Improved Ant Colony Optimization for robot navigation

Abstract

This paper involves the design and development of an improved Ant Colony Optimization algorithm for the purpose of robot navigation. The algorithm is able to calculate optimum paths for the robot to travel on to perform the tasks of goal-seeking, wall-following and obstacle avoiding, where the efficiency of a path is determined based on the length of the path. This algorithm is based upon earlier research done by Mehtap Kose and improves on his work by simplifying the algorithm equations, expanding the size of the simulation environment, increasing the task capabilities of the robot, as well as testing the algorithm in real time on an autonomous mobile robot. The paper also enhances the usefulness of the ACO algorithm by designing and creating a user friendly ACO graphical user interface and also conducts further research on the workings of the algorithm by conducting systematic testing and simulations. The successful completion of this paper proves the feasibility of employing the concepts of Ant Colony Optimization in robot navigation to solve real world problems.