% File: multipath_sim.m
% Matlab Simulink IEEE 802.15.3a compliant channel model
% Author: Tim Becker (MSEE Degree, December 2004)
% Advisor: Robert Morelos-Zaragoza. San Jose State University
function [sys,x0,str,ts] = multipath_sim(t,x,u,flag,T,nT,LAMBDA,lambda,GAMMA,gamma,sigma1_dB,sigma2_dB,std_shadow,LOSflag)
% Differs from multipath_sim2.m in that h begins at t=0 for NLOS channels.
v = 10000;
switch flag,
% Initialization
case 0,
[sys,x0,str,ts]=mdlInitializeSizes(T,nT,v);
% Outputs
case 3,
sys=mdlOutputs(t,x,u,T,LAMBDA,lambda,GAMMA,gamma,sigma1_dB,sigma2_dB,std_shadow,LOSflag,v);
% Unhandled flags
case {1, 2, 4, 9}
sys = [];
% Unexpected flags
otherwise
error(['Unhandled flag = ',num2str(flag)]);
end
% end multipath_sim
%=============================================================================
% mdlInitializeSizes
% Return the sizes, initial conditions, and sample times for the S-function.
%=============================================================================
function [sys,x0,str,ts]=mdlInitializeSizes(T,nT,v)
sizes = simsizes;
sizes.NumContStates = 0;
sizes.NumDiscStates = 0;
sizes.NumOutputs = v;
sizes.NumInputs = 0;
sizes.DirFeedthrough = 0;
sizes.NumSampleTimes = 1; % at least one sample time is needed
sys = simsizes(sizes);
% initialize the initial conditions
x0 = [];
% str is always an empty matrix
str = [];
% initialize the array of sample times
ts = [T*nT*1e-9 0];
% end mdlInitializeSizes
%=============================================================================
% mdlOutputs
% Return the block outputs.
%=============================================================================
function sys=mdlOutputs(t,x,u,T,LAMBDA,lambda,GAMMA,gamma,sigma1_dB,sigma2_dB,std_shadow,LOSflag,v)
% returns the real impulse response of the
% indoor multipath channel as modeled by Intel in
% IEEE 802.15-02/279r0-SG3a and is based on the
% Saleh-Valenzuela model with lognormal fading.
% T is the time resolution (nsec).
sigma1 = 10^(sigma1_dB/20);
sigma2 = 10^(sigma2_dB/20);
Omega0 = 1;
if (strcmp(LOSflag,'on'))
T1 = [0];
else
T1 = [randn^2/(2*LAMBDA)+randn^2/(2*LAMBDA)];
end
while T1(end) <= 10*GAMMA
dT = randn^2/(2*LAMBDA)+randn^2/(2*LAMBDA);
T1 = [T1 T1(end)+dT];
end
for n = 1:length(T1)
tau(n) = {[T1(n)]};
while tau{n}(end) <= tau{n}(1)+10*gamma
dt = randn^2/(2*lambda)+randn^2/(2*lambda);
tau(n) = {[tau{n} tau{n}(end)+dt]};
end
mu(n) = {(10*log(Omega0)-10*T1(n)/GAMMA-10*(tau{n}-T1(n))/gamma)/log(10)-(sigma1^2+sigma2^2)*log(10)/20};
beta(n) = {10.^((sigma1*randn(size(mu{n}))+sigma2*randn(size(mu{n}))+mu{n})/20)};
alpha(n) = {beta{n}.*sign(rand(size(beta{n}))-0.5)}; % Intel version, +/-1
%alpha(n) = {beta{n}.*exp(-j*2*pi*rand(size(beta{n})))}; % cell array of channel coefficients
end
maxtime = max(tau{1}); % find maximum arrival time index in cell array tau
for n = 2:length(T1)
if max(tau{n}) > maxtime
maxtime = max(tau{n});
end
end
N = 32; % oversampling factor
Nfs = N/T;
h = zeros(1,floor(maxtime*Nfs)+1); % initialize impulse response vector
for n = 1:length(T1)
tauN{n} = floor(tau{n}*Nfs); % quantized time indices
end
for n = 1:length(T1)
for m = 1:length(tau{n})
h(tauN{n}(m)+1-tauN{1}(1)) = h(tauN{n}(m)+1-tauN{1}(1))+alpha{n}(m); % only line different from multipath_sim2.m
end
end
h = N*resample(h,1,N);
maxtime = ceil((10*GAMMA+10*gamma)/T); % maximum arrival time for channel
h = h(1:maxtime); % concatenate h to maximum channel length
E = sum(h.*conj(h)); % compute total channel energy
h = h./sqrt(E); % normalize total energy to 1
fac = 10^(std_shadow*randn/20);
h = h.*fac;
if(length(h)<v)
h = [h zeros(1,v-length(h))];
elseif(length(h)>v)
h = h(1:v);
end
%length(h)
sys = real(h);
% end mdlOutputs
