in gr_ofdm_insert_preambles cc 4 Main aim Inserting pre modulated symbols Preamble before each payload to ease the synchronization operation specially knowing start of the frame Inputs and Outputs It has two inputs and two outputs The first input is a stream of vectors of gr complex fft_length which are the modulated symbols of the payload And the second input is a stream of characters the LSB of each indicates whether the corresponding symbol in first input is the first symbol of the payload or not It is 1 if the corresponding symbol is the first symbol otherwise it is 0 While the first output is a stream of vectors of gr_complex fft_length these include the preamble symbols and the payload symbols And the second output is a stream of characters the LSB of each indicates whether the corresponding symbol in the first input is the first symbol of a packet 1 e the first symbol of the preamble It is 1 if the corresponding symbol is the first symbol otherwise it is 0 Operation and flowchart The information in the preamble is a known sequence of symbols that is constructed from a vector of ones and negative ones that has been randomly generated and stored in the file ofdm py The first input port will contain OFDM symbols and the secon
52. ess The block pfb_ clock sync ccf is used to build the two filter banks as mentioned before It is also used to perform the whole timing recovery process applying the input to the both filter banks estimating error and tuning to the right filter bank The implementation of this block is found in the file gr_pfb_clock_sync_ccf cc The input to this block is the complex signal in baseband The timing recovery process is applied to it and then the output is the complex signal corrected 4 Costas loop Its main aim is phase recovery It is responsible for the correction of the phase error The implementation of this block is found in the file gr_costas_loop_cc cc The input to this block is the complex received signal This block considers that the frequency offset correction and the timing recovery is performed In BPSK the real part of the output signal is the baseband BPSK signal and the imaginary part is the error signal This error signal the imaginary part is used to estimate a value for the phase error Then this value is updated through the control loop Two cases will arise e Positive error phase correction in CW direction signal e Phase_error e Negative error phase correction in CCW direction signal e The error signal should be multiplied by the real part or the sign of the real part to perform sign correction For more illustration look at figure 3 12 j phase_error Figure 3 12 a Phase error in CCW direc
53. ferent frequency and the lowest frequency is DC and the iFFT output is the summation of all N sub carriers Thus the IFFT block provides a simple way to modulate data onto N orthogonal subcarriers The block of N output samples from the 1FFT make up a single OFDM symbol gt The block itself is used both for the computation of the FFT and the iFFT just by editing its parameters It also allows us to set other parameters such as the window used if the FFT should be shifted and FFT size 4 OFDM cyclic prefix It is defined in gr_ofdm_cyclic_prefixer cc Main aim Adding cyclic prefix to OFDM symbols Inputs and Outputs It has one input and one output It the OFDM symbols as its input resulting in an output symbols with cyclic prefix Operation After the fourier transformation the redundancy belonging to the cyclic prefix must be added to the OFDM signal It does a very simple job It takes the OFDM symbol from its input port and copies the last symbols the 43 number of symbols to copy is specified in the cp_size parameter at the beginning of the symbol thus creating a symbol with a size that is the sum of the size of the entered symbol and the size of the cyclic prefix 4 6 OFDM demodulator 4 6 1 Motivation As shown in figure 4 7 the demodulator consists of many C functions callable by Python Like the modulator it is firstly defined in the file ofdm py The inner structure of the demodulator consists of two ba
54. ffort A large community of developers and users have contributed to a substantial code base and provided many practical applications for the hardware and software The powerful combination of flexible hardware open source software and a community of experienced users make it the ideal platform for your software radio development 1 6 2 Motherboard and internal construction Figure 1 3 shows a typical graph for USRP Motherboard The USRP has 4 high speed analog to digital converters ADCs each at 12 bits per sample 64MSamples sec There are also 4 high speed digital to analog converters DACs each at 14 bits per sample 128MSamples sec These 4 input and 4 output channels are connected to an Altera Cyclone EP1C12 FPGA The FPGA in turn connects to a USB2 interface chip the Cypress FX2 and on to the computer The USRP connects to the computer via a high speed USB2 interface only and will not work with USB1 1 So in principle we have 4 input and 4 output channels if we use real sampling However we can have more flexibility and bandwidth if we use complex IQ sampling Then we have to pair them up so we get 2 complex inputs and 2 complex outputs The USB controller contains the firmware that defines its behavior and the USB endpoints The firmware also takes care of loading the FPGA bit stream The FPGA handles the high bandwidth computations and reduces the data rate to something we can send over the USB 2 0 The Analog Device chip is
55. go to have Header state which calls demapper function again to produce bytes again A Side note as said before the demodulation of OFDM symbols are performed using PLL To understand more let us assume e px is the phase locked loop coefficient px Ipxle e Es the accumulated error e fis the frequency shift e is the phase gain e f is the frequency gain e is the equalizer gain e Xy be the closest constellation point to Yp Also X Ci when euclidean distance metric is minimum i e min Y C where C are from the constellation points 1 lt i lt M of size M Where Y is the received symbols after previous module operation Algorithm Step 1 Initialize 9 0 e 0 f 0 a 0 25 fy 7 4 eg 0 05 Step 2 Initialize p p1 p2 pel where py po py 1 04 0 0 Step 3 Pass on all received subcarriers and perform the following Update with previous calculated PLL coefficients Yp Ye xe X pp and phase error Update accumulated error and making phase ecc Vox Xp ee lele correction gradually by multiplying with conjugate of nearest symbol Also check if Yell gt 0 001 ifyesdo px px g X Xk Yk pr otherwise it is considered as not useful and no correction is needed f Note that P Ph Ey X X Y D represents updating of PLL coefficients 53 5 tep 4 Ue argle take phase of accumulated error
56. grades As such SDR can really be considered an enabling technology that is applicable across a wide range of areas within the wireless industry SDR enabled devices can be dynamically programmed in software to reconfigure the characteristics of equipment In other words the same piece of hardware can be modified to perform different functions at different times SDR performs significant amounts of signal processing in a general purpose computer or a reconfigurable piece of digital electronics or the combination of both SDR is where all the signal manipulations and processing works in radio communication are done in software instead of hardware So in SDR signal will be processed in digital domain instead in analog domain as in the conventional radio The digitization work will be done by a device called the Analog to Digital Converter ADC Fig 1 1 shows the concept of Software Defined Radio This figure shows that the ADC process is taking place after the Front End FE circuit FE is used to down convert the signal to the lower frequency called an Intermediate Frequency IF this is needed due to the limitation of the speed of current Commercial of The Shelf COTS ADC The ADC will digitize signal and pass it to the baseband processor for further processes demodulation channel coding source coding and etc In conventional radio all this 2 processes are done in hardware In this project we seek to explore the viability of using GNU Rad
57. he root raised cosine filtering and performs a differential coherent detection or phase shift demodulation The input data consists of a complex sampled signal at baseband frequency and the output is a big endian stream of bits packed 1 bit Byte The parameters that are necessary here are the number of samples per symbol the excess in bandwidth that refers to the roll off factor of the root raised cosine filter the Costa s alpha parameter the mu factor and its gain etc These parameters will be explained on greater detail later As shown in figure 3 6 the block diagram of the DPSK demodulator consists internally of seven sub blocks each will be discussed in detail Rec signal Automatic gain control Frequency locked loop Poly phase filter bank Costas loop Estimated data Differential decoder Constellation receiver mapper Figure 3 6 Data flow through the demodulation block 1 AGC It stands for Automatic Gain Control In this block the signal is first multiplied by a constant to scale the signal from full range to 1 Then an automatic gain control is performed The procedure consists of a calculation of the maximum power level among the samples Until this point a simple rescaling is performed The AGC works by measuring a smoothed power history then adjusting an output gain to achieve constant power Its important parameters are Attack rate How fast the AGC decreases
58. hich indicates that the block takes input as 8 byte complex values and produces an output in 4 byte floating point values Similarly the block gr_multiply_const_vss would take a vector of 2 byte short integers and produce an output in the same format Other possible data types include b for 1 byte integers and i for 4 byte integer values A new GNU Radio signal processing block is defined 5 by deriving from the base class gr_block or one of its subclasses gr_sync_decimator gr_sync_interpolator gr_sync_block or gr_hier_block2 in C Then a SWIG interface is defined for this block which enables it to be constructed and connected from Python At the core of the signal processing block are two member functions forecast and work Forecast returns an estimate of how many units of the input data is required for this module to produce a given number of output units and work is the function that does the actual computation on the input data and produces an output This signal processing blocks framework abstracts away the complexity of how one might schedule the work of multiple signal processing blocks on the computer 1 4 2 GNU Radio installation In this project Ubuntu 10 10 was used and we are going to show steps of installation of GNU on it 1 Install the following pre requisite packages from Terminal sudo apt get y install libfontconfig1 dev libxrender dev libpulse dev A swig g automake autoconf libtool python dev libfftw3
59. ied_tones 200 bw float occupied_tones float fft_length 2 0 tb bw 0 08 AM Blocks AM self chan_coeffs gr firdes low_pass 1 0 gain 1 0 sampling rate bw tb midpoint of trans band tb width of trans band er firdes WIN_HAMMING filter type 65 if name__ main def get_samp_rate self return self samp_rate def set_samp_rate self samp_rate self samp_rate samp_rate A parser OptionParser option_class eng_option usage prog options options args parser parse_args tb top_block tb Run True fig pylab figure 1 sp fig add_subplot 1 1 1 p2 sp plot tb chan_coeffs b linewidth 2 label channel filter taps sp legend pylab show 4 8 3 PN based synchronization Code is divided into two parts First one is responsible for writing the needed data to certain files The second one is responsible for reading the data and plotting it First code writing data to files from gnuradio import digital from gnuradio import eng_notation from gnuradio import gr from gnuradio eng_option import eng_option from gnuradio gr import firdes from grc_gnuradio import blks2 as grc_blks2 from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import numpy import wx 66 class top_block grc_wxgui top_block_gui def mt self grc_wxgui top_block_gui mt self title Top Block ANN Variables AA self s
60. ignal are not aligned with the data clock used to generate the analog waveform Our approach for solving the timing recovery problem works by setting up two filter banks one filter bank contains the signal s pulse shaping matched filter such as a root raised cosine filter where each branch of the filter bank contains a different phase of the filter Figure 3 8 shows a typical four phase poly phase taps The second filter bank contains the derivatives of the filters in the first filter bank Thinking of this in the time domain the output of the first filter bank is shaped as the autocorrelation function We want to align the output signal to be sampled at exactly the peak of the autocorrelation function Amplitude 100 120 140 160 180 Tap Number 20 40 60 Figure 3 8 Four phase poly phase taps The derivative of the autocorrelation contains a zero at the maximum point of it Furthermore the region around the zero point is relatively linear We make use of this fact to generate the error signal This error signal is used to tune the selection of the filters in the filter bank depending on the output of the derivative filter bank as shown in figure 3 9 Positive Slope Advance Timing ei A Derivative Zero Slope Negative Slope Hold Timing Retard Timing Figure 3 9 Estimating error process 30 A Positive 41 Positive But using the derivative filter Derivative i Ki Correlation
61. io an open source SDR implementation and the Universal Software Radio Peripheral USRP an SDR hardware platform to transmit and receive the OFDM signal with BPSK modulation Quality of Service QoS in terms of Packet Received Ratio PRR on the data transmitted will then be investigated and analyzed Software defined radio nowadays is a tool that helps the wireless and mobile communications industry in many aspects For your knowledge the term SDR was introduced by Joseph Mitola from MITRE Corporation in 1991 His first paper on SDR was published in 1992 at IEEE National Telesystems Conference Though the concept was first proposed in 1991 software defined radios have their origins in the defense sector since the late 1970 s in both the U S and Europe Software Figure 1 1 Module diagram of a SDR sender and receiver 1 3 2 Basic principle and difference to analog radios The basic principle of SDR is the reduction to the minimum of the hardware dedicated to signal processing parts and its translation into software that should be runnable by an all purpose commodity PC The signal should be generated digitally and dealt with in the PC as much as possible undergoing modulators filters FFT blocks and even amplifiers all of them done in software until the signal is ready to be sent Then the software gives place to the hardware which has the function of transforming the digital samples in an analog signal and modulating the baseband
62. ions args parser parse_args tb top_block tb Run True fig pylab figure 1 sp fig add_subplot 1 1 1 p2 sp plot tb gr_vector_sink_x_0 data b linewidth 1 2 label preamble symm real part fig2 pylab figure 2 sp2 fig2 add_subplot 1 1 1 p2 sp2 plot tb gr_vector_sink_x_1 data b linewidth 1 2 label preamble symm imag part sp legend pylab show 4 8 2 BW and channel filter from gnuradio import digital from gnuradio import eng_notation from gnuradio import gr from gnuradio import window from gnuradio eng_option import eng option from gnuradio gr import firdes from gnuradio wxgui import fftsink2 from grc_gnuradio import blks2 as grc_blks2 62 from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import numpy import wx class top_block grc_wxgui top_block_gui def init self grc_wxgui top_block_gui init self title Top Block HE Variables DEE self samp_rate samp_rate 32000 E E H E E a Blocks E E E E E E E H EH HE HEE EH E EHEHEH self wxgui_fftsink2_0 fftsink2 fft_sink_c self GetWin baseband_freq 0 y_per_div 10 y_divs 10 ref_level 0 ref_scale 2 0 sample_rate samp_rate fft_size 1024 fft_rate 15 average False avg alpha None title FFT Plot peak_hold False self Add self wxgui_fftsink2_0 win 63 self random_source_x_0 gr vector_source_b map int numpy random randint 0 2
63. large PAPR which results channels in reducing the power efficiency or 2 For slow channels loading is easier the RF amplifier than single carrier 3 Sensitive to doppler shift 3 OFDM is robust against narrow band interference as it will affect only portion of subcarrier and may overcome it by using coding Table 4 1 Advantages and disadvantages of OFDM comapred with SC 4 3 Objective This phase from our project implements a communication system that uses an OFDM multiplexing method for various reasons First of all the prototype built for this project 35 had the objective of simulating the lower layers of the ones used by next generation mobile communications 4G and some of the most important technologies in that field which are WiMAX and LTE They both use OFDM as an important part of them WiMAX uses OFDM as its multiplexing method in its physical layer LTE uses OFDM for its downlink which is from the base station to the terminal and a pre coded version of OFDM called Single Carrier Frequency Division Multiple Access SC FDMA OFDM is also used in many other technologies like ADSL digital radio Digital Audio Broadcasting DAB terrestrial digital TV Digital Video Broadcasting Terrestrial DVB T and terrestrial mobile TV Digital Video Broadcasting Handheld DVB H Again the main goal is achieving a complete OFDM transceiver transmitter and receiver through USRP 2 and USRP N 210 boards one
64. lation table which we will map from is done here by generating random symbols depending on the pre defined modulation scheme For our case generated symbols may be 1 or 1 BPSK 3 The Mapping Process itself and building the symbols Then the mapping process comes now it seems late a bit but preparations are important too Building a single symbol First Initialize all bins to zero to set unused carriers After that checking that we still have messages is done and we don t exceed number of occupied carriers so far Then taking a single message and working on it byte by byte via d_msgbytes then for each byte passing through it too bit by bit Then test to make sure we can fit bits of the M ary then take the number of bits of M ary at a time from the byte to add to the symbol Also for each taken bits enter the constellation table and map it them to one of the pre defined symbols But what would happen if we couldn t fit the number of bits of M ary If we can t fit the number of bits of M ary store them for the next byte by saving residual bits of this message where residual bits must be smaller than number of bits of M ary Take the residual bits fill out with info from the new byte and put them in the symbol also from the constellation table When will we end Check on End Of File EOF must be done each iteration if EOF 1 then packet was finished free message and go to work again to consume another packet Oth
65. ming synchronization for OFDM In Communications IEEE Transactions on 45 12 1997 Pp 1613 1621 DOI 10 1109 26 650240 URL http dx doi org 10 1109 26 650240 3 Alexandra Martinez Torio Software Defined S Band Ground Station Transceiver for Satellite Communications Institute of Telecommunications E389 Department of Electrical Engineering Vienna University of Technology 4 Tuan Ta Synchronization in OFDM Department of Electrical and Computer Engineering University of Maryland at College Park September 2010 5 Grace R Woo Demonstration and Evaluation of Co channel DPSK Source Separation Master of Science in Media Arts and Science Massachusetts Institute of Technology June 2007 6 Patrick Ellis amp Scott Jaris Advisors Dr In Soo Ahn amp Dr Yuefeng Lu Implementation of Software Defined Radio Using USRP Boards May 10 2011 7 GNU Radio URL http www gnu org software gnuradio 8 Dawei Shen GNU Radio Installation Guide Step by Step May 19 2005 9 Fredric j harris Multirate Signal Processing for Communication Systems 10 Blossom Eric Johnathan Corgan Matt Ettus and Tom Rondeau GNU Radio 2006 Web 27 Feb 2011 11 Fredric harris Band Edge Filtering and Processing for Timing and Carrier Recovery Communications and Signal Processing Institute San Diego State University San Diego California 12 Wen Li Senior System Engineer Advanced Analog P
66. n C 1 5 3 Python Don t get afraid of the name stay tuned the python will not hurt you The story of this mysterious name is at the same time he began implementing python Guido van Rossum was also reading the published scripts from Monty Python s Flying Circus a BBC comedy series from the seventies in the unlikely case you didn t know It occurred to him that he needed a name that was short unique and slightly mysterious so he decided to call the language Python The most important thing in the programming language is the name A language will not succeed without a good name About Python it is a script language is used to connect the signal processing blocks together In Python the necessary signal sources sinks and processing blocks are selected and configured with the correct parameters The flow of data through the flow graph exists out of data in one of the following data types _ Byte 1 byte of data 8 bits _ Short 2 byte integer _ Int 4 byte integer _ Float 4 byte floating point _ Complex 8 bytes As mentioned previously all sources sinks and blocks are implemented as classes in C This results no matter how complicated the radio is in good readable Python code The real heavy load is done in C Figure 1 2 shows an example of a Python code As you can see it is not difficult to create a radio in Python usr bin env python from gnuradio import gr from gnuradio import audio def build_graph
67. n odd frequencies of the preamble and its time domain representation resulting in symmetric signal This Simulation is done for the PN sequence existing in the file ofdm py which represents the preamble after adding zeros in odd frequencies Figure 4 17 shows this symmetry in time domain in real part as well as imaginary part preamble symm preamble symm real part 100 200 300 400 500 600 Real part Imaginary part Figure 4 17 Preample symmetry 4 7 2 BW and channel filter The following figure 4 18 shows the frequency domain representation of t transmitted data with fft_length 512 occupied_tones 200 cyclic_prefix 128 and Sampling rate 32 k sample second 56 B W calculations Symbol time per subcarrier 512 32000 second 0 016 second What is the reciprocal of previous quantity Left for the reader Bandwidth of the signal occupied carriers Symbol time per subcarrier Then B W 200 0 016 12 5 KHz as shown in the figure Or you may simply say that the BW Sample rate EE 12 5 KHz as in previous method Amplitude dB 100 10 20 30 40
68. ncy USB Universal Service Bus LO Local Oscillator FM Frequency Modulation SNR Signal To Noise Ratio FIR Finite Impulse Response BW Band Width FDM Frequency Division Multiplexing CDM Code Division Multiplexing PAPR Peak To Average Power Ratio SC FDMA Single Carrier Frequency Division Multiple Access DAB Digital Audio Broadcasting DVB T Digital Video Broadcasting Terrestrial DVB H Digital Video Broadcasting Handheld CP Cyclic Prefix QAM Quadrature Amplitude Modulation PSK Phase Shift Keying IFFT Inverse Fast Fourier Transform ISI Inter Symbol Interference ICI Inter Carrier Interference FFT Fast Fourier Transform BPSK Binary Phase Shift Keying FEC Forward Error Correction LSB Least Significant Bit CRC Cyclic Redundancy Check PLL Phase Locked Loop GRC Gnu Radio Companion BER Bit Error Rate MSB Most Significant Bit FLL Frequency Locked Loop CW Clock Wise CCW Counter Clock Wise QPSK Quadrature Phase Shift Keying EOF End Of File PN Pseudo Noise Acknowledgments First and foremost thanks God the most merciful and most beneficent to who we relate any success in achieving any work in our life It gives us the greatest pleasure to express our deepest attitude and warmest thanks to Dr Mohamed Khairy Lecturer of Communications Electronics and Electrical Communications Department Faculty of Engineering C
69. nerates the corresponding Python code and library so that these classes and functions can be called from Python Most default signal blocks are already created in the GNU Radio or by third parties So you only touch the C environment to create your own special signal processing blocks It requires some experienced C skills to create these blocks If you do not have much C experience try to use signal blocks that are already part of the GNU Radio library or try to find code from third parties 1 6 USRP 1 6 1 Motivation The Universal Software Radio Peripheral or USRP pronounced usurp was designed as a low cost board solely for the purpose of running GNU radio applications and allowing general purpose computers to function as high bandwidth software radios Fully developed by Matt Ettus it is a very flexible platform and can be used to implement real time applications In essence it serves as a digital baseband and IF section of a radio communication system You may say it is the bridge between the software world and the RF world The basic design philosophy behind the USRP has been to do all of the waveform specific processing like modulation and demodulation on the host CPU All of the high speed general purpose operations like digital up and down conversion decimation and interpolation are done on the FPGA The true value of the USRP is in what it enables engineers and designers to create on a low budget and with a minimum of e
70. nt part for the GNU Radio users All the ADCs and DACs are connected to the FPGA This piece of FPGA plays a key role in the USRP USRP 2 system Basically what it does is to perform high bandwidth math and to reduce the data rates to something you can squirt over USB2 0 GE ON USRP USRP 2 respectively The standard FPGA configuration includes digital down converters DDC implemented with 4 stages cascaded integrator comb CIC filters Also it includes digital up converters DUC implemented with 4 stages cascaded integrator comb CIC filters CIC filters are very high performance filters using only adds and delays The DDC and DUC each contain 2 halfband filters The high rate one has 7 taps and the low rate one has 31 taps For spectral shaping and out of band signals rejection 4 Daughter boards On the mother board there are two slots One of these slots is for TX and the other is for RX Each daughter board slot has access to ADC DAC The daughter boards are used to hold the RF receiver interface or tuner and the RF transmitter Every daughterboard has an DC EEPROM 24LC024 or 24LC025 onboard which identifies the board to the system This allows the host software to automatically set up the system properly based on the installed daughterboard The EEPROM may also store calibration values like DC offsets or IQ imbalances If this EEPROM is not programmed a warning message is printed every time USRP software is run There are several kind
71. ntenna system e Tier 3 Ideal Software Radio ISR ISRs provide dramatic improvement over an SDR by eliminating the analog amplification or heterodyne mixing prior to digital analog conversion Programmability extends to the entire system with analog conversion only at the antenna speaker and microphones e Tier 4 Ultimate Software Radio USR USRs are defined for comparison purposes only It accepts fully programmable traffic and control information and supports a broad range of frequencies air interfaces amp applications software It can switch from one air interface format to another in milliseconds use GPS to track the user location store money using smartcard technology or provide video so that the user can watch a local broadcast station or receive a satellite transmission Cognitive radio CR is a form of wireless communication in which a transceiver can intelligently detect which communication channels are in use and which are not and instantly move into vacant channels while avoiding occupied ones This optimizes the use of available radio frequency RF spectrum while minimizing interference to other 1 4 GNU Radio 1 4 1 Concepts and architecture GNU Radio is primarily developed using the GNU Linux operating system but Mac OS and Windows are also supported In GNU Radio a radio system is represented as a directed signal flow graph where graph vertices are known as signal processing blocks and edges indicate a connection
72. o bit domain from string domain e M ary modulation Here identifying number of M ary bits used is done Besides that it checks whether number of produces symbols are true two in BPSK and their values are true as well But how all work together from the name all work through the WORK function Basically it is the most important function in that block and every one too As its name reveals it makes all things all work together It takes the coming symbols from previous module and produces the output in accordance with previous functions slicer demapper PLL and discussions Inside thw Work function the following steps are done 1 Sync search Looks for flag indicating beginning of the packet If successfully found set up for header decode and go to enter have sync which discussed before 2 Have sync Only demodulate after getting the preamble signal It calls the demapper function and inherently calls the slicer function which discussed 52 before both returning demapped output symbols and number of bytes produced Firstly it gets the header if it is OK it completes the packaging of received packet and finally pass it to the output queue In contrary if header isn t correct it represents a packet loss So we go out of the loop and see another packet The process is being repeated till all received data are consumed 3 If one packet is still not filled with bytes returned We
73. ption usage prog options options args parser parse_args tb top_block tb Run True Second code Read data from files and plotting it import pylab from gnuradio import eng_notation from gnuradio import gr from gnuradio eng_option import eng_option from gnuradio gr import firdes from grc_gnuradio import wxgui as grc_wxgul from optparse import OptionParser import wx class top_block grc_wxgui top_block_gui 68 if def mt self grc_wxgui top_block_gui mt self title Top Block ANN Variables AM self samp_rate samp_rate 32000 ANN Blocks ANN self gr_vector_sink_x_0 gr vector_sink_b 1 self gr_throttle_O gr throttle gr sizeof_char 1 samp_rate self gr_file_source_0 gr file_source gr sizeof_char 1 current directory False rer RER RER Connections H PATITE self connect self gr_file_source_0 0 self gr_throttle_0 0 self connect self gr_throttle_0 0 self gr_vector_sink_x_0 0 def get_samp_rate self return self samp_ rate def set_samp_rate self samp_ rate name self samp_ rate samp_rate U main 1 parser OptionParser option_class eng_option usage prog options options args parser parse_args tb top_block tb Run True fig pylab figure 1 sp fig add_subplot 1 1 1 69 p2 sp plot tb gr_vector_sink_x_0 data 50000 r linewidth 2 label peaks signals sp legend
74. r than the expected delay spread such that multipath component for one symbol can t interfere with the next symbol The guard time consists of no signal However the problem of ICI Inter Carrier Interference would arise ICI is the cross talk between different subcarriers which means they are no longer orthogonal To eliminate ICI OFDM symbols are cyclically extended in the guard time as shown in figure 4 3 This ensures that delayed replicas of the OFDM symbol always have an integer number of cycles within the FFT interval And that we call CP Figure 4 1 Signal in the presence of ISI Figure 4 2 Signal in the presence of GT tZ Figure 4 3 The effect of adding a cyclic prefix to the signal 37 4 5 OFDM modulator 4 5 1 Motivation The OFDM modulator exits in python file ofdm py Also the OFDM demodulator exists in the same python file OFDM modulator modulates an OFDM stream based on the options FFT length occupied tones and cyclic prefix length this block creates OFDM symbols using a specified modulation scheme here is BPSK It consists of four main blocks OFDM Mapper OFDM Insert Preamble FFT Reverse and OFDM Cyclic Prefix each of them will be discussed in details in next subsection As shown in figure 4 4 it interconnects some blocks that are defined in C GN A CF adder ofdm_mapper insert_preambles cyclic_prefixer Cupar C C Figure 4 4 OFDM Modulator Hierarchy Looking at the definition of the mod
75. roduct Group and Jason Meiners Introduction to phase locked loop system modeling 13 Thomas Schmidt CRC Generating and Checking 14 Kushal Shah Audio Streaming over FM band between USRP1 and USRP2 Phase IT Project Report 75 15 Oussama Sekkat The FPGA 16 Arief Marwanto Mohd Adib Sarijari Norsheila Fisal Sharifah Kamilah Syed Yusof Rozeha A Rashid Experimental Study of OFDM Implementation 17 Jesper M Kristensen ppt GNU Radio An Introduction 18 Hlaing Minn Member IEEE Vijay K Bhargava Fellow IEEE and Khaled Ben Letaief Fellow IEEE A Robust Timing and Frequency Synchronization for OFDM Systems 19 Firas Abbas Simple User Manual for Gnuradio 3 1 1 20 Andreas Muller DAB Software Receiver Implementation 21 Fredric J Harris Senior Member IEEE and Michael Rice Senior Member IEEE Multirate Digital Filters for Symbol Timing Synchronization in Software Defined Radios 22 Tuan Ta Synchronization in OFDM 23 J eff Feigin Practical Costas loop design 24 Dawei Shen Writing a Signal Processing Block for GNU Radio 25 http gnuradio org doc doxygen index html 26 http gnuradio org doc doxygen index html 27 Firas Abbas Simple Gnuradio User Manual http www scribd com doc 9688074 Simple Gnuradio User Manual v10 28 http www tutorialspoint com python index htm 29 GRC tutori
76. rough the demodulation block 27 Figure 3 7 Carrier recovery DOCS isis 28 Figure 3 8 Four phase poly phase EG 30 Figure 3 9 Estimating error PIC 30 Figure 3 10 Taking the derivative output only is not enough 31 Figure 3 11 Whole system block V ageapt cid idea 31 Figure 3 12 Phase error ii Dour halts ai pi 32 Figure 3 13 Block diagram of costas l00p ss 33 Figure 4 1 Signal in the presence DIS A 37 Figure 4 2 Signal in the presence of GT ii te 37 Figure 4 3 The effect of adding a cyclic prefix to the agnal 37 Figure 44 OFDM Modulator Hierarchy rai ie BA ee 38 Figure 4 5 OFDM E 39 Figure 4 6 Flowchart of OFDM insert preample algorithm 42 Figure 4 7 OFDM demodulator hierarchy sitiada 44 Figure 4 8 Block diagram OFDM receiver block ssssesssesesessesseseresrersersresrerseeseese 45 Figure 4 9 Time domain signal of channel filter 0 eee eeeeeneeceeeeneeeeeeeeeaeees 46 Figure 4 10 Distribution of preamble in frequency domain ooococcnocccnocaninncconanannninnnons 47 vil Figure 4 11 Preamble symmetry in time nina 47 Figure 4 12 Timing metric of Schmid and Cox method 48 Figure 4 13 Block diagram of OFDM symbol synchronization module 48 Figure 4 14 Internal structure of OFDM synchronization block A 49 Figure 4 15 Correlation between of halves iii ii 49 Figure 4 16 State machine representing OFDM frame sink block 51 Figure 4 17 Preamples MU a iris 56 Figure 4 18 Transmitted OFDM signal unie
77. s 57 Figure 4 19 Time domain representation for filter taps sesesessssseesseessessseessseeessee 57 Figure 4 20 Output of the timing Metric 2 2 diane idas 58 Figure 4 21 Output of the matched filter ri ee BA Aes 58 LR en e e e GEE da 58 Figure 4 23 Peak detection Process pia AA 59 viii List of Symbols and Abbreviations SDR Software Defined Radio OFDM Orthogonal Division Multiplexing WIMAX Worldwide Interoperability For Microwave Access LTE Long Tem Evolution USRP Universal Software Radio Peripheral DBPSK Differential Binary Phase Shift Keying IEEE Institute of Electrical and Electronics Engineers QOS Quality Of Service PRR Packet Received Ratio PC Personal Computer RF Radio Frequency SCR Software Controlled Radio ISR Ideal Software Radio USR Ultimate Software Radio CR Cognitive Radio GPS Global Positioning System OS Operating System GRC Gnu Radio Companion FFT Fast Fourier Transform FPGA Field Programmable Gate Array DAC Digital To Analog Converter ADC Analog To Digital Converter DSP Digital Signal Processing GE Giga Ethernet DDC Digital Down Converter DUC Digital Up Converter CIC Cascaded Integrator Comb TX Transmitter RX Receiver AGC Automatic Gain Controller LFTX Low Frequency Transmitter LFRX Low Frequency Receiver VHF Very High Frequency ix UHF Ultra High Freque
78. s of daughter boards available now such as 15 A Basic TX RX Daughterboards Each has two SMA connectors that can be used to connect external up down tuners or signal generators We can treat it as an entrance or an exit for the signal without affecting it Some form of external RF front end is required The ADC inputs and DAC outputs are directly transformer coupled to SMA connectors 50Q impedance with no mixers filters or amplifiers The BasicTX and BasicRX give direct access to all of the signals on the daughterboard interface Each of the Basic TX RX boards has logic analyzer connecters for the 16 general purpose IOs These pins can be used to help debugging your FPGA design by providing access to internal signals B Low Frequency TX RX Daughterboards The LFTX and LFRX are very similar to the BasicTX and BasicRX respectively with 2 main differences Because the LFTX and LFRX use differential amplifiers instead of transformers their frequency response extends down to DC The LFTX and LFRX also have 30 MHz low pass filters for anti aliasing C TVRX Daughterboards This is a receive only daughter board It is a complete VHF and UHF receiver system based on a TV tuner module The RF frequency ranges from 50MHz to 860MHz with an IF bandwidth of 6MHz All tuning and AGC functions can be controlled from software Typical noise figure is 8 dB This board is the only USRP daughterboard which is NOT MIMO capable D DBSRX Daughter
79. s the received tone in frequency domain As expected for a single tone we received two approximately deltas at frequency and its counterpart of sent tone 22 Chapter 3 Digital Environment 3 1 Motivation We are going in this phase of the project to show practical digital modulation scheme s implemented on USRP boards under GRC Successful transmission and reception of bits or packets is successfully accomplished with reasonable BER We will discuss transmission path as well as reception path in details in the coming sub sections 3 2 Transmission Path For transmission path the scenario set up is basically constituted of three main different blocks random vector source DPSK modulator multiply const and and USRP sink as shown in figure 3 1 Now we will discuss the operation of each block Options ID top block Generate Options W GUI UHD USRP Sink Samp Rate Spa 250k Cho Center Freq Hz 1 836 D r Gain dB 0 Vector Source Vector 0 0 0 1 1 0 1 1 Repeat No Multiply Const Constant 3 2168k Figure 3 1 DBPSK transmission path 3 2 1 Random Vector source Random source This block as the beginning of the chain generates a random digital signal It provides us a certain number of byte type samples which values range from O to in our case 128 this number is not significant for the meaning of the measurements Since it is a data source it has no data input but one output
80. s using it The implemented prototype communication system has been a prove of concept that has shown that it is possible to run a communication system of a certain complexity all in software with not much more than commodity equipment The prototype has shown a very good behaviour in some parts of the system such as the synchronization and has also shown weaknesses in other parts like the return channel or the equalization xii Chapter 1 Introduction This chapter gives an overview of the objectives of this project as well as the different tasks that will be necessary to fulfill these objectives The chapter is divided into the motivation section 1 1 and the tasks and contents section 1 2 In the motivation we explain the interests that made us decide for this project introducing also the context in which it has been thought The tasks and contents section will explain the different proposed tasks for this project and will locate them in the different chapters and sections of the project 1 1 Motivation This project has been conceived as the result of putting many ideas in the practice It started with the objective of covering and gaining experience in some fields of interest in wireless communications such as SDR or the implementation of communication systems on USRP boards The first idea that starts shaping this project is the importance that SDR has acquired in the last years As of today it has become a tool that allows researchers gre
81. seband signal and the remaining task is to calculate its instantaneous frequency If we integrate frequency we get phase or angle Conversely differentiating phase with respect to time gives frequency These are the key insights we use to build the demodulator We used the gr quadrature_demod_cf block for computing the instantaneous frequency of the baseband signal We approximate differentiating the phase by determining the angle between adjacent samples 2 De emphasizer The second block in the chain is the deemphasizer deemph is an instance of the class fm_deemph fm_deemph is also a hierarchical block defined in fm_emph py What is de emphasis Let s introduce it briefly It has been theoretically proven that in an FM detector the power of the output noise increases with the frequency quadratically However for most practical signals such as human voice and music the power of the signal decreases significantly as frequency increases As a result the signal to noise ratio SNR at the high frequency end usually becomes unacceptable To circumvent this effect people introduce pre emphasis and de emphasis into FM systems At the transmitter we use proper pre emphasis circuitry to manually amplify the high frequency components and do the reverse operation at the receiver to recover the original power distribution of the signal As a result we effectively improve the SNR 21 In the analog world a simple first order RLC circuit us
82. sic modules corresponding to the two blocks that are defined in ofdm py First one is ofdm_receiver which takes care of the synchronization and equalization of the CRC signal while the second module is ofdm_frame_sink which mainly represents a state machine that de maps the symbols into bits checks the validity of the synchronized frames and sends them to a superior layer by adding them to the queue of received data packets OFDM Demod Modulation BPSK FFT Length 512 Occupied Tones 200 Cyclic Prefix Length 128 SNR 10 ofdm_receiver ofdm_frame_sink Figure 4 7 OFDM demodulator hierarchy Each of them will be discussed later in details Its input comes in most cases from the output of the channel It demodulates a received OFDM stream based on the options FFT length occupied tones and CP length Its output port will contain the demodulated signal And the output data will be sent to a handler via a callback function At the end of the demodulator chain the output data in the output queue 44 the queue of received packets from ofdm_frame_sink module will be sent to a handler via a callback function It is a function running in a separate thread that monitors the queue of demodulated data packets Once a new packet enters this queue it will be taken by this function that will check if data in the packet is correct or not by looking at the CRC32 code Afterwards it will call the previously mentioned callback
83. signal to the desired carrier frequency to be sent The last stage would be sending it to the antenna The concept of SDR is very different to the traditional radios that we have been using until now Traditional radios rely on dedicated hardware for all its functions and each hardware part has a very concrete and fixed function The same processor in the PC used by SDR will take care of all the signal processing and it will be the software the one responsible of dictating the function that will be computed Not having dedicated hardware has a very important advantage in relation to traditional radios All parameters that the radio uses are set in software and they are all configurable through software This makes the development and research of new applications a lot easier faster and cheaper as we can use software radio as a prototype where we can 3 test all kinds of variations and configurations We will not need to make or order hardware in order to try new configurations or variations of any kind Another advantage is the possibility of having a radio that can work as many radios as the same device can use different radio technologies without any change in the hardware This application was one of the first objectives of software radio but for a number of reasons that we will explain in the next paragraphs it is still not a very interesting solution Software radio has some hardware requirements It will require some ADC and DAC hardware as
84. sor_cc The input to this block is the complex received signal after performing phase frequency and timing recovery on it The output is the phase differences 6 Constellation decoder This block is responsible for finding the closest constellation point and taking the decision The decision is taken based on the minimum distance calculated The implementation of this block exists in the file gr_constellation_decoder_cb cc The input to this block is the complex data the phase differences After making decision the output is a byte representing the corresponding constellation point In our case BPSK output will be 1 or 1 7 Symbol mapper It takes the constellation points coming from the constellation decoder and performs symbol mapping The implementation of this block exists in the file map_bb cc Its input is a group of bytes representing the constellation points And finally we get the output as a group of bytes each byte represents one of the transmitted bits 33 3 3 3 Vector sink Graphically or in a file depending on its data format sinks are used to visualized the resulting signal at the end of the chain Input parameters are depending on the kind of input data Particularly in our scenario since we want to see the resulting byte type data after the demodulator block we used a vector sink where the date recovered will be easily shown in python file thus we can calculate BER 34 Chapter 4 OFDM 4 1 Introduction
85. ss was run through consuming one and every packet to produce as many OFDM symbols of FFT length to hold the full packet Thus as long as messages are pushed the OFDM symbols are produced each of FFT size 512 bits you know The concept 40 of occupied carriers is used here where the used carriers are only 200 out of 512 And the remainder is padded with zeros from both sides left and right 1 Checking and Preparation Checking that occupied carriers lt FFT length Otherwise the process couldn t be completed Then filling out carriers to occupied carriers length by starting with the string FE7F with equivalent binary 1111111001111111 you know Then a loop is considered to pad F 1111 from both sides till reaching a maximum of 7 bits is remaindered At this case we pad ones too from both sides with the remaindered length and it will be in equivalent for odd number of bits e g 7 bits will result in padding four bits to the left and three bits to the right After that padding zeros for the remainder length of the FFT size from both sides too is done Finally passing on each occupied bit a bit represents ONE and push it in the map with a known index Of course it was done after moving to bit domain from string domain 2 M ary Modulation Scheme and Constellation Table Identifying the M ary modulation scheme which will be used is done here Actually the used modulation scheme is BPSK Also building the constel
86. the FFT output for the correlation in case there is a large frequency shift in the data It assumes that the fine frequency shift has already been corrected and that the samples fall in the middle of one FFT bin It then uses one of those known symbols to estimate the channel response over all subcarriers and does simple 1 tap equalization on all subcarriers This corrects the phase and amplitude distortion caused by the channel 50 B OFDM frame sink It is defined in digital_ofdm_frame_sink cc It is the second module in the OFDM demodulator It is a state machine that demaps the symbols into bits checks the validity of the synchronized frames from first module ofdm_receiver and sends them to a superior layer by adding them to the queue of received data packets This module has in fact fewer complexities than its preceding one Recall it looks more or less like a state machine as shown in figure 4 16 This state machine has three states the first one is sync search in which the algorithm looks for the flag in the second input indicating the beginning of the frame looks for the flag indicating beginning of packet looking for frame start Flag found frame start frame start flag found flag found e nque ue frame A OK HAVE HEADER Figure 4 16 State machine representing OFDM frame sink block find and check header header OK Once it is found it will arrive in the have sync state
87. the second of preamble which is used for frequency error correction 3 Frequency Modulator It is defined in a C file gr_frequency_modulator_fc cc The input to this block is the frequency error correction value outputted from synchronization block which is T Af T as explained before And T T N where Te is the time between two samples and N is the FFT length So the input is TAf NT Now we want to generate e Z IT which we can use to eliminate the frequenc g q y offset But GEI as we deal with a discrete samples so we want to D generate e7 TAfkTe And this block is used to generate the frequency offset corresponding to each sample 4 OFDM Sampler It is defined in a C file digital_ofdm_sampler cc The ofdm_sampler block will use that frequency offset to sample the signal that comes straight from the channel The output of the ofdm_sampler is the signal sampled in vectors of the size matches to FFT size without existence of CP 5 FFT Forward It is also defined in C in the file gr_fft_vcc cc It takes a vector of complex values outputted from the sampler with FFT size length and computes the FFT 6 Frame Acquisition It is defined in C in the file digital_ofdm_frame_acquisition cc It takes a vector of complex values of the signal in the frequency domain which is the output of the FFT of a received OFDM symbol and finds the start of a frame based on two known symbols It also looks at the surrounding bins in
88. the second state you know Then it will let the preamble through Afterwards the algorithm will demap the symbols and check the bytes corresponding to the header of the data The header is built in a way that the first and the last half are exactly the same so this is the way it validates the header If the header is correct the state machine will move to the next state which is called have header There the algorithm will demap the rest of the frame and insert the resulting data bits in the output queue As explained before that queue is monitored by a thread that will take the data validate it and send the results to the upper layers of the system Throughout the duration of the have sync and have header states the algorithm will constantly look for beginning of frame flags in its 51 input port In case it finds one it will detect an error and reset the state machine to the sync search state Moving to the algorithm a bit Now we will discuss some important but not all functions used such as In enter have sync State is assigned to have sync At the beginning of state have sync resetting PLL is done In enter have header State is assigned to have header As explained before in packet format header consists of two 16 bit shorts in network byte order payload length is lower 12 bits and whitener offset is upper 4 bits Here in have header state the two fields are got Slicer Takes a stream of complex re
89. tion Figure 3 12 b Phase error in CW direction Figure 3 12 Phase error As shown in figure 3 12 a we have phase rotation error in CCW direction So the error signal the imaginary part is positive and the loop generates a phase recovery to rotate the constellation in the CW direction But in figure 3 12 b we have phase rotation error in CW direction So the error signal the imaginary part should be negative in order to generate a phase recovery which makes a rotation in the CCW direction But the value of the imaginary part is also positive So we multiply by the real part or the sign of the real part to perform sign correction Finally figure 3 13 shows a complete block diagram of costas loop 32 Rec_signal Error Real Imaginary LE e7j phase_error Figure 3 13 Block diagram of costas loop Control loop Phase error Phase error recovery signal 5 Differential decoder It is similar to the differential encoder as mentioned before in transmission path But here it makes differential decoding based on phase change of symbols Differential decoding is done by multiplying each sample of the input by the conjugate of the previous sample so that we can get the phase difference which is actually the data This can be illustrated through the following code segment for int i 0 i lt noutput_items i out i in i conj in i 1 The implementation of this block exists in the file diff_pha
90. to lock multiple systems together for MIMO Compatibility with all the same daughter boards as the original USRP Configuration flash SD card Figure 1 6 USRP 2 and its motherboard 13 And the following table provides comparison between USRP1 USRP2 and USRP N210 USRP 1 USRP 2 USRP N210 Interface USB 2 0 Gigabit Ethernet Gigabit Ethernet FPGA Altera EP1C12 Xilinx Spartan 3 Xilinx Spartan 2000 3A DSP 3400 8 MHz 16 bits 25 MHz E 16 bits 25 MHz 16 bits RF Bandwidth to from host Cost 700 1400 1400 ADC Samples 12 bit 64 MS s 14 bit 100 MS s 14 bit 100 MS s DAC Samples 14 bit 128 MS s 16 bit 400 MS s 16 bit 400 MS s Daughterboard capacity 2 TX 2 RX 1 TX 1 RX 1 TX 1 RX SRAM None 1 Megabyte 1 Megabyte Power 6V 3A 6V 3A 6V 3A Table 1 1 Comparison between USRP1 USRP2 and USRP N210 As our project is implemented on USRP2 and USRP N210 figure shows the internal construction and blocks of USRP2 or USRP N210 as there is no big differences except some improvements in USRP N210 over USRP 2 such as more capable FPGA and reprogrammable over Ethernet instead of SD card in USRP 2 FPGA Xilinx Spartan 3A DSP DDC Decim Ethernet PHY e d DUC UHD Inte Network Driver p pais wer Command amp Control Dac Data Streaming ADC DAC Clk MIMO Expansion 32 bit RISC GPIO SPI L rocessor P TEXO ADC DAC Cik
91. to the most important standards for wireless communications in this new generation called 4G in which the main standards are the IEEE 802 16 Worldwide Interoperability for Microwave Access WiMAX and the 3GPP s Long Term Evolution LTE 1 2 Tasks and contents For this project communication systems will be implemented These communications systems will include Analog environment example s digital modulation techniques such DBPSK and a physical layer example based on the OFDM multiplexing method thus being comparable to actual technologies such as WiMAX or LTE The first part of this report is the theoretical part in which the technologies that will be 1 used during the implementation are explained We will try to project the role that these technologies will play in the implementation throughout the theoretical explanation also we intended to introduce the concepts and ideas behind the chosen software toolkit for the implementation GNU Radio We will explain the structure of GNU Radio with the objective of giving the reader a general idea of the possibilities of GNU Radio This part is included in chapter 1 This chapter can be specially useful for people interested in starting a SDR project that are considering GNU Radio as their platform In chapter 2 we go deeply inside communication systems the best and easiest start is dealing with analog environment So in this phase we exhibited analog simple educational example to make th
92. to update with it Step 5 f f fx0 update frequency error Step 6 00 Ok f Og X H update phase error due to frequency and phase error Steps 4 5 and 6 are done after finishing one complete OFDM symbol by passing on all its subcarriers this means that updating PLL coefficients is done at each iteration subcarrier But the values are assigned in fact at each OFDM symbol and hence the corrections is done on OFDM symbols domain as an average errors and corrections and not on subcarriers domain 4 7 A point by point simulation Actually this sub section is not only very educational for beginners but also it completes the whole picture for a professional communication engineer Simulations are done using python programming language And the figures are plotted using pylab module Also these simulations are done without any USRP boards used only simulation through GRC to clarify and verify the function of each block The practical system using USRPs will come later stay tuned 4 7 1 Mapper e Aim Showing an example for header and scrambled data of OFDM symbols Simulation is done based on input 0 0 0 0 0 0 with length 512 occupied tones 200 and FFT length 512 The input is chosen to be all zeros because the scrambled data are added to the input data xor for each bit So the output will be the scramble data because it is already added to zeros e The following data is representing the first
93. ually suffices for pre emphasis and de emphasis 3 Audio Filter Maybe you are wondering where we pick out the station of interest from the digitized frequency band Actually this is done explicitly by the channel filter in wfm_rcy py and implicitly by the digital down converter DDC on the USRP Recall that DDC can be regarded as a low pass FIR filter followed by a down sampler As a result of these two operations the target station is picked out then spread out in the digital spectrum after decimation Because we choose an appropriate decimation rate and channel filter bandwidth we have isolated our station of interest To keep our life simple we just design a mono receiver using a low pass filter to select only channel signal 2 3 3 Band pass filter We used it to filter the received signal into our band only and hence correction reception is guaranteed The parameters inside this block can be controlled manually such as low cut off frequency of the filter high cut off frequency of the filter and transition BW 2 3 4 Rational resampler To reach the sample rate of audio sink in used PC we used this block to do this task thus we can enter the audio sink with adequate sample rate 2 3 5 Audio sink This block is responsible for listening to frequency of the sent tone Its sample rate differs from one PC to another that s why we used a rational resampler block before it 2 3 6 FFT sink This block simply display
94. ulator in the python file we can see that it has no direct input and it has one output which is the complex modulated signal at baseband The modulator includes a send function that takes as the parameter the payload of size 512 byte which will be sent from upper level thread a queue of messages each message contains a payload from data stream Then the Send PKT function sends these data to the first module of the modulator ofdm_mapper This payload will be converted into a data packet by calling a function make _packet from the python module ofdm packet_utils This function will calculate and add CRC as well as header addition both to the payload 38 4 5 2 Packet structure and construction As we said in previous sub section the word packet Now we will discuss it in details its structure as well as construction process of it Basically the packet structure as shown in figure 4 5 contains three fields are header payload 512 byte and CRC32 The Header Payload header contains two equal length fields each one of 2 byte that comprises e a offset and payload with CRC length Fa Ka y The two equal length fields are used for d W identifying the packet at the receiver Whitened scrambled Moreover the payload with CRC fields coded payload oded payload are whitened scrambled which Length Log d sem Offset involves bit wise XOR operation with gt q known pseudo random PN sequen
95. vector_sink_b 1 self gr_vector_sink_x_0 gr vector_sink_f 1 self gr_throttle_O gr throttle gr sizeof_float 1 samp_rate self digital_ofdm_sync_pn_0 digital ofdm_sync_pn 512 128 True self digital_ofdm_mod_0 grc_blks2 packet_mod_b digital ofdm_mod options grc_blks2 options modulation bpsk fft_length 512 occupied_tones 200 cp_length 128 pad_for_usrp False log None verbose None payload_length 0 THHHHRHH RRHH RRHH RRHH RH E E H a Connections 71 rer RARE self connect self digital_ofdm_mod_0 0 self digital_ofdm_sync_pn_0 0 self connect self gr_throttle_0 0 self gr_vector_sink_x_0 0 self connect self digital_ofdm_sync_pn_0 0 self gr_throttle_0 0 self connect self digital_ofdm_sync_pn_0 1 self gr_vector_sink_x_1 0 self connect self random_source_x_0 0 self gr_vector_sink_x_2 0 self connect self random_source_x_0 0 self digital_ofdm_mod_0 0 def get_samp_rate self return self samp_rate def set_samp_rate self samp_rate self samp_rate samp_rate A if name main parser OptionParser option_class eng_option usage prog options options args parser parse_args tb top_block tb Run True Second code Reading data and plotting them from gnuradio import digital import pylab from gnuradio import eng_notation from gnuradio import gr from gnuradio eng_option import eng
96. ved signal coming through the GE link from the USRP 2 motherboard This signal is a complex digitalized signal with a sample rate of 640 kHz down converted to baseband 20 2 3 2 Wide band FM receiver It s defined in a python file wfm_rcv py It has one input down converted base band signal and one output the demodulated audio It has many blocks implemented in C demodulating block de emphasizer and audio filter 1 Demodulating The first block in the wide band FM receiver chain is fm_demod an instance of gr quadrature_demod_cf To understand the work done within it we should know about how FM signals are generated With FM the instantaneous frequency of the transmitted waveform is varied as a function of the input signal The instantaneous frequency at any time is given by f t k m t fc Where m t is the input signal k is a constant that controls the frequency sensitivity and fc is the the carrier frequency To recover m t two steps are needed First we need to remove the carrier fc which leaves us with a baseband signal that has an instantaneous frequency proportional to the original message m t The second step is to compute the instantaneous frequency of the baseband signal Thus our challenge is to find a way to remove the carrier and compute the instantaneous frequency Removing the carrier has been done on the FPGA via the digital down converter DDC Thus the signal coming into the guts has already become a ba
97. well as RF module dedicated to the modulation to the desired transmission or receiving frequency These elements and a PC with enough computational power to run the software are the only hardware that we need If we need to speak in numbers we should differentiate if the radio will be used for narrow band applications or for wide band For narrow band applications a regular Pentium PC should have more than enough capacity to meet the requirements If however the application that we want to implement uses up a bigger part of the spectrum for example a receiver of multiple FM channels at the same time the requirements get much bigger and we might need powerful PCs in order to process all the data that we are using in the required time In our application the aim is implementing a communications system based on OFDM which is also a wide band application Not all aspects of software radio are positive and there is no free lunch There are also a number of challenges that have been reducing the use of SDR to a limited number of applications The first of them all is the power consumption of a SDR device As we have seen in the hardware requirements we will in many cases need powerful computers to run an application This means that we will need an amount of power that would never be achievable by a handheld device The power consumption of SDR devices is not comparable to the power needed by the radios that nowadays work in hardware Another important draw

Implementation of a wireless OFDM system using USRP 2

Contents

Download Pdf Manuals

Related Search

Related Contents