why do I get this error while using qammod? "Expected input number 1, X, to be an array with all of the values < 4."

N=256;
% Number of subcarriers or size of IFFT/FFT
N_data_symbol=128;
% Number of symbol to IFFT
GI=N/4;
% Guard interval 1/4,1/8,1/16,. . .
M=4;
% Modulation 2:BPSK, 4:QPSK, (8,16,32,64,128,256):QAM
L=16;
% Channel length
N_Iteration=500;
% Number of iteration
SNR=[0:1:15];
% Signal to Noise Ratio in dB
for i=1:length(SNR)
    snr=SNR(i)
    for (k=1:N_Iteration)
        tx_bits=randi(N_data_symbol,1,M);
        % Input Bit Streams to Transmit
        % Modulation
        tx_bits_Mod=qammod(tx_bits,M);
        input_symbol=[zeros((N-N_data_symbol)/2,1); 
            tx_bits_Mod;
            zeros((N-N_data_symbol)/2,1)];
        
        % IFFT and Circular Prefix Addition
        ofdm_symbol_ifft=ifft(input_symbol,N);
        % Guard Interval insertion (CP)
        guard_symbol=ofdm_symbol_ifft(N-GI+1:N);
        % Add the cyclic prefix to the ofdm symbol
        ofdm_symbol=[guard_symbol; ofdm_symbol_ifft];
        %sig_pow=ofdm_symbol'.*conj(ofdm_symbol);
        ofdm_spectrum=ofdm_symbol;
        %ofdm_spectrum=[ofdm_spectrum; ofdm_symbol];
        % h=randn(L,1)+j*randn(L,1); % Generate random channel
        h(1:L,1)
        %h=h./sum(abs(h)); % Normalization
        h=1; %AWGN
        y1=filter(h,1,ofdm_symbol);
        %y=x*h
        % Adding AWGN Noise
        y=awgn(y1,snr,'measured');
        %y=x*h+n
        % Remove Cyclic prefix
        rx_symbol=y(GI+1:N+GI);
        % The FFT of the time domain signal after the removal of
        cyclic prefix
        rx_symbol_fft=fft(rx_symbol,N);
        % Equalization
        %H_f=fft(h,N);
        %G=1./H_f;
        rx_equalized_zp=rx_symbol_fft;
        rx_equalized=rx_equalized_zp(((N-N_data_symbol)/2)+1:(N+N_data_symbol)/2);
        % Demodulate
        rx_bits_zp=qamdemod(rx_equalized,M);
        rx_bits=rx_bits_zp;
        % Comparison
        % Bit Error Rate computation
        [nErr bErr(i,k)]=symerr(tx_bits,rx_bits);
    end
end

snr_theo=10.^(SNR/10);
Theory_awgn=0.5*erfc(sqrt(snr_theo));
semilogy(SNR,mean(bErr'),'b',SNR,Theory_awgn,'ro–');

The errors I'm getting are as follows:

Error using qammod

Expected input number 1, X, to be an array with all of the values < 4.

Error in qammod>validateInput (line 288)

validateattributes(x, {'numeric'}, {'real','integer','>=',0,'<',M}, mfilename, 'X', 1);

Error in qammod (line 95)

validateInput(x, M, bitInput, outputDataType);

Error in Experimental_DCOFDM_2 (line 21)

tx_bits_Mod=qammod(tx_bits,M);

Answer 1

[Chunru]

A few bugs in program. The major one is generating input signa which should bel: tx_bits=randi([0 M-1], [N_data_symbol,1]);

N=256;
% Number of subcarriers or size of IFFT/FFT
N_data_symbol=128;
% Number of symbol to IFFT
GI=N/4;
% Guard interval 1/4,1/8,1/16,. . .
M=4;
% Modulation 2:BPSK, 4:QPSK, (8,16,32,64,128,256):QAM
L=16;
% Channel length
N_Iteration=500;
% Number of iteration
SNR=[0:1:15];
% Signal to Noise Ratio in dB
for i=1:length(SNR)
    snr=SNR(i)
    for (k=1:N_Iteration)
        %tx_bits=randi(N_data_symbol,1,M);
        tx_bits=randi([0 M-1], [N_data_symbol,1]);
        % Input Bit Streams to Transmit
        % Modulation
        tx_bits_Mod=qammod(tx_bits,M);
        input_symbol=[zeros((N-N_data_symbol)/2,1); 
            tx_bits_Mod;
            zeros((N-N_data_symbol)/2,1)];
        
        % IFFT and Circular Prefix Addition
        ofdm_symbol_ifft=ifft(input_symbol,N);
        % Guard Interval insertion (CP)
        guard_symbol=ofdm_symbol_ifft(N-GI+1:N);
        % Add the cyclic prefix to the ofdm symbol
        ofdm_symbol=[guard_symbol; ofdm_symbol_ifft];
        %sig_pow=ofdm_symbol'.*conj(ofdm_symbol);
        ofdm_spectrum=ofdm_symbol;
        %ofdm_spectrum=[ofdm_spectrum; ofdm_symbol];
        % h=randn(L,1)+j*randn(L,1); % Generate random channel
        %h(1:L,1)
        %h=h./sum(abs(h)); % Normalization
        h=1; %AWGN
        y1=filter(h,1,ofdm_symbol);
        %y=x*h
        % Adding AWGN Noise
        y=awgn(y1,snr,'measured');
        %y=x*h+n
        % Remove Cyclic prefix
        rx_symbol=y(GI+1:N+GI);
        % The FFT of the time domain signal after the removal of
        % cyclic prefix
        rx_symbol_fft=fft(rx_symbol,N);
        % Equalization
        %H_f=fft(h,N);
        %G=1./H_f;
        rx_equalized_zp=rx_symbol_fft;
        rx_equalized=rx_equalized_zp(((N-N_data_symbol)/2)+1:(N+N_data_symbol)/2);
        % Demodulate
        rx_bits_zp=qamdemod(rx_equalized,M);
        rx_bits=rx_bits_zp;
        % Comparison
        % Bit Error Rate computation
        [nErr bErr(i,k)]=symerr(tx_bits,rx_bits);
    end
end

snr = 0

snr = 1

snr = 2

snr = 3

snr = 4

snr = 5

snr = 6

snr = 7

snr = 8

snr = 9

snr = 10

snr = 11

snr = 12

snr = 13

snr = 14

snr = 15

snr_theo=10.^(SNR/10);
Theory_awgn=0.5*erfc(sqrt(snr_theo));
semilogy(SNR,mean(bErr'),'b', SNR,Theory_awgn, 'ro-');

[Hanojhan Rajahrajasingh]

Thank you so much.

[Hanojhan Rajahrajasingh]

In the above code channel is set to 1 (h), if I try to uncomment the below line of code and generate a random channel then blue curve representing the simulated value becomes a straight line. What is causing this issue? Why can't the channel be used?

% h=randn(L,1)+j*randn(L,1); % Generate random channel