Autonomous Lane Change System for Intelligent Vehicle Based on Model Curve Fitting

Abstract

In this paper, we present a novel lane change system for intelligent vehicles combined with the model curve fitting method to plan a safe-feasible path in high speed. According to the existing literature, the current virtual potential function (VPF) performs well in real-time collision avoidance. However, this method also exists a flaw that it is based on the point mass model to produce the path in which could be non-safe for intelligent vehicles to track. To improve this method to be more practical, we propose a novel model curve fitting (MCF) method to consider the dynamics of the ground vehicles that can plan a safer path for intelligent vehicles to track. We simulated the overall system framework in MATLAB/Simulink with vehicle dynamic lateral motion model. From the simulation results, we confirm that the MCF method combined with VPF is highly effective compared to the traditional VPF method.