Optimizing autonomous path tracking using population-based metaheuristic and pure pursuit control

Abstract

Autonomous driving encompasses a comprehensive system comprising perception, decision-making, planning, and movement control subsystems. Within this framework, path tracking stands as a pivotal task, involving the precise and secure guidance of a vehicle along a predefined trajectory. Traditional path tracking methodologies often hinge on rule-based or parameter-tuning approaches, which may lack adaptability in intricate and dynamic scenarios. In this study, we present an approach employing a population-based metaheuristic for optimizing autonomous path tracking via a pure pursuit controller, notable for its streamlined control. The experimental assessments serve to elucidate the efficacy of the used algorithm, albeit within the confines of a single illustrative case, accompanied by relevant performance metrics.