Dynamic Robot Path Planning Using Improved Max-Min Ant Colony Optimization

This paper presents a method of using an improved version of the Max-Min Ant Colony Optimization (ACO) algorithm for use in dynamic global robot path planning. A modified Bug2 algorithm was used to determine the initial best path on a map. After running the Max-Min ACO algorithm to find the optimal path, the path was checked to see if it crossed any previously unknown obstacles, and if so the route was recalculated from the obstacle to the goal. The previously found best path was saved to help subsequent runs find the optimal solution faster. This algorithm was tested using simulations, and it was determined that it performed well in finding the average shortest path. It also resulted in greatly reduced processing times.