Moving Obstacle Avoidance with Potential Function Using Estimated Velocity of Obstacle
Yusaku Koike, Koichi Hidaka · 2024
This paper proposes a moving obstacle avoidance for a two-wheel mobile robot using the estimated velocity of the obstacle. The robot avoids fixed obstacles such as shelves and desks in an indoor environment with walls and moves to a target point. The method is that an avoidance path is generated using an artificial potential method using estimated velocity, and the robot tracks the path. UKF estimates the Velocity of the obstacle. To validate the effectiveness of our proposed method, we conducted avoidance simulations in two distinct scenarios. The first scenario involved avoiding opposing obstacles, while the second scenario simulated obstacles moving across the robot’s path near the goal. These simulations provided valuable insights into the performance of our method under different conditions.