HELP! How can i solve this array incompatible size issue

I was trying to combine two signals and caculate the bit error rate and plot the graph but it shows error at line 30 with Arrays have incompatible sizes for this operation. Can anyone help me please. I am new to matlab. Below is the code.

clc;
clear all;
close all;
% Parameters
fs = 1000;          % Sampling frequency
fc = 100;           % Carrier frequency for BPSK modulation
T = 1;              % Duration of the signals
t = 0:1/fs:T-1/fs;  % Time vector
% SNR for channel in dB
SNRdB = 0 : 32;
% Get SNR  Valur, SNRdB = 10*log(10) (signalpower/noise power)
SNR = 10.^(SNRdB/10);
%Data bits
N = 10^5;
x = randsrc(1,N,[0,1]);
%BPSK data generation
bpsk_modulated  = 1-2*x;
% Generate Interference Signal (e.g., sinusoidal interference)
interference_frequency = 50; % Frequency of the interference signal
interference_amplitude = 0.5; % Amplitude of the interference signal
interference_signal = interference_amplitude * sin(2*pi*interference_frequency*t);
% Combine BPSK Signal with Interference
size(bpsk_modulated)

ans = 1x2

1 100000

size(interference_signal)

ans = 1x2

1 1000

x_BPSK = bpsk_modulated + interference_signal;

Arrays have incompatible sizes for this operation.

%AWGN noise
n = randn(1,N);
for k=1:length(SNR);
    %Assume that noise follow the gaussian distribution, the (0,1). here
    %the veriance (sigma squre = 1 and mean = 0), so that noise power = 1
    y = (sqrt(SNR(k)) * x_BPSK) + n;
    % We need to find which bits has changed with noise ( -1 -> +1 , +1 ->
    % -1) in the transmitted channel
    noisy_bits = y.*x_BPSK;
    % Get indices of the currupted bits by noise
    index_currupted = find((noisy_bits)<0);
    
    % Get no of currupted bits by noise
    NumOfError_bits(k) = length(find(index_currupted));
end
%BER calculation
ber= NumOfError_bits/N;
%Simulation results
figure;
%Practical
prac = semilogy(SNRdB, ber, 'b*-','linewidth', 1);
hold on;
%Theoritical
theoritical = qfunc(sqrt(SNR));
theo = semilogy(SNRdB,theoritical,' r+-','linewidth',1);
xlabel("SNR in dB");
ylabel("Bit Error Rate(BER)");
legend ([prac theo], {'Practical','Theoritical'});
grid on;
datacursormode on;

[Fangjun Jiang]

see the output of size() prior to the error.

The size of the two matrix is different.

[WeiQuan]

Hello Fangjun, are you able to guide me on that? How can i change the two matrix to be in same size?

Answer 1

[Chunru]

Make the change: N = length(t);

If you need larger N, thant T in your code to a larger value.

% Parameters
fs = 1000;          % Sampling frequency
fc = 100;           % Carrier frequency for BPSK modulation
T = 1;              % Duration of the signals
t = 0:1/fs:T-1/fs;  % Time vector
% SNR for channel in dB
SNRdB = 0 : 32;
% Get SNR  Valur, SNRdB = 10*log(10) (signalpower/noise power)
SNR = 10.^(SNRdB/10);
%Data bits
%N = 10^5;
N = length(t);
x = randsrc(1,N,[0,1]);
%BPSK data generation
bpsk_modulated  = 1-2*x;
% Generate Interference Signal (e.g., sinusoidal interference)
interference_frequency = 50; % Frequency of the interference signal
interference_amplitude = 0.5; % Amplitude of the interference signal
interference_signal = interference_amplitude * sin(2*pi*interference_frequency*t);
% Combine BPSK Signal with Interference
%whos
x_BPSK = bpsk_modulated + interference_signal;
%AWGN noise
n = randn(1,N);
for k=1:length(SNR);
    %Assume that noise follow the gaussian distribution, the (0,1). here
    %the veriance (sigma squre = 1 and mean = 0), so that noise power = 1
    y = (sqrt(SNR(k)) * x_BPSK) + n;
    % We need to find which bits has changed with noise ( -1 -> +1 , +1 ->
    % -1) in the transmitted channel
    noisy_bits = y.*x_BPSK;
    % Get indices of the currupted bits by noise
    index_currupted = find((noisy_bits)<0);
    
    % Get no of currupted bits by noise
    NumOfError_bits(k) = length(find(index_currupted));
end
%BER calculation
ber= NumOfError_bits/N;
%Simulation results
figure;
%Practical
prac = semilogy(SNRdB, ber, 'b*-','linewidth', 1);
hold on;
%Theoritical
theoritical = qfunc(sqrt(SNR));
theo = semilogy(SNRdB,theoritical,' r+-','linewidth',1);
xlabel("SNR in dB");
ylabel("Bit Error Rate(BER)");
legend ([prac theo], {'Practical','Theoritical'});
grid on;
datacursormode on;

[WeiQuan]

Thanks Chunru!.. May i ask why dose my practical value wont decrease at all?

[Chunru]

You need a very large N in order to compute the small BER. For example, if you want to have a data point with BER of 1e-10, N shoud be in the order of 1e10. It is unrealistic to do simulation of this type with very small BER (eg. BER = 1e-20, N=O(1e20); Imagine how much memory it uses.).

Try to plot your BER Vs SNR in a reasonable range.

[WeiQuan]

Hey Chunru may i know which parameters should i change to have a reasonable range ?

[Chunru]

You can set N=1e9 if you have >8GB or RAM (adjust it according to your hardware). In this case, plot your BER curve up to 1e-9 (or so).