A Comparison of Path Planning Algorithms for Robotic Vacuum Cleaners

Abstract: Household robotics is on the rise with robotic vacuum cleaners taking the lead. One of the challenges in creating great robotic vacuum cleaners is path planning for coverage of the room. For instance, the robot might not know how the environment looks, or where in the room the robot is. Three path planning algorithms are tested: the random path algorithm, the snaking algorithm and the spiral algorithm. In our simulations the random algorithm was initially faster than the snaking algorithm. However, the random algorithm's performance worsened and it performed similar to the snaking algorithm in the end. The spiral algorithm performed worse than the other tested algorithms. The random algorithm showed a more stable performance than both the snaking and spiral algorithms. However, we believe that the spiral algorithm could be complemented with other algorithms. Errors in the results could possibly come from the method used in the software to detect wall collisions. Future research could improve the software and also cover collision with objects in rooms and rooms of other shapes. We concluded that the random algorithm performed best in our simulations and questioned the practical use of more advanced algorithms.