ELEC350 Assignment 5

Size: px

Start display at page:

Download "ELEC350 Assignment 5"

Imogen Holly Powell
6 years ago
Views:

1 ELEC350 Assignment 5 Instructor: Prof. Peter F. Driessen Marker: Peng Lu You are given a sound file in.wav format containing a binary FSK signal with noise. You are asked to implement a receiver and identify the data bits and decode the text message, using ascii code transmitting least significant bit first. Each student is given a different sound file, available here. Choose the file with the last 2 digits of your student number. The message starts with an ascii space, and the first letter of the message is capitalized. If the text is short, the message may repeat. There may be some random appearing symbols at the. A test file is given here containing the message "This is a test message " repeated several times. Playing the soundfile and listening to the FSK will help your work. You can read the file into Matlab using the "wavread" command. Decode the message using an AM type receiver, e.g. a bandpass filter on one of the two tones, followed by an envelope detector. A better receiver will use two bandpass filters and two envelope detectors (one for each tone), and the outputs are compared. Include documented Matlab source code, and plotted receiver output with description/interpretation/comparison. The message will a readable text message. If you do not get valid ascii characters, then you have made an error, and will not get credit for correctly decoding the message.

2 Answer: Fig.1. Energy spectrum of signals at different stages Fig. 1 shows the energy spectrum of FSK modulated signal, bandpass filtered signals and signal after envelop detector. For the FSK modulated signals, two frequency bands exist. After passing through the bandpass filter, the two bands are separated. The function of envelop detector is to remove carrier signal so that the modulated low frequency signal can be extracted.

Fig.2 waveforms of FSK modulated signal and bandpass filtered signal Fig. 2 shows the waveforms of FSK modulated signal and bandpass filtered signals.

3 Fig.2 waveforms of FSK modulated signal and bandpass filtered signal Fig. 2 shows the waveforms of FSK modulated signal and bandpass filtered signals. The bandpass filter divides the modulated signal into high frequency band and low frequency band. As there exists only one frequency band at the same time in the modulated signal, high and low frequency band signals are also separated in the time domain after passing through the bandpass filter.

4 Fig. 3 envelope detector output Fig. 3 shows the waveforms of envelope detector outputs. As there is only one frequency band at the same time, the high and low frequency band signals are inverse after passing through the envelope detector. The final decision is made upon the differences between them.

5 Appix of Source Code (take the file fskaudio1.wav as an example) % This program is to decode FSK audio signal. The only parameter that need to be changed % according to specific situation is symbol_sample, which is the number % of samples contained in a bit. This value can be obtained by observing % the output signal from the envelope detector. % Input: FSK modulated audio file filename.wav % Output: ASCii string of original message % Author: Peng Lu, maker for Elec350, fall term, 2006 clear all; close all; % Read data from audio file % [data,fs,num_bits]=wavread('test.wav'); [data,fs,num_bits]=wavread('fskaudio1.wav'); % energy spectrum density of the signal spectrum=abs(fft(data)).^2; interval=fs/(length(data)-1); % frequency sampling interval bandwidth=60; % assumed signal bandwidth used for central frequence estimation num_sample=round(bandwidth/interval); % number of samples contained in assumed bandwidth % obtain center frequency of high and low freq. bands spect_temp=spectrum; pos_half=floor(length(spect_temp)/2); [val_1 pos_1]=max(spect_temp(1:pos_half)); spect_temp(pos_1-round(num_sample/2):pos_1+round(num_sample/2))=0; % wipe out significant freq. components [val_2 pos_2]=max(spect_temp(1:pos_half)); if(pos_1>pos_2) freq_usb=(pos_1-1)*interval; freq_lsb=(pos_2-1)*interval; else freq_usb=(pos_2-1)*interval; freq_lsb=(pos_1-1)*interval; bandwidth=freq_usb-freq_lsb; % real bandwidth of signal %design BPF [USB_num,USB_den]=butter(5,[freq_USB-bandwidth/2 freq_usb+bandwidth/2]/(fs/2)); [LSB_num,LSB_den]=butter(5,[freq_LSB-bandwidth/2 freq_lsb+bandwidth/2]/(fs/2)); % pass audio signal through the BPF data_usb=filter(usb_num,usb_den,data); data_lsb=filter(lsb_num,lsb_den,data); interval=fs/(length(data_usb)-1); % frequency sampling interval

6 % design an envelope detector, which is a concatenation of rectifier % and low pass filter % There are two types of rectifier: half wave rectifier and full wave rectifier % In half wave rectification, either the positive or negative half of the AC wave is % passed while the other half is blocked. % A full wave rectifier converts the whole of the input waveform to one % of constant polarity (positive or negative) at its output by reversing % the negative (or positive) portions to the alternating current waveform. USB_temp=data_USB; LSB_temp=data_LSB; USB_temp=abs(USB_temp); % full wave rectification of high freq. band % USB_temp(USB_temp<0)=0; % half wave rectification of high freq. band LSB_temp=abs(LSB_temp); % full wave rectification of low freq. band % LSB_temp(LSB_temp<0)=0; % half wave rectification of low freq. band [num,den]=butter(10,bandwidth*2/(fs/2),'low'); % low pass filter data_usb_filter=filter(num,den,usb_temp); data_lsb_filter=filter(num,den,lsb_temp); diff=data_usb_filter-data_lsb_filter; x_ranges=[0:interval:fs]; figure subplot(4,1,1); plot(x_ranges,spectrum); xlabel('frequency(hz)'); title('energy spectrum of FSK modulated signal'); subplot(4,1,2); plot(x_ranges,abs(fft(data_usb)).^2,'r'); xlabel('frequency(hz)'); title('energy spectrum of high frequency band'); subplot(4,1,3); plot(x_ranges,abs(fft(data_lsb)).^2,'g') xlabel('frequency(hz)'); title('energy spectrum of low frequency band'); subplot(4,1,4); plot(abs(fft(diff).^2)); xlabel('frequency(hz)'); title('energy spectrum of envelope detector output'); ranges=[2e4:2e4+4e3]; x_ranges=ranges/fs; figure subplot(3,1,1); plot(x_ranges,data(ranges))

7 title('waveforme of FSK modulated signal'); subplot(3,1,2); plot(x_ranges,data_usb(ranges),'r') title('waveforme of high frequency signal'); subplot(3,1,3); plot(x_ranges,data_lsb(ranges),'g') xlabel('time(s)'); title('waveforme of low frequency signal'); ranges=[2e4:2e4+4e3]; x_ranges=ranges/fs; figure subplot(3,1,1); plot(x_ranges,data_usb_filter(ranges)) title('envelope detector output (high frequency band)'); subplot(3,1,2); plot(x_ranges,data_lsb_filter(ranges),'r') title('envelope detector output (low frequency band)'); subplot(3,1,3); plot(x_ranges, diff(ranges)); xlabel('time(s)'); title('difference between high and low frequency band output'); symbol_sample=260; % number of samples in a bit % chip synchronization [val,pos_max]=max(abs(diff(1:symbol_sample*10))); for shift=1:symbol_sample energy(shift)=sum(abs(diff(pos_max+shift-1:symbol_sample:pos_max+shift-1+10*symbol_sampl e))); [val,shift_max]=max(energy); chip_sync=pos_max+shift_max-1; % a sample that have the maximum power chip_sync=mod(chip_sync,symbol_sample); % search within symbol_sample bit_seq=diff(chip_sync:symbol_sample:length(diff)); % sample signal to obtain bit sequence bit_seq=sign(bit_seq); % make decision with threshold equal to 0 %symbol synchronization pilot=[ ]'; % pilot symbol: space for k=1:length(bit_seq)-15 symbol_corr(k)=sum(bit_seq(k:k+length(pilot)-1).*pilot);

8 k=1; while (symbol_corr(k)<8) k=k+1; symbol_sync=k; % find the first ASCii code for space % record the position where symbols start str_index=1; temp=zeros(1,8); while k<length(bit_seq)-7 binary=bit_seq(k:k+7); % get a symbol binary=binary>0; % convert it to 0 and 1 sequence dec_val=binary(8)*2^7+binary(7)*2^6+binary(6)*2^5+binary(5)*2^4+binary(4)*2^3+... binary(3)*2^2+binary(2)*2^1+binary(1); % convert it to decimate msg_string(str_index)=char(dec_val); % convert it to a character k=k+8; str_index=str_index+1; % next symbol fprintf('message=%s%n',msg_string)

Laboratory 5: Spread Spectrum Communications

Laboratory 5: Spread Spectrum Communications Cory J. Prust, Ph.D. Electrical Engineering and Computer Science Department Milwaukee School of Engineering Last Update: 19 September 2018 Contents 0 Laboratory