An Efficient Lateral Manoeuvre Planner for Automated Vehicles

Harikirshnan Vijayakumar,Dezong Zhao,Jianglin Lan,Jingjing Jiang,Wenjing Zhao,Daxin Tian,Yuanjian Zhang,Dachuan Li

DOI: https://doi.org/10.2139/ssrn.4161344

2022-01-01

SSRN Electronic Journal

Abstract:In automated driving, lateral manoeuvre planning plays a crucial role as it is responsible for generating a feasible path for lane change or obstacle avoidance. Moreover, the lateral path should be generated without delay and be comfortable for the passengers. The work proposes an efficient planner module for lateral manoeuvres in automated driving, considering the objectives including computational efficiency and passenger comfort. The work proposes a novel method of generating a lower order parametric curve for the lateral movements in automated driving. Additionally, the work proposes a replanning trajectory which guides the vehicle back to the original lane if the lateral movement is blocked by the surrounding vehicle. The planner module is completed by including a computationally efficient velocity profile according to the manoeuvres performed. The generated path and velocity from the planner are tracked using a nonlinear model predictive controller due to its ability to handle nonlinearity in the prediction model, constraints, and cost function. Finally, the work proposes using a separate planner and controller module for a forced lane change scenario by considering separate objectives of lane changing and lane keeping. The simulation results show the proposed method’s computational efficiency and curvature compatibility to a higher order curve. The advantages of the proposed method to other similar curves in the literature include shorter path length and better control over the curve when the same comfort parameters are considered.