Trajectory planner based on fifth-order polynomials applied for lane changing in autonomous driving

Abstract

The control of autonomous cars is currently a topic of significant interest, primarily owing to their inherent advantages. These cars have the potential to enhance safety, decrease fuel consumption, thereby reducing travel costs, improve driving comfort, and alleviate traffic congestion. The control of car dynamics remains an open subject, with numerous researchers and companies striving to enhance existing strategies or develop novel ones. The developed solutions should aim to assist the driver while simultaneously ensuring safer driving conditions. Therefore, this paper proposes a solution for controlling the lateral of a car using model predictive control (MPC) algorithm includes decision making and trajectory planning. The initial step involves modeling the dynamics by constructing a nonlinear model based on the laws of physics, specifically considering the lateral aspects of a vehicle. Secondly, a trajectory planner is designed to ensure a safe path in the presence of the front car on the road, incorporating safety constraints. The host car is capable of executing lane shifts, overtaking, or braking maneuvers to avoid collisions with the front car. Thirdly, the trajectory generated by the planner is tracked using a Model Predictive Control (MPC) algorithm. The effectiveness of the proposed algorithm has been assessed through simulation.