802.11ad waveform generation with beamforming -凯发k8网页登录

introduction

ieee 802.11ad [ ] defines the directional multi-gigabit (dmg) transmission format operating at 60 ghz. to overcome the large path loss experienced at 60 ghz, the ieee 802.11ad standard is designed to support directional beamforming. by using phased antenna arrays you can apply an antenna weight vector (awv) to focus the antenna pattern in the desired direction. each packet is transmitted on all array elements, but the awv applies a phase shift to each element to steer the transmission. the quality of a communication link can be improved by appending optional training fields to dmg packets, and testing different awvs at the transmitter or receiver. this process is called beam refinement.

a dmg packet consists of the following fields:

the stf and ce fields form the preamble. the preamble, header, and data fields of a dmg packet are transmitted with the same awv. for transmitter beam refinement training, up to 64 training (trn) subfields can be appended to the packet. each trn subfield is transmitted using a different awv. this allows the performance of up to 64 different awvs to be measured, and the awv for the preamble, header, and data fields to be refined for subsequent transmissions. ce subfields are periodically transmitted, one for every four trn subfields, among the trn subfields. each ce subfield is transmitted using the same awv as the preamble. to allow the receiver to reconfigure agc before receiving the trn subfields, the trn subfields are preceded by agc subfields. for each trn subfield, an agc subfield is transmitted using the same awv applied to the individual trn subfield. this allows a gain to be set at the receiver, suitable to measuring all trn subfields. the diagram below shows the packet structure with four agc and trn subfields numbered and highlighted. therefore, four awvs are tested as part of beam refinement. the same awvs are applied to agc and trn subfields with the same number.

this example simulates transmitter training by applying different awvs to each of the training subfields to steer the transmission in multiple directions. the strength of each training subfield is evaluated at a receiver by evaluating the far-field plane wave to determine which transmission awv is optimal. this simulation does not include a channel or path loss.

this example requires wlan toolbox and phased array system toolbox.

waveform specification

the waveform is configured for a dmg packet transmission with single-carrier modulation (sc) physical layer, a 100-byte physical layer service data unit (psdu), and four transmitter training subfields. the four training subfields allow four awvs to be tested for beam refinement. using the function , create a dmg configuration object. a dmg configuration object specifies transmission parameters.

dmg = wlandmgconfig;
dmg.mcs = 1;              % single-carrier modulation
dmg.traininglength = 4;   % use 4 training subfields
dmg.packettype = 'trn-t'; % transmitter training
dmg.psdulength = 100;     % bytes

beamforming specification

the transmitter antenna pattern is configured as a 16-element uniform linear array with half-wavelength spacing. using the objects (phased array system toolbox) and (phased array system toolbox), create the phased array and the awvs. the location of the receiver for evaluating the transmission is specified as an offset from the boresight of the transmitter.

receiveraz = 6; % degrees off the transmitter's boresight

a uniform linear phased array with 16 elements is created to steer the transmission.

n = 16;                      % number of elements
c = physconst('lightspeed'); % propagation speed (m/s)
fc = 60.48e9;                % center frequency (hz)
lambda = c/fc;               % wavelength (m)
d = lambda/2;                % antenna element spacing (m)
txarray = phased.ula('numelements',n,'elementspacing',d);

the awvs are created using a (phased array system toolbox) object. five steering angles are specified to create five awvs, one for the preamble and data fields, and one for each of the four the training subfields. the preamble and data fields are transmitted at boresight. the four training subfields are transmitted at angles around boresight.

% create a directional steering vector object
steeringvector = phased.steeringvector('sensorarray',txarray);
% the directional angle for the preamble and data is 0 degrees azimuth, no
% elevation, therefore at boresight. [azimuth; elevation]
preambledataangle = [0; 0]; 
% each of the four training fields uses a different set of weights to steer
% to a slightly different direction. [azimuth; elevation]
trnangle = [[-10; 0] [-5; 0] [5; 0] [10; 0]];
% generate the weights for all of the angles
weights = steeringvector(fc,[preambledataangle trnangle]);
% each row of the awv is a weight to apply to a different antenna element
preambledataawv = conj(weights(:,1)); % awv used for preamble, data and ce fields
trnawv = conj(weights(:,2:end));      % awv used for each trn subfield

using the helper function, the array response shows the direction of the receiver is most aligned with the direction of training subfield trn-sf3.

plotarrayresponse(txarray,receiveraz,fc,weights);

generate baseband waveform

use the configured dmg object and a psdu filled with random data as inputs to the waveform generator, . the waveform generator modulates psdu bits according to a format configuration.

% create a psdu of random bits
s = rng(0); % set random seed for repeatable results
psdu = randi([0 1],dmg.psdulength*8,1);
% generate packet
tx = wlanwaveformgenerator(psdu,dmg);

apply weight vectors to each field

a (phased array system toolbox) object is created to apply the awvs to the waveform, combine the radiated signal from each element to form a plane wave, and determine the plane wave at the angle of interest, receiveraz. each portion of the dmg waveform tx is passed through the radiator with a specified set of awvs, and the angle at which to evaluate the plane wave.

radiator = phased.radiator;
radiator.sensor = txarray;        % use the uniform linear array
radiator.weightsinputport = true; % provide awv as argument
radiator.operatingfrequency = fc; % frequency in hertz
radiator.combineradiatedsignals = true; % create plane wave
% the plane wave is evaluated at a direction relative to the radiator
steerangle = [receiveraz; 0]; % [azimuth; elevation]
% the beamformed waveform is evaluated as a plane wave at the receiver
planewave = zeros(size(tx));
% get indices for fields
ind = wlanfieldindices(dmg);
% get the plane wave while applying the awv to the preamble, header, and data
idx = (1:ind.dmgdata(2));
planewave(idx) = radiator(tx(idx),steerangle,preambledataawv);
% get the plane wave while applying the awv to the agc and trn subfields
for i = 1:dmg.traininglength
    % agc subfields
    agcsfidx = ind.dmgagcsubfields(i,1):ind.dmgagcsubfields(i,2);
    planewave(agcsfidx) = radiator(tx(agcsfidx),steerangle,trnawv(:,i));
    % trn subfields
    trnsfidx = ind.dmgtrnsubfields(i,1):ind.dmgtrnsubfields(i,2);
    planewave(trnsfidx) = radiator(tx(trnsfidx),steerangle,trnawv(:,i));
end
% get the plane wave while applying the awv to the trn-ce
for i = 1:dmg.traininglength/4
    trnceidx = ind.dmgtrnce(i,1):ind.dmgtrnce(i,2);
    planewave(trnceidx) = radiator(tx(trnceidx),steerangle,preambledataawv);
end

evaluate the beamformed waveform

the helper function plots the magnitude of the beamformed plane wave. when evaluating the magnitude of the beamformed plane wave we can see that the fields beamformed in the direction of the receiver are stronger than other fields.

plotdmgwaveform(planewave,dmg,'beamformed plane wave with fields highlighted');

rng(s); % restore random state

conclusion

this example showed how to generate an ieee 802.11ad dmg waveform and apply awvs to different portions of the waveform. the example uses wlan toolbox to generate a standard-compliant waveform, and phased array system toolbox to apply the awvs and evaluate the magnitude of the resultant plane wave in the direction of a receiver.

selected bibliography