Reference no: EM132711495
ELEC86303 Wireless Communication Systems
Project - Adaptive Channel Equalization
Based on the equalizer system in Figs. 1 & 2, consider the following:
1) Consider the input as bipolar sequence given by: a = (randn(ni, 1)>0) *2-1; ni=10000; where ni is the length of the input sequence.
2) Consider a simple channel given as: h= [ 0.05 -0.063 0.088 -0.126 -0.25 0.9047 0.25 0 0.126 0.038 0.088];
Part I: Training-Mode Adaptive Equalizer
1. Consider a length 11 adaptive filter in Fig. 1 and initiate the taps to zeros.
2. Consider an additive white Gaussian noise with variance 0.001 as: a. v= sqrt(0.001)*randn(ni,1);
3. Consider the delay that would result from the adaptive filter delay.
4. Use LMS algorithm with proper step size μ.
5. Plot the learning curve (the error in dB vs. the sample number).
6. Provide a simple decision device to be applied to the filter output to provide a clear signal that should resemble the transmitted signal. Compare the transmitted signal to the signal after the decision device. To compare, you need to observe the transmitted signal but those samples at the end equal to the delay of the filter, and compare that to the signal after the decision devices starting from the sample equal to the delay plus one to the length of the other signal. Compute the Bit-Error-Rate (BER) = error count/total number of samples encountered.
7. Provide plots of the impulse response of the channel, the received signal, u, and the signal the output of the filter y.
8. Select a different value of μ (preferably of less value) and repeat the above procedure and comment on the MSE and the convergence rate.
Part II: Blind-Mode Adaptive Equalizer
1. Consider a length 11 adaptive filter in Fig. 2 and initiate the taps to zeros but the one at the middle (close to the middle) as 0.5.
2. Consider an additive white Gaussian noise with SNR = 35 dB as follows:
snrdb = 35; % SNR (in dB)
SNR = 10.^(snrdb/10); v = randn(length(ah),1); v = v*sqrt(1/SNR)*(sqrt(0.5));
3. Use CMA algorithm (so you need to calculate the dispersion factor γ2).
4. Plot the learning curve (the error in dB vs. the sample number).
5. Provide a simple decision device to be applied to the filter output to provide a clear signal just like in point 4, Part I, and follow the same steps to provide s measure of the BER.
6. Provide plots of the impulse response of the channel, the received signal, u, and the signal the output of the filter y.
7. Provide plots of the impulse response of the channel, the received signal, u, and the signal the output of the filter y.
8. Select a different value of m (preferably of less value) and repeat the above procedure and comment on the MSE and the convergence rate.
Write a report describing your methodology and results with your comments and Matlab code.