Signal Processing

Array Response Vector

Q: How is the array response vector used in beamforming?

To steer the array beam toward angle theta_0, the beamforming weight vector is set equal to the array response vector for that angle: w = a(theta_0). The array output y = w^H * x, where x is the received signal vector. When the weights match the steering vector, signals from theta_0 add coherently (constructive interference) and the array gain equals N (the number of elements). Signals from other directions add with random phases and are suppressed. For adaptive beamforming, the weights are optimized using the data covariance matrix to simultaneously steer toward the desired signal and place nulls toward interferers.

Q: What is the array manifold?

The array manifold is the set of all array response vectors as the angle sweeps through all possible directions. For a uniform linear array, the manifold is a curve in N-dimensional complex space parameterized by the angle theta. For a 2D planar array, it is a surface parameterized by azimuth and elevation. The shape of the manifold determines the array's ability to resolve closely spaced sources. If two angles produce nearly identical steering vectors, the array cannot distinguish them. This is why element spacing, array geometry, and mutual coupling all affect angular resolution.

Q: How does the steering vector relate to DOA estimation?

Direction-of-arrival (DOA) algorithms like MUSIC and ESPRIT use the steering vector to search for the angles that best explain the received data. MUSIC computes the noise subspace from the data covariance matrix, then evaluates 1/||P_noise * a(theta)||^2 for all candidate angles. Peaks in this pseudo-spectrum indicate source directions. The resolution depends on how quickly the steering vector changes with angle (related to the array aperture) and the SNR. For a ULA with N elements and half-wavelength spacing, the angular resolution is approximately 2/(N*cos(theta)) radians.

A complex N-dimensional vector whose entries describe the relative phase (and optionally amplitude) at each element of an antenna array for a plane wave arriving from a specific direction θ. Also called the steering vector or array manifold vector, it is the fundamental mathematical object in phased array processing: beamforming weights, direction-of-arrival (DOA) estimation algorithms (MUSIC, ESPRIT), and MIMO channel models are all expressed in terms of the array response vector.

Category: Signal Processing

Also called: Steering Vector

Dimension: N (number of elements)

Understanding the Array Response Vector

When a plane wave arrives at an antenna array from angle θ, each element receives the signal at a slightly different time due to the geometric path-length difference. For a uniform linear array (ULA) with element spacing d, element n receives the signal with a time delay of n×d×sin(θ)/c relative to the reference element. In the frequency domain, this time delay becomes a phase shift of n×kd×sin(θ), where k = 2π/λ.

The array response vector collects these phase shifts into a single vector: a(θ) = [1, e^{jkd sinθ}, e^{j2kd sinθ}, ..., e^{j(N−1)kd sinθ}]^T. This vector completely describes how the array "sees" a signal from direction θ. To steer the beam toward θ₀, the beamformer sets its weights w = a(θ₀), maximizing the output for signals from that direction.

Steering Vector for ULA

Array response vector:
a(θ) = [1, e^jψ, e^j2ψ, ..., e^j(N−1)ψ]^T
where ψ = (2πd/λ) × sin(θ)

Beamformer output:
y = w^H × x = a^H(θ₀) × x
Array gain at θ₀ = N (coherent combining)

MUSIC pseudo-spectrum:
P_MUSIC(θ) = 1 / ||P_N × a(θ)||²
where P_N = noise subspace projector

Example: 8-element ULA at 10 GHz, d = λ/2 = 15 mm. Beamwidth ≈ 2/(8×cosθ) rad ≈ 14.3° at broadside.

Array Geometries and Steering Vectors

Geometry	Steering Vector Parameters	Angular Coverage	DOA Estimation
Uniform Linear (ULA)	θ only (1D)	180° (half-space)	Azimuth only
Uniform Circular (UCA)	θ (azimuth, 360°)	360° azimuth	Azimuth, ambiguous elevation
Uniform Rectangular (URA)	θ, φ (2D)	Hemisphere	Azimuth + elevation
Conformal	Geometry-dependent	Application-specific	Full 3D (with calibration)

Common Questions

Frequently Asked Questions

How is the array response vector used in beamforming?

Set w = a(θ₀) to steer toward θ₀. Signals from that direction add coherently (gain = N). Other directions add with random phases and are suppressed. Adaptive beamforming optimizes weights using the covariance matrix to simultaneously steer toward the signal and null interferers.

What is the array manifold?

The set of all steering vectors as angle sweeps through all directions. For a ULA, it is a curve in N-dimensional complex space. If two angles produce nearly identical steering vectors, the array cannot resolve them. Element spacing, geometry, and mutual coupling all affect resolution capability.

How does the steering vector relate to DOA estimation?

MUSIC evaluates 1/||P_N×a(θ)||² for all candidate angles. Peaks indicate source directions. Resolution depends on how quickly a(θ) changes with angle (proportional to array aperture). A ULA with N elements at λ/2 spacing resolves approximately 2/(N×cosθ) radians.

Related Terms

← ARQ Attention Mechanism (RF) →