Conventional Beamforming

Uses for Beamformers

You can use a beamformer to spatially filter the arriving signals. Accentuating or attenuating signals that arrive from specific directions helps you distinguish between signals of interest and interfering signals from other directions.

Support for Conventional Beamforming

You can implement a narrowband phase shift beamformer using phased.PhaseShiftBeamformer. When you use this object, you must specify these aspects of the situation you are simulating:

  • Sensor array

  • Signal propagation speed

  • System operating frequency

  • Beamforming direction

For wideband beamformers, see Wideband Beamforming.

Narrowband Phase Shift Beamformer with a ULA

Construct a ULA with 10 elements. Assume the carrier frequency is 1 GHz and set the array element spacing to be one-half the carrier frequency wavelength.

fc = 1e9;
lambda = physconst('LightSpeed')/fc;
hula = phased.ULA('NumElements',10,'ElementSpacing',lambda/2);

The ULA sensors are isotropic antenna elements (see phased.IsotropicAntennaElement). Set the frequency range of the antenna elements to position the carrier frequency in the middle of the operating range.

hula.Element.FrequencyRange = [8e8 1.2e9];

Simulate a test signal. For this example, use a simple rectangular pulse.

t = linspace(0,0.3,300)';
testsig = zeros(size(t));
testsig(201:205)= 1;

Assume the rectangular pulse is incident on the ULA from an angle of 30 degrees azimuth and 0 degrees elevation. Use the collectPlaneWave method of the ULA object to simulate reception of the pulse waveform from the specified angle.

angle_of_arrival = [30;0];
x = collectPlaneWave(hula,testsig,angle_of_arrival,fc);

x is a matrix with ten columns. Each column represents the received signal at one of the array elements.

Corrupt the columns of x with complex-valued Gaussian noise. Reset the default random number stream for reproducible results. Plot the magnitudes of the received pulses at the first four elements of the ULA.

rng default
npower = 0.5;
x = x + sqrt(npower/2)*(randn(size(x))+1i*randn(size(x)));
plot(t,abs(x(:,1))); title('Element 1 (magnitude)');
axis tight;  ylabel('Magnitude');
plot(t,abs(x(:,2))); title('Element 2 (magnitude)');
axis tight; ylabel('Magnitude');
plot(t,abs(x(:,3))); title('Element 3 (magnitude)');
axis tight; xlabel('Seconds'); ylabel('Magnitude');
plot(t,abs(x(:,4))); title('Element 4 (magnitude)');
axis tight; xlabel('Seconds'); ylabel('Magnitude');

Construct your phase-shift beamformer. Set the WeightsOutputPort property to true to output the spatial filter weights.

hbf = phased.PhaseShiftBeamformer('SensorArray',hula,...

Apply the step method for the phase shift beamformer. The step method computes and applies the correct weights for the specified angle. The phase-shifted outputs from the ten array elements are then summed.

[y,w] = step(hbf,x);

Plot the magnitude of the output waveform along with the original waveform for comparison.

plot(t,abs(testsig)); axis tight;
title('Original Signal'); ylabel('Magnitude');
plot(t,abs(y)); axis tight;
title('Received Signal with Beamforming');
ylabel('Magnitude'); xlabel('Seconds');

To examine the effect of the beamforming weights on the array response, plot the array normalized power response both with—and without—the beamforming weights.

azang = -180:30:180;
title('Array Response without Beamforming Weights');
title('Array Response with Beamforming Weights');

Related Examples

Was this topic helpful?