Adaptive Backstepping of Synergistic Hybrid Feedbacks with Application to Obstacle Avoidance