Due to the huge traffic flow and complex weaving, unsignalized roundabout is one of the most challenging scenarios for autonomous driving. Though cooperative autonomous driving has the potential to address this challenge, solving the multi-vehicle cooperation problem usually can't meet the real-time requirement. To efficiently solve this problem, this paper proposes a self-organizing cooperative control framework considering both “whom to cooperate with” and “how to cooperate”. At the high level, based on the self-organization theory and traffic safety theory, a finite state machine (FSM) is designed for connected automated vehicles (CAVs) to switch states, which reduce the size of the multi-vehicle cooperation problem by selecting allies and forming organizations real-time. At the low level, to solve the multi-vehicle cooperation problem, model predictive control (MPC) is used to minimize an objective function related to safety, efficiency and comfort of CAVs. Finally, simulations and micro-vehicle experiment are conducted to verify the effectiveness of the proposed method. The result demonstrates the superiority of the method in decreasing calculation time and enhancing traffic safety by comparing with an existing benchmark.
A Self-organizing Cooperative Control Framework for Connected Automated Vehicles at Unsignalized Roundabouts
2024-09-24
1903106 byte
Conference paper
Electronic Resource
English
Unsignalized Intersections: Roundabouts
Springer Verlag | 2020
|Omnidirectional Video Instrumentation For Unsignalized Intersections and Roundabouts
British Library Conference Proceedings | 1997
|When Do Drivers Yield to Cyclists at Unsignalized Roundabouts?
Transportation Research Record | 2019
|