UGV Path Planning based on an Improved Continuous Ant Colony Optimisation Algorithm
Jing Liu, Aya Hussein, Sreenatha Gopalarao Anavatti, Matthew Garratt, Hussein A. Abbass · 2021 IEEE Symposium Series on Computational Intelligence (SSCI) · 2021
Path planning has always been an essential component of autonomy for Unmanned Ground Vehicles. In this paper, we present an improved continuous ant colony optimisation algorithm with differential evolution operator and local search, namely LIACODER, to solve the path planning problem with improved accuracy. A transformed coordinate system is introduced, based on which a solution repair method is presented to accelerate the convergence speed and facilitate the search process for the most feasible and optimal path. Experiments are conducted in both abstract and physical environments to compare LIACODERto classical path planning algorithms such as A*, D* and other state-of-the-art swarm optimisation algorithms. The superior performance of LIACODERin terms of finding feasible path with lower cost are validated.