DGLF-MPF: A Destination-Guided Motion Planning Framework for Autonomous Vehicles in Lane-Free Traffic
点击次数：
影响因子：7.9
DOI码：10.1016/j.trc.2025.105440
发表刊物：Transportation Research Part C: Emerging Technologies
关键字：Lane-free trafficIntelligent Agent ModelMotion planningAutonomous VehiclesDestination-GuidedTraffic flow simulation
摘要：Lane-Free Traffic (LFT) emerges as a promising transportation strategy, allowing autonomous vehicles (AVs) to move without the limitations of traditional lane-based systems. However, integrating short-term destination requirements, such as navigating ramps, into the real-time motion planning of non-connected AVs remains a major challenge. To address this, we propose the Destination-Guided, Lane-Free AV Motion Planning Framework (DGLF-MPF), which integrates lane-free following (LFF) with the intelligent agent model (IAM) based on social force theory. This framework incorporates short-term destination requirements to guide AV behavior planning, enabling efficient navigation in dynamic LFT. Calibration with the CitySim dataset confirms that DGLF-MPF accurately simulates realistic vehicle movements in on-ramp and off-ramp scenarios, effectively capturing personalized driving behaviors. Simulations in multi-ramp lane-free environments show that, compared to traditional lane-based systems, DGLF-MPF significantly improves traffic flow efficiency and space utilization. It showcases an obvious self-organizing phenomenon in complex ramp-mainline interactions. By incorporating dynamic avoidance and longitudinal deceleration mechanisms, the framework reduces traffic conflicts during merging and diverging, thereby enhancing driving comfort and traffic flow stability. These results provide strong support for implementing LFT management in more complex, real-world traffic scenarios.
论文类型：期刊论文
论文编号：105440
期号：182
是否译文：否
发表时间：2025-11-15
收录刊物：SCI、EI