Interference Reduction Design for Improved Multitarget Detection in ISAC Systems

TL;DR

This work tackles multitarget detection in ISAC by designing a waveform that maximizes the intended-user signal power while suppressing both inter-user and sensing interference. It casts the design as a non-convex optimization and makes it tractable through a symmetric idempotent sensing covariance matrix $C$ and a distribution-based interference model, solved with non-disciplined and disciplined convex programming plus successive convex approximation (SCA). Key contributions include a distribution-based interference cancellation framework, a constructive idempotent covariance design, and a practical SCA-based algorithm that converges to stable waveforms. Compared with state-of-the-art DFRC schemes, the proposed method demonstrates superior beampattern accuracy and stronger interference suppression, along with reduced computational complexity. The approach holds promise for more reliable multitarget detection in ISAC systems and supports scalable integration of sensing within high-rate communications.

Abstract

The advancement of wireless communication systems toward 5G and beyond is spurred by the demand for high data rates, exceedingly dependable low-latency communication, and extensive connectivity that aligns with sensing requisites such as advanced high-resolution sensing and target detection. Consequently, embedding sensing into communication has gained considerable attention. In this work, we propose an alternative approach for optimizing integrated sensing and communication (ISAC) waveform for target detection by concurrently maximizing the power of the communication signal at an intended user and minimizing the multi-user and sensing interference. We formulate the problem as a non-disciplined convex programming (NDCP) optimization and we use a distribution-based approach for interference cancellation. Precisely, we establish the distribution of the communication signal and the multi-user communication interference received by the intended user, and thereafter, we establish that the sensing interference can be distributed as a centralized Chi-squared if the sensing covariance matrix is idempotent. We design such a matrix based on the symmetrical idempotent property. Additionally, we propose a disciplined convex programming (DCP) form of the problem, and using successive convex approximation (SCA), we show that the solutions can reach a stable waveform for efficient target detection. Furthermore, we compare the proposed waveform with state of the art radar-communication waveform designs and demonstrate its superior performance by computer simulations.

Figures (3)

• Figure 1: System model
• Figure 2: Data structure, convergence and sample prediction.
• Figure 3: Beampatterns comparison