Hierarchical automatic lane-changing motion planning: from self-optimum to local-optimum
This work addresses traffic efficiency and safety in autonomous driving, though it appears incremental as it builds on existing hierarchical planning and MPC methods.
The researchers tackled the problem of minimizing lane change impacts on traffic flow by proposing a hierarchical automatic lane-changing algorithm that achieves local optimum, which reduced total loss for nearby vehicles and increased traffic speed and throughput in the lane change area.
In order to minimize the impact of lane change (LC) maneuver on surrounding traffic environment, a hierarchical automatic LC algorithm that could realize local optimum has been proposed. This algorithm consists of a tactical layer planner and an operational layer controller. The former generates a local-optimum trajectory. The comfort, efficiency, and safety of the LC vehicle and its surrounding vehicles are simultaneously satisfied in the optimization objective function. The later is designed based on vehicle kinematics model and the Model Predictive Control (MPC), which aims to minimize the tracking error and control increment. Combining macro-level and micro-level analysis, we verify the effectiveness of the proposed algorithm. Our results demonstrate that our proposed algorithm could greatly reduce the impact of LC maneuver on traffic flow. This is reflected in the decrease of total loss for nearby vehicles (such as discomfort and speed reduction), and the increase of traffic speed and throughput within the LC area. In addition, in order to guide the practical application of our algorithm, we employ the HighD dataset to validate the algorithm. This research could also be regarded as a preliminary foundational work to develop locally-optimal automatic LC algorithm. We anticipate that this research could provide valuable insights into autonomous driving technology.