Vehicle Localization and False Targets Elimination in 6G Integrated Sensing and Communication Systems

Enabling safe and efficient driving in complex traffic environments, accurate vehicle localization is paramount for advanced transportation management systems. A densely deployed roadside unit (RSU) infrastructure is evolving from a communication-only network to one with integrated sensing and communication (ISAC) capabilities, offering essential status information for vehicles. However, the precise localization of vehicles is challenging due to the presence of multiple echoes from numerous vehicles and false targets generated by multipath effects. To overcome this issue, this paper proposes a novel location parameter estimation method using orthogonal time frequency space (OTFS) signals under rapidly time-varying channels. In particular, we utilize the wideband capabilities of 6G systems to model vehicles as extended targets. Meanwhile, the high-resolution range profiles (HRRP) are introduced to uniquely characterize each vehicle, which allows us to accurately associate each path with their corresponding vehicles, thereby eliminating false targets and enhancing localization accuracy. Additionally, we design a sparse Bayesian learning (SBL) algorithm to estimate unknown parameters iteratively. Simulation results demonstrate that the proposed scheme achieves better NMSE and localization performance in intricate traffic scenarios.