Fig. 2
From: Robot movement planning for obstacle avoidance using reinforcement learning
Illustration of perpendicular compensation in APF direction. When repulsive and attractive forces nearly oppose each other, a perpendicular correction is introduced, steering the robot away from local minima.
