Optimization of a mobile robot's path tracking

The optimal path tracking is essential for the efficient motion of a mobile robot in its workspace. That's why several optimization methods have used such as the ant colony, the simulated annealing and the genetic algorithms. In our work we have focused on the evolutionary method which is the genetic algorithms, to solve the path-planning problem. This method was enhanced by the use of the artificial potential field in order to reach the goals correctly. In dead the workspace of the studied mobile robot is a free space, equipped by artificial beacons. Some of them mark a set of surveyed sites. The site that hasn't visited yet represents a goal to reach. Whereas the mobile robot must be away form these were visited. The hybridization of the genetic algorithms by the artificial potential field approach has successfully carried out. However, we have noticed that the hybrid genetic algorithm path planning run in the previous work, generated better path in ulterior generation than in the final one. In this paper, we aim to improve this hybrid algorithm by using a back-tracking on the previous generation.