Development of Obstacle Avoidance Control System Using Modified Attractive Equations of Gaussian Potential Function

Navigation system is a necessary system for autonomous vehicle and many papers proposes the techniques for trajectory tracking and obstacle avoidance. Gaussian potential function is well-known for trajectory tracking control system with obstacle avoidance using 2D lidar. The disadvantage is that it relies on local minima, which is unable to deviate the moving path in some cases that vehicle and target is moving in the same direction, and obstacle is in between because the temptation of attractive potential field is too high. Then, the trajectory tracking control with obstacle avoidance using a modified attractive function has been introduced that conventional attractive function of gaussian potential function is modified in order to improve the navigation system. The simulation is performed by CARLA Simulator and the motion of the obstacle avoidance using the modified function is considered by the sway motion during the tracking and the safety during the lateral motion, which is considered by lateral acceleration of the vehicle, its value in simulation case is less than other types of attractive function. Meanwhile, the modified function also maintains the safety distance gap between the vehicle and the obstacle to not avoid at very close range from obstacle, which probably cause the collision.