Abstract:The presence of interfering scatterers fundamentally changes the design principle for MIMO sensing. Unlike the target-only case, where MIMO sensing sequence design reduces to optimizing the transmit sample covariance, this paper shows that scatterer-induced signal-dependent interference makes the Bayesian Fisher information depend on the full temporal precoding sequence. Consequently, for the MIMO sensing problem with scatterers using the Bayesian Cramér-Rao lower bound (BCRLB) as the objective, the entire sensing sequence must be designed explicitly, instead of just the precoding matrix. This paper considers such a precoding sequence design problem under hardware constraint for MIMO sensing for estimating the azimuth angles of multiple targets based on the prior information of both the targets and the scatterers. We formulate a worst-case BCRLB minimization across multiple target angles, yielding a max-min fractional program under constant-modulus or constant-norm hardware constraints. We further develop a constant-norm linear transform that converts the ratio objectives into linear forms, leading to an iterative algorithm with closed-form precoder updates. The framework extends to joint precoder-combiner design and multi-stage sensing with adaptive prior refinement. Numerical results demonstrate the effectiveness and the efficiency of the proposed algorithm, revealing sweeping-like beampatterns that illuminate target angular regions while suppressing interference from the scatterers.
From: Yiming Liu [view email]
[v1]
Mon, 15 Jun 2026 18:37:30 UTC (15,085 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。