干扰(通信)
计算机科学
物理
电子工程
噪音(视频)
电磁干扰
信号处理
信噪比(成像)
噪声测量
算法
邻道干扰
探测理论
雷达跟踪器
控制理论(社会学)
匹配滤波器
带宽(计算)
作者
Boyu Teng,Xiaojun Yuan
标识
DOI:10.1109/lwc.2026.3708142
摘要
High-accuracy wireless localization is a key enabler of integrated sensing and communication (ISAC), but its performance degrades under strong non-line-of-sight (NLoS) propagation and co-channel interference. Existing methods typically approximate interference as spatially white Gaussian noise, leading to model mismatch and biased line-of-sight (LoS) estimation. In this letter, we propose a localization framework that explicitly suppresses structured interference. A score-based generative model is employed to capture the non-Gaussian and spatially correlated interference prior in a data-driven manner. Based on this prior, a score-based turbo message passing (STMP) algorithm is developed for joint interference suppression and localization. Simulation results demonstrate that the proposed method significantly outperforms conventional maximum-likelihood schemes and achieves localization accuracy close to the interference-free Cramér–Rao bound.
科研通智能强力驱动
Strongly Powered by AbleSci AI