Uplink Multi-RSU Cooperative Sensing Strategy for Integrated Sensing and Communication System

Integrated Sensing and Communication (ISAC) is expected to be a key technology for next-generation wireless networks due to its efficient use of spectrum and device resources. In internet of vehicles (IoV) networks, however, mutual interference can substantially degrade sensing accuracy in multi-user, multi-target scenarios. Moreover, the limited sensing capability of a single roadside unit (RSU) further highlights the need for cooperative sensing among multiple RSUs. This paper investigates multi-RSU cooperative sensing within an uplink ISAC framework. Firstly, the stochastic geometry is employed to model the distribution of communication user equipments (COM-UEs), and a closed-form expression for interference power is derived. Aiming at mitigate the mutual interference effectively, we then investigate receive beamforming design to minimize the sensing beampattern matching error while meeting communication requirements. To address time offset (TO) and carrier frequency offset (CFO) due to asynchronous transceivers, we propose a cross-path cross-correlation (CPCC) scheme between line-of-sight (LOS) and non-LOS (NLOS) signals, effectively reducing ambiguity in range and velocity estimation. Furthermore, a symbol-level fusion method for multi-RSUs’ sensing information is proposed. The phase information matrices extracted from echo signals across multi-RSUs are conducted independent matching and weighted fusion to enhance cooperative sensing accuracy. Simulations show that our approach outperforms single-RSU sensing, with improvements of 52.3% and 64.1% in localization and velocity estimation, respectively. Compared to data-level fusion, the symbol-level approach yields further enhancements of 44.8% in localization and 64.6% in velocity estimation.