Dynamic IRS Allocation for Spectrum-Sharing MIMO Communication and Radar Systems

TL;DR

This work tackles spectrum sharing between cellular communications and radar sensing in a MIMO setting using a single intelligent reflecting surface (IRS). It proposes a WMMSE-based joint optimization that simultaneously tunes transmit/receive beamformers, IRS phase shifts, and a dynamic allocation of IRS elements between communication ($N_c$) and sensing ($N_s$) via $\eta = \frac{1}{1+(\gamma_c/\gamma_s)^{\beta}}$, with $N_c=N\eta$ and $N_s=N(1-\eta)$. The framework employs a block coordinate descent strategy to iteratively solve for beamformers, phase shifts, and element allocation, demonstrating significant gains in both communication and sensing SINRs over fixed allocations. Results indicate that a single IRS with dynamic allocation can outperform setups using multiple IRSs with fixed allocation, offering reduced deployment cost and signaling overhead while enabling robust JCAS performance. The approach integrates cross-interference considerations into beamforming and phase-shift design, and future work addresses imperfect CSI and decentralized allocation.

Abstract

This paper investigates the use of intelligent reflecting surfaces (IRS) to assist cellular communications and radar sensing operations in a communications and sensing setup. The IRS dynamically allocates reflecting elements to simultaneously localize a target and assist a user's communication. To achieve this, we propose a novel optimization framework that jointly addresses beamforming design and IRS element allocation. Specifically, we formulate a Weighted Minimum Mean Square Error (WMMSE)-based approach that iteratively optimizes the transmit and receive beamforming vectors, IRS phase shifts, and element allocation. The allocation mechanism adaptively balances the number of IRS elements dedicated to communication and sensing subsystems by leveraging the signal-to-noise-plus-interference-ratio (SINR) between the two. The proposed solution ensures efficient resource utilization while maintaining performance trade-offs. Numerical results demonstrate significant improvements in both communication and sensing SINRs under varying system parameters.

Figures (4)

• Figure 1: MIMO communication and radar sensing network setup assisted by a single IRS.
• Figure 2: Comparison of fixed and dynamic IRS element allocation.
• Figure 3: Effect of SINRs on the allocation of IRS elements with varying number of BS antennas.
• Figure 4: Effect of SINRs on the allocation of IRS elements with varying number of total IRS elements.