Metaheuristic Optimization in Path Planning of Autonomous Vehicles under the ATOM Framework

Abstract

Autonomous transportation systems are embarking our lives at an increasing pace. Over the past few years, several commercially available vehicles are incorporated with increasing levels of autonomy. The Autonomous Transportation Operating Modules (ATOM) framework is proposed to organize and coordinate the development as well as testing of these autonomous systems. One of the most important modules is the path planning of the vehicle, and finding the optimal path between two points is of great importance as it is directly to power saving of the battery. Metaheuristic optimization techniques are widely used to solve complex problems in an acceptable time interval. In this study, three metaheuristic approaches; simulated annealing, particle swarm and ant colony optimization are investigated to find the optimal path between two points in a static environment. The results of the PSO outperformed the other two, however both PSO and ACO can be used in path planning after enhancing the computational requirements of ACO to make it faster. These results open the door for investigating the implementation of these approaches on the embedded level for further demonstration and testing on real experimental platforms.