Digital antenna array

Lore & Background

The history of the digital antenna array began with the theory of multichannel analysis in the 1920s, which evolved into three-channel antenna analyzers by the 1940s. By the late 1950s, effective signal processing in radars led to the use of electronic computers. In 1957, Ben S. Meltont and Leslie F. Bailey published an article on using algebraic operations for signal processing with electronic circuits or analog computers. In 1960, the idea of using high-speed computers for direction finding was first implemented to locate earthquake epicenters, notably by B. A. Bolt and independently by Flinn at the Australian National University. Although these experiments used data input cards, they were a decisive step toward the DAA. The remaining challenge—direct digital data input from sensors without punch cards—was solved after 1962 in the former USSR, addressing super-Rayleigh resolution of emission sources.

The core approach in DAA is digital beamforming, performed after analog-to-digital conversion in receivers or before digital-to-analog conversion in transmitters. Digital signals can be transformed and combined in parallel to produce multiple output signals. Before beamforming, channel characteristics are corrected using a test source or heterodyne signal, applicable to both receiving and transmission channels. Limitations on direction-of-arrival accuracy and interference suppression are associated with jitter in ADCs and DACs.

Signal processing methods include the maximum likelihood beamformer, which models noise as stationary Gaussian white noise and the signal as deterministic but unknown. The Bartlett beamformer extends conventional spectral analysis, with power given by v^H R v. The Capon (MVDR) beamformer uses the inverse of the covariance matrix for better resolution but higher complexity, though GPU computing is narrowing the gap. The MUSIC beamformer decomposes the covariance matrix into signal and noise subspaces, using the noise subspace in the denominator for improved DOA estimation. An alternative is the ESPRIT algorithm. An important trend is the use of artificial intelligence technologies in DAA signal processing.

Reader's Guide

The digital antenna array represents a significant evolution in antenna technology, enabling simultaneous multi-directional signal estimation and adaptive integration times for both distant and fast-moving objects. Its development, rooted in multichannel analysis theory from the 1920s, culminated in practical realization after 1962 under Vladimir Varyukhin in the USSR. The DAA's digital beamforming approach, performed after ADC or before DAC, allows parallel processing and flexible signal combination, with channel correction via test sources or heterodyne signals. Various beamforming algorithms—Bartlett, Capon (MVDR), MUSIC, and ESPRIT—offer trade-offs between resolution and computational complexity, with GPU advances enabling real-time Capon processing. The integration of artificial intelligence is an emerging trend. DAAs are applied in radars, massive MIMO communications, sonars, and medical ultrasound sensors. The technology's legacy includes improved direction-of-arrival estimation and interference suppression, though accuracy is limited by ADC/DAC jitter.

Did You Know?

• The theory of digital antenna arrays began emerging in the 1920s as multichannel analysis.
• The first practical use of high-speed computers for direction finding was in 1960 to locate earthquake epicenters.
• Digital beamforming in DAA can be performed after ADC in receivers or before DAC in transmitters.
• The MUSIC beamformer uses the noise subspace of the covariance matrix for better DOA estimation than Capon.