Detection in Analog Sensor Networks with a Large Scale Antenna Fusion Center

TL;DR

This paper investigates distributed detection of Gaussian signals in large-scale antenna fusion centers, optimizing sensor transmission gains and revealing benefits of multiple antennas for maintaining detection performance.

Contribution

It introduces a convex optimization framework for sensor gain design in large-antenna fusion centers, showing how multiple antennas improve detection robustness at lower transmit power.

Findings

Constant detection probability with decreasing sensor gain as antennas increase

Multi-antenna fusion centers outperform single-antenna ones at lower power levels

Both setups achieve similar detection bounds at high transmit power

Abstract

We consider the distributed detection of a zero-mean Gaussian signal in an analog wireless sensor network with a fusion center (FC) configured with a large number of antennas. The transmission gains of the sensor nodes are optimized by minimizing the ratio of the log probability of detection (PD) and log probability of false alarm (PFA). We show that the problem is convex with respect to the squared norm of the transmission gains, and that a closed-form solution can be found using the Karush-Kuhn-Tucker conditions. Our results indicate that a constant PD can be maintained with decreasing sensor transmit gain provided that the number of antennas increases at the same rate. This is contrasted with the case of a single-antenna FC, where PD is monotonically decreasing with transmit gain. On the other hand, we show that when the transmit power is high, the single- and multi-antenna FC both asymptotically achieve the same PD upper bound.