This paper develops a bi-level control framework that considers the mixed traffic flow of autonomous vehicles (AVs) and human-driven vehicles (HVs) in transport networks. Our framework consists of a multi-class dynamic traffic assignment at the upper level to determine the optimal traffic flow splits for vehicles, while an end-to-end trajectory planning algorithm for AVs is incorporated into the lower level to attain the eco-driving strategy in the mixed traffic environment. The macroscopic decisions (e.g. traffic flow splits) at the upper level can directly affect the progression of the mixed traffic flows, while microscopic decisions (e.g. trajectory profiles) at the lower level can provide realistic feedback (e.g. link supply capacities) to guide the search direction of the upper level and ultimately improve the obtained solution. Besides, we also introduce an effective solution method for this framework that solves the mixed-integer linear programming models at the upper and lower levels. Numerical results indicate that even a low penetration rate of AVs can significantly reduce fuel consumption. Furthermore, AVs can reduce the total travel time of traffic users, eventually mitigating the congestion in the networks.
Optimal Trajectory Planning Framework for a Mixed Traffic Network
2022-10-08
478042 byte
Conference paper
Electronic Resource
English
Framework for Aircraft Trajectory Planning Toward an Efficient Air Traffic Management
Online Contents | 2012
|Optimal Trajectory Planning for Robot
Trans Tech Publications | 2011
|Trajectory Planning Framework Combining Replanning and Hierarchical Planning
Transportation Research Record | 2023
|British Library Conference Proceedings | 2022
|