A dynamic lane-changing trajectory planning model for automated vehicles. (October 2018)

• Record Type: Journal Article
• Authors: Yang, Da; Zheng, Shiyu; Wen, Cheng; Jin, Peter J.; Ran, Bin
• Is Part Of: Transportation research. Volume 95(2018)
• Page Start: 228
• Page End: 247
• Publication Date: 2018-10
• Subjects: Lane-changing -- Automated vehicles -- Trajectory planning -- Reaction time; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2018.06.007
• Languages: English
• ISSNs: 0968-090X

Abstract

Highlights: A dynamic lane-changing trajectory planning model is proposed. Lane-changing vehicles can adapt to the real-time change of surrounding traffic. The lane-changing trajectory curve is a time-independent cubic polynomial function. Lane-changing can be finished smoothly, efficiently and safely using the new model. The planned velocity and trajectory can be well-tracked by the automated vehicles. Abstract: This paper focuses on the lane-changing trajectory planning (LTP) process in the automatic driving technologies. Existing studies on the LTP algorithms are primarily the static planning method in which the states of the surrounding vehicles of a lane-changing vehicle are assumed to keep unchanged in the whole lane-changing process. However, in real-world traffic, the velocities of the surrounding vehicles change dynamically, and the lane-changing vehicle needs to adjust its velocity and positions correspondingly in real-time to maintain safety. To address such limitations, the dynamic lane-changing trajectory planning (DLTP) model is proposed in the limited literature. This paper proposes a novel DLTP model consisting of the lane-changing starting-point determination module, trajectory decision module and trajectory generation module. The model adopts a time-independent polynomial trajectory curve to avoid the unrealistic assumptions on lane-changing velocities and accelerations in the existing DLTP model. Moreover, a rollover-avoidance algorithm and a collision-avoidance algorithm containing a reaction time are presented to guarantee the lane-changing safety of automated vehicles, even in an emergent braking situation. The field lane-changing data from NGSIM data are used to construct a real traffic environment for lane-changing vehicles and verify the effectiveness of the proposed model, and CarSim is applied to investigate the traceability of the planned lane-changing trajectories using the proposed model. The results indicate that an automated vehicle can complete the lane-changing process smoothly, efficiently and safely following the trajectory planned by the proposed model, and the planned velocity and trajectory can be well-tracked by automated vehicles.