Decentralized Localization of Distributed Antenna Array Elements Using an Evolutionary Algorithm

TL;DR

This paper presents a decentralized method for localizing distributed antenna array elements using an evolutionary algorithm, enabling coherent operations like beamforming in complex environments with high precision.

Contribution

It introduces a novel combination of two-way time transfer, multidimensional scaling, and differential evolution to accurately localize array elements without direct link measurements.

Findings

Achieved a mean localization error of 0.82 mm in laboratory tests.

Supported distributed beamforming up to 24.3 GHz.

Demonstrated effective localization with 34 dB SNR.

Abstract

Distributed phased arrays have recently garnered interest in applications such as satellite communications and high-resolution remote sensing. High-performance coherent distributed operations such as distributed beamforming are dependent on the ability to synchronize the spatio-electrical states of the elements in the array to the order of the operational wavelength, so that coherent signal summation can be achieved at any arbitrary target destination. In this paper, we address the fundamental challenge of precise distributed array element localization to enable coherent operation, even in complex environments where the array may not be capable of directly estimating all nodal link distances. We employ a two-way time transfer technique to synchronize the nodes of the array and perform internode ranging. We implement the classical multidimensional scaling algorithm to recover a decentralized array geometry from a set of range estimates. We also establish the incomplete set of range estimates as a multivariable non-convex optimization problem, and define the differential evolution algorithm which searches the solution space to complete the set of ranges. We experimentally demonstrate wireless localization using a spectrally-sparse pulsed two-tone waveform with 40 MHz tone separation in a laboratory environment, achieving a mean localization error vector magnitude of 0.82 mm in an environment with an average link SNR of 34 dB, theoretically supporting distributed beamforming operation up to 24.3 GHz.