What Is Array Signal Processing?

Array signal processing is a branch of signal processing concerned with the acquisition, analysis, and interpretation of signals captured by multiple spatially distributed sensors operating simultaneously. By combining the outputs of individual sensors in time and space, array processing systems can determine the direction of signal sources, separate simultaneous signals, and improve signal-to-noise ratio beyond what a single sensor can achieve. The sensors, called array elements, may be antennas, microphones, hydrophones, accelerometers, or any transducer sensitive to the wavefield of interest, and they are arranged in geometries ranging from uniform linear arrays to two-dimensional planar, circular, or fully three-dimensional configurations. The field draws on statistical estimation theory, linear algebra, adaptive filtering, and electromagnetic and acoustic wave propagation.

Array signal processing originated in radar and sonar research during the mid-twentieth century, where the need to locate targets in azimuth and elevation motivated the development of phased-array systems. The mathematical framework generalized rapidly to acoustic, seismic, and wireless communications applications, and it now constitutes a central tool in systems from mobile base station antennas to medical ultrasound imaging.

Direction-of-Arrival Estimation

Direction-of-arrival (DOA) estimation is one of the fundamental problems in array signal processing: given the outputs of an array of sensors, estimate the angles from which impinging wavefronts originate. When a narrowband signal arrives at an array from a particular direction, each element receives the same waveform but with a phase shift determined by the element's position relative to the wavefront and the signal's wavelength. Classical methods such as the delay-and-sum beamscanner form a conventional beam and sweep it in angle, producing a spatial spectrum with a peak at the signal direction. High-resolution subspace methods achieve finer angular discrimination by decomposing the array covariance matrix into signal and noise subspaces: the MUSIC algorithm (Multiple Signal Classification) constructs a noise subspace projection that exhibits deep nulls in the directions of incident signals, and ESPRIT (Estimation of Signal Parameters via Rotational Invariance Technique) exploits a translational invariance structure present in certain array geometries. A tutorial survey of classical and modern DOA estimation methods reviews these and related approaches in depth.

Adaptive Arrays and Beamforming

Adaptive arrays adjust their spatial response dynamically in response to the observed signal environment, rather than using fixed steering weights. The minimum variance distortionless response (MVDR) beamformer, also called the Capon beamformer, computes weights that minimize total output power subject to a unity-gain constraint in the look direction, automatically placing nulls toward interferers without requiring prior knowledge of their directions. Linearly constrained minimum variance (LCMV) methods extend this approach by enforcing multiple simultaneous constraints. Adaptive arrays are foundational to spatial multiplexing in modern cellular systems: massive MIMO (multiple-input multiple-output) deployments use arrays of hundreds of antenna elements to serve many users simultaneously in the same frequency band. Time-of-arrival estimation, which measures the propagation delay of a signal from source to each element, complements angle estimation and enables full three-dimensional source localization when arrays are calibrated and synchronized.

Acoustic Arrays and Transducers

In acoustic applications, sensor arrays are built from microphones, hydrophones, or piezoelectric transducers arranged to capture sound fields. Microphone arrays in teleconferencing and voice assistant systems apply delay-and-sum beamforming to enhance speech from a target direction and suppress room reverberation and background noise. Underwater acoustic arrays (towed arrays or fixed seafloor arrays) use hydrophones to detect and track submarines, marine mammals, and seismic events. Acoustic transducers in medical ultrasound imaging operate as transmit-and-receive arrays, where the phased-array ultrasound transducer steers and focuses the acoustic beam electronically without mechanical scanning, enabling real-time volumetric imaging of internal organs. Seismic sensor arrays similarly apply array processing to detect and characterize underground sources with sub-wavelength resolution.

Applications

Array signal processing has applications across many engineering and scientific domains, including:

• Radar target detection, tracking, and angle estimation
• Sonar and underwater acoustic surveillance
• Wireless base station beamforming and massive MIMO communications
• Speech enhancement and noise suppression in microphone arrays
• Medical ultrasound imaging using phased-array transducers
• Radio astronomy and very-long-baseline interferometry (VLBI)