MATLAB实现限幅和滤波的PAPR分布和BER.zip

% CCDF_of_clipped_filtered_OFDM_signal.m % Plot Fig. 7.16 clear, clf SNRdBs=[0:10]; N_SNR=length(SNRdBs); % SNR[dB] vector Nblk=100; CRs=[0.8:0.2:1.6]; N_CR=length(CRs); gss='*^<sd>v'; b = 2; M = 2^b; % Number of bits per QAM symbol and Alphabet size N = 128; Ncp = 0; % FFT size and CP size (GI length) fs = 1e6; L = 8; % Sampling frequency and Oversampling factor Tsym=1/(fs/N); Ts=1/(fs*L); % OFDM symbol period and Sampling period fc = 2e6; wc = 2*pi*fc; % Carrier frequency t = [0:Ts:2*Tsym-Ts]/Tsym; if b==1, Mod='BPSK'; elseif b==2, Mod='QPSK'; else Mod=[num2str(M) 'QAM']; end % const = qammod([0:M-1],M); % QAM modulation A = modnorm(qammod([0:M-1],M),'avpow',1); % Normalization factor mdmod = modem.qammod('M',M, 'SymbolOrder','Gray','InputType','Bit'); mddem = modem.qamdemod('M',M, 'SymbolOrder','Gray','OutputType','Bit'); %Hd = euqiripple_filter_fin; Fcoeff = Hd.numerator; Lcoeff = length(Fcoeff); Fs=8; Norder=104; dens=20; % Sampling frequency, Order, and Density factor of filter FF=[0 1.4 1.5 2.5 2.6 Fs/2]; % Stopband/Passband/Stopband frequency edge vector WW=[10 1 10]; % Stopband/Passband/Stopband weight vector h = firpm(Norder,FF/(Fs/2),[0 0 1 1 0 0],WW,{dens}); % BPF coefficients Clipped_errCnt = zeros(size(CRs)); ClippedFiltered_errCnt = zeros(size(CRs)); CF = zeros(1,Nblk); CF_c = zeros(N_CR,Nblk); CF_cf = zeros(N_CR,Nblk); % ------------- Iteration with increasing SNRdB --------------------% ber_analytic = berawgn(SNRdBs-10*log10(b),'qam',M); kk1=1:(N/2-Ncp)*L; kk2=kk1(end)+1:N/2*L+N*L; kk3=kk2(end)+[1:N*L/2]; z = [2:0.1:16]; len_z = length(z); for i = 1:N_SNR SNRdB = SNRdBs(i); for ncf = 0:2 % no/clip/clip&filter if ncf==2, m=ceil(length(h)/2); else m=1; end for cr = 1:N_CR if ncf==0&cr>1, break; end CR = CRs(cr); nobe = 0; for nblk = 1:Nblk %(i) msgbin = randint(b,N); % binary squences X = A*modulate(mdmod,msgbin); % 4QAM (QPSK) mapper X(1) = 0+j*0; % DC subcarrier not used x = IFFT_oversampling(X,N,L); x_b = add_CP(x,Ncp*L); x_b_os = [zeros(1,(N/2-Ncp)*L), x_b, zeros(1,N*L/2)]; x_p = sqrt(2)*real(x_b_os.*exp(j*2*wc*t)); if ncf>0, x_p_c = clipping(x_p,CR); x_p=x_p_c; % clipping if ncf>1, x_p_cf = ifft(fft(h,length(x_p)).*fft(x_p)); x_p=x_p_cf; end end if i==N_SNR, CF(nblk) = PAPR(x_p); end y_p_n = [x_p(kk1) awgn(x_p(kk2),SNRdB,'measured') x_p(kk3)]; % add Noise(AWGN) y_b = sqrt(2)*y_p_n.*exp(-j*2*wc*t); Y_b = fft(y_b); y_b_z = ifft(zero_pasting(Y_b)); y_b_t = y_b_z((N/2-Ncp)*L+m+[0:L:(N+Ncp)*L-1]); Y_b_f = fft(y_b_t(Ncp+1:end),N)*L; Y_b_bin = demodulate(mddem,Y_b_f); nobe = nobe + biterr(msgbin(:,2:end),Y_b_bin(:,2:end)); %Clipped_errCnt(cr) = Clipped_errCnt(cr) + nobe; end % End of the nblk loop if ncf==0, ber_no(i) = nobe/Nblk/(N-1)/b; elseif ncf==1, ber_c(cr,i) = nobe/Nblk/(N-1)/b; else ber_cf(cr,i) = nobe/Nblk/(N-1)/b; end if i==N_SNR for iz=1:len_z, CCDF(iz) = sum(CF>z(iz))/Nblk; end if ncf==0, CCDF_no = CCDF; break; elseif ncf==1, CCDF_c(cr,:) = CCDF; else CCDF_cf(cr,:) = CCDF; end end end end end subplot(221), semilogy(z,CCDF_no), grid on, hold on for cr = 1:N_CR gs = gss(cr); subplot(221), semilogy(z,CCDF_c(cr,:),[gs '-'], z,CCDF_cf(cr,:),[gs ':']), hold on subplot(222), semilogy(SNRdBs,ber_c(cr,:),[gs '-'], SNRdBs,ber_cf(cr,:),[gs ':']), hold on end semilogy(SNRdBs,ber_no,'o', SNRdBs,ber_analytic,'k'), grid on