Antenna array geometry for mobile ad-hoc networks

Small antenna arrays are advantageous for ad-hoc networks. Typically antenna array geometry is designed using simplified objective measures to make analysis tractable or maximize search efficiency. This paper investigates array design using fitness functions more tightly related to network performance. Particle swarm optimization (PSO) is used to design planar antenna arrays with minimum mean-square error or noise-to-signal ratio. Fitness functions are averaged across look angle to include effects of random signal and interference directions, resulting in a better measure of performance in real-world conditions. We found that irregular “T” or “diamond” geometries can outperform more common arrays such as Y or square arrays.