gusucode.com > wlan工具箱matlab源码程序 > wlan/wlan/wlanPacketDetect.m
function [startOffset,M] = wlanPacketDetect(x,chanBW,varargin)
%WLANPACKETDETECT OFDM packet detection using the L-STF
%
% STARTOFFSET = wlanPacketDetect(X,CHANBW) returns the offset from the
% start of the input waveform to the start of the detected preamble using
% auto-correlation. Only OFDM modulation is supported.
%
% STARTOFFSET is an integer scalar indicating the location of the start
% of a detected packet as the offset from the start of the matrix X. If
% no packet is detected an empty value is returned.
%
% X is the received time-domain signal. It is an Ns-by-Nr matrix of real
% or complex samples, where Ns represents the number of time domain
% samples and Nr represents the number of receive antennas.
%
% CHANBW is a string describing the channel bandwidth and must be 'CBW5',
% 'CBW10', 'CBW20' 'CBW40', 'CBW80' or 'CBW160'.
%
% STARTOFFSET = wlanPacketDetect(..., OFFSET) specifies the offset to
% begin the auto-correlation process from the start of the matrix X. The
% STARTOFFSET is relative to the input OFFSET when specified. It is an
% integer scalar greater than or equal to zero. When unspecified a value
% of 0 is used.
%
% STARTOFFSET = wlanPacketDetect(...,OFFSET,THRESHOLD) specifies the
% threshold which the decision statistic must meet or exceed to detect a
% packet. THRESHOLD is a real scalar greater than 0 and less than or
% equal to 1. When unspecified a value of 0.5 is used.
%
% [STARTOFFSET,M] = wlanPacketDetect(...) returns the decision statistics
% of the packet detection algorithm of matrix X. When THRESHOLD is set to
% 1, the decision statistics of the complete waveform will be returned
% and STARTOFFFSET will be empty.
%
% M is a real vector of size N-by-1, representing the decision statistics
% based on auto-correlation of the input waveform. The length of N
% depends on the starting location of the auto-correlation process till
% the successful detection of a packet.
%
% Example 1:
% % Detect a received 802.11n packet at 20dB SNR.
%
% cfgHT = wlanHTConfig; % Create packet configuration
% SNR = 20; % In decibels
% tgn = wlanTGnChannel('LargeScaleFadingEffect','None');
%
% % Generate transmit waveform
% txWaveform = wlanWaveformGenerator([1;0;0;1],cfgHT);
%
% fadedSig = tgn(txWaveform);
% rxWaveform = awgn(fadedSig,SNR,0);
%
% startOffset = wlanPacketDetect(rxWaveform,cfgHT.ChannelBandwidth);
%
% Example 2:
% % Detect a received 802.11a packet without channel impairments
%
% cfgNonHT = wlanNonHTConfig; % Create packet configuration
%
% % Generate transmit waveform
% txWaveform = wlanWaveformGenerator([1;0;0;1],cfgNonHT, ...
% 'WindowTransitionTime', 0);
%
% % Delay the signal by appending zeros at the start
% rxWaveform = [zeros(20,1);txWaveform];
%
% startOffset = wlanPacketDetect(rxWaveform, ...
% cfgNonHT.ChannelBandwidth,0,0.99);
% sprintf('Packet start offset: %d',startOffset)
%
% Example 3:
% % Return the decision statistics of the input waveform. The waveform
% % consists of five WLAN packets. The decision statistics shows five
% % peaks. Each peak corresponds to the detection of a single packet.
%
% cfgNonHT = wlanNonHTConfig;
% txWaveform = wlanWaveformGenerator([1;0;0;1],cfgNonHT, ...
% 'NumPackets',5,'IdleTime',20e-6);
% [~,M] = wlanPacketDetect(txWaveform,cfgNonHT.ChannelBandwidth,0,1);
% figure; plot(M);
%
% See also wlanCoarseCFOEstimate, wlanFieldIndices.
% Copyright 2016 The MathWorks, Inc.
%#codegen
narginchk(2,4);
nargoutchk(0,2);
% Check if M is requested
if nargout==2
M = [];
end
validateattributes(x,{'double'},{'2d','finite'},mfilename,'signal input');
startOffset = [];
if isempty(x)
return;
end
if nargin == 2
threshold = 0.5; % Default value
offset = 0; % Default value
elseif nargin == 3
threshold = 0.5;
validateattributes(varargin{1}, {'double'}, ...
{'integer','scalar','>=',0}, mfilename, 'offset');
offset = varargin{1};
else
validateattributes(varargin{1}, {'double'}, ...
{'integer','scalar','>=',0}, mfilename, 'offset');
validateattributes(varargin{2}, {'double'}, ...
{'real','scalar','>',0,'<=',1}, mfilename, 'threshold');
offset = varargin{1};
threshold = varargin{2};
end
% Validate Offset value
coder.internal.errorIf(offset>size(x,1)-1, ...
'wlan:wlanPacketDetect:InvalidOffsetValue');
% Validate Channel Bandwidth
coder.internal.errorIf(~any(strcmp(chanBW,{'CBW5','CBW10','CBW20', ...
'CBW40','CBW80','CBW160'})), ...
'wlan:wlanPacketDetect:InvalidChBandwidth');
Td = 0.8e-6; % Time period of a short training symbol for 20MHz
switch chanBW
case 'CBW40'
symbolLength = Td/(1/40e6);
case {'CBW80'}
symbolLength = Td/(1/80e6);
case {'CBW160'}
symbolLength = Td/(1/160e6);
otherwise % 'CBW5', 'CBW10', 'CBW20'
symbolLength = Td/(1/20e6); % In samples
end
lenLSTF = symbolLength*10; % Length of 10 L-STF symbols
lenHalfLSTF = lenLSTF/2; % Length of 5 L-STF symbols
inpLength = (size(x,1)-offset);
% Append zeros to make the input equal to multiple of L-STF/2
if inpLength<=lenHalfLSTF
numPadSamples = lenLSTF-inpLength;
else
numPadSamples = lenHalfLSTF*ceil(inpLength/lenHalfLSTF)-inpLength;
end
padSamples = zeros(numPadSamples,size(x,2));
% Process the input waveform in blocks of L-STF length. The processing
% blocks are offset by half the L-STF length.
numBlocks = (inpLength+numPadSamples)/lenHalfLSTF;
if nargout==2
% Define decision statistics vector
DS = coder.nullcopy(zeros(size(x,1)+length(padSamples)-offset-2*symbolLength+1,1));
if numBlocks > 2
for n=1:numBlocks-2
% Update buffer
buffer = x((n-1)*lenHalfLSTF+(1:lenLSTF)+offset,:);
[startOffset,out] = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
DS((n-1)*lenHalfLSTF+1:lenHalfLSTF*n,1) = out(1:lenHalfLSTF);
if ~(isempty(startOffset))
% Packet detected
startOffset = startOffset+(n-1)*lenHalfLSTF;
DS((n-1)*lenHalfLSTF+(1:length(out)),1) = out;
% Resize decision statistics
M = DS(1:(n-1)*lenHalfLSTF+length(out));
return;
end
end
% Process last block of data
blkOffset = lenHalfLSTF*(numBlocks-2);
buffer = [x(blkOffset+1+offset:end,:);padSamples];
[startOffset,out] = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
if ~(isempty(startOffset))
startOffset = startOffset+blkOffset; % Packet detected
end
DS(blkOffset+1:end,1)= out;
M = DS(1:end-length(padSamples));
else
buffer = [x(offset+1:end,:);padSamples];
[startOffset,out] = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
M = out;
end
else
if numBlocks > 2
for n=1:numBlocks-2
buffer = x((n-1)*lenHalfLSTF+(1:lenLSTF)+offset,:); % Update buffer
startOffset = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
if ~(isempty(startOffset))
startOffset = startOffset+(n-1)*lenHalfLSTF; % Packet detected
return;
end
end
% Process last block of data
blkOffset = lenHalfLSTF*(numBlocks-2);
buffer = [x(blkOffset+1+offset:end,:);padSamples];
startOffset = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
if ~(isempty(startOffset))
startOffset = startOffset+blkOffset; % Packet detected
end
else
buffer = [x(offset+1:end,:);padSamples];
startOffset = correlateSamples(buffer,symbolLength,lenLSTF,threshold);
end
end
end
function [packetStart,Mn] = correlateSamples(rxSig,symbolLength,lenLSTF,...
threshold)
% Estimate the start offset of the preamble of the receive WLAN packet,
% using auto-correlation method [1,2].
% [1] OFDM Wireless LANs: A Theoretical and Practical Guide 1st Edition
% by Juha Heiskala (Author),John Terry Ph.D. ISBN-13:978-0672321573
% [2] OFDM Baseband Receiver Design for Wireless Communications by
% Tzi-Dar Chiueh, Pei-Yun Tsai. ISBN: 978-0-470-82234-0
correlationLength = lenLSTF-(symbolLength*2);
pNoise = eps; % Adding noise to avoid the divide by zero
weights = ones(symbolLength,1);
index = 1:correlationLength+1;
packetStart = []; % Initialize output
% Shift data for correlation
rxDelayed = rxSig(symbolLength+1:end,:); % Delayed samples
rx = rxSig(1:end-symbolLength,:); % Actual samples
% Sum output on multiple receive antennas
C = sum(filter(weights,1,(conj(rxDelayed).*rx)),2);
CS = C(symbolLength:end)./symbolLength;
% Sum output on multiple receive antennas
P = sum(filter(weights,1,(abs(rxDelayed).^2+...
abs(rx).^2)/2)./symbolLength,2);
PS = P(symbolLength:end)+pNoise;
Mn = abs(CS).^2./PS.^2;
N = Mn > threshold;
if (sum(N) >= symbolLength*1.5)
found = index(N);
packetStart = found(1)-1;
% Check the relative distance between peaks relative to the first
% peak. If this exceed three times the symbol length then the
% packet is not detected.
if sum((found(2:symbolLength)-found(1))>symbolLength*3)
packetStart = [];
end
end
end
% [EOF]