Answers are grounded in this paper and cite their sources.
Cooperative perception (CoP) relies on the exchange of sensor data and intermediate perception results between connected and automated vehicles (CAVs) and roadside units (RSUs). Evaluating CoP models therefore requires a simulation environment that combines vehicle dynamics and perception with explicit vehicle-to-everything (V2X) communication models. The CAVISE simulation environment integrates CARLA and OpenCDA with SUMO and Artery for this purpose. Building on earlier CAVISE work, this paper describes the CAVISE API (CAPI), an interface architecture that coordinates bidirectional data exchange between OpenCDA and Artery, synchronized with simulation steps. Its modular message structure supports heterogeneous CoP payloads while preserving transmission outcomes for each agent produced by the network simulator. CAPI is described in the context of the complete CAVISE architecture. Six CoP scenarios are evaluated under three communication configurations: (1) direct exchange without network simulation, (2) Artery using the standard INET propagation model, and (3) Artery using the GEMV2 propagation model. In 4cars_coperception, AP@0.7 changes from 0.214 under direct exchange to 0.094 under INET and 0.097 under GEMV2. Additional cooperative driving automation scenario using an autonomous intersection management model is also evaluated to show the extensibility of CAPI to different CAV applications. Current scalability limitations of the integrated simulation workflow are also identified.
No takes yet. Share an insight, caveat, or question.
Khoroshilov et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: