Optimal path planning for waiter robot using modified ant colony optimization

The waiter robot is utilized to give a quick service for customer's satisfaction in the restaurant. In performing its task, the waiter robot facing problems such as many obstacles consist of tables, chairs, or customers along the path in the workspace. Therefore, it has capable of path planning before moving to the goal. A traditional algorithm such as A-star and Dijkstra algorithm can solve path planning for the environment like this easily, they produce the shortest path but not safely motion. Modified Ant Colony Optimization (M-ACO) algorithm is presented in this paper used to get optimal path planning for waiter robot in the restaurant. We evaluate the M-ACO algorithm for the waiter robot by optimizing some parameters to observe the performance of this approach in terms safely motion and path length. At the same time, we also examine A-star and Dijkstra algorithms in the same working environment, so we can compare the performance of each approach. The experiment results for case 1 show that M-ACO with global transfer factor (P0) value of 0.9 gives path length 79.5264 cm without crashing any obstacles while A-star and Dijkstra give path length 73.907 cm with crashing any obstacles, and another experiment result for case 2 show that M-ACO with P0 value 0.9 give path length 99.745 cm without crashing any obstacles while Astar and Dijkstra give path length 84.5617 cm with crashing any obstacles.