Eigenvalue-Based Detection in MIMO Systems for Integrated Sensing and Communication

TL;DR

This paper introduces an eigenvalue-based detection method for MIMO ISAC systems that jointly optimizes sensing and communication performance, providing closed-form error expressions and outperforming traditional schemes.

Contribution

It develops a novel eigenvalue-based detector with closed-form error probabilities and formulates a joint power and threshold optimization for enhanced sensing and communication.

Findings

Proposed eigenvalue-based detector yields accurate detection probabilities.

Joint optimization significantly reduces total detection error.

Simulation shows outperforming of traditional CFAR-based schemes.

Abstract

This paper considers a MIMO Integrated Sensing and Communication (ISAC) system, where a base station simultaneously serves a MIMO communication user and a remote MIMO sensing receiver, without channel state information (CSI) at the transmitter. Existing MIMO ISAC literature often prioritizes communication rate or detection probability, typically under constant false-alarm rate (CFAR) assumptions, without jointly analyzing detection reliability and communication constraints. To address this gap, we adopt an eigenvalue-based detector for robust sensing and use a performance metric, the total detection error, that jointly captures false-alarm and missed-detection probabilities. We derive novel closed-form expressions for both probabilities under the eigenvalue detector, enabling rigorous sensing analysis. Using these expressions, we formulate and solve a joint power allocation and threshold optimization problem that minimizes total detection error while meeting a minimum communication rate requirement. Simulation results demonstrate that the proposed joint design substantially outperforms conventional CFAR-based schemes, highlighting the benefits of power- and threshold-aware optimization in MIMO ISAC systems.