构建云控智驾测试平台，真实模拟通信延迟对安全与舒适的影响

Testing cloud-controlled intelligent connected vehicles (ICVs) requires simulation environments that faithfully emulate both vehicle behavior and realistic communication latencies. This paper proposes a latency-aware co-simulation platform integrating CarMaker and Vissim to evaluate safety and comfort under real-world vehicle-to-cloud (V2C) latency conditions. Two communication latency models, derived from empirical 5G measurements in China and Hungary, are incorporated and statistically modeled using Gamma distributions. A proactive conflict module (PCM) is proposed to dynamically control background vehicles and generate safety-critical scenarios. The platform is validated through experiments involving an exemplary system under test (SUT) across six testing conditions combining two PCM modes (enabled/disabled) and three latency conditions (none, China, Hungary). Safety and comfort are assessed using metrics including collision rate, distance headway, post-encroachment time, and the spectral characteristics of longitudinal acceleration. Results show that the PCM effectively increases driving environment criticality, while V2C latency primarily affects ride comfort. These findings confirm the platform's effectiveness in systematically evaluating cloud-controlled ICVs under diverse testing conditions.