estChanResp.m

function [hfd,h,snr] = estChanResp(r,xfd,opt)
% estChanResp:  Estimates the channel response using frequency-domain
% correlation
% 
% The model is that the TX repeatedly sends x=ifft(xfd) and the RX 
% receives samples
%     r = h*x + w
% where h is the channel impulse response and w is noise.  
%
% Parameters
% ----------
% r:  RX complex baseband samples of length nfft
% xfd:  TX samples in frequency-domain
% nleft, nright:  Number of samples to the left and right of peak
%     used for signal energy computation
% normToNoise:  Boolean flag indicating that the energy / tap
%     should be normalized to the noise energy estimate.
%     In this case abs(h(k))^2 represents the SNR per tap.
%
% Returns
% -------
% hfd:   Frequency domain channel estimate.  hfd(k) is the estimate
%    of the channel frequency response at f(k) = k/nfft*fsamp
% h:  Time-domain channel impulse response estimate.
% snr:  Total snr in dB.
arguments
    r (:,1) double;
    xfd (:,1) double;
    
    opt.nleft (1,1) {mustBeInteger} = 8; 
    opt.nright (1,1) {mustBeInteger} = 8;
    opt.normToNoise (1,1) = false;

end

% Create empty outputs until they are set.
% You can delete this code when you have set the variables
hfd = [];
h = [];
snr = [];

% TODO:  Take the FFT of r 
%    rfd = ...

% TODO:  Estimate the channel frequency response by dividing by the 
% FFT of x, xfd
%    hfd = ...

% TODO:  Compute the time-domain response
%    h = ...


% TODO:  Find the peak location in h
% Then, circularly shift h so that the peak is at position opt.nleft
%    h = circshift(...)

% TODO:  Compute the SNR
%   Enoise = ...
%   Esig = ...
%   snr = ... (in dB)



% TODO:  Normalize to the noise level
%    h = ...
if opt.normToNoise

end

end