Obstacle Avoidance with DS

This lecture covers obstacle avoidance using Dynamical Systems (DS) in the context of learning and adaptive control for robots. The instructor explains the main properties for modulating dynamics, ensuring the robot avoids obstacles while reaching its goal. Various concepts are discussed, such as the modulation matrix, assumptions, and desiderata for real-time obstacle avoidance. The lecture also delves into the challenges of avoiding spurious fixed points and the importance of stability guarantees using contraction theory. The presentation concludes with demonstrations and theoretical insights on avoiding concave obstacles with DS, emphasizing the significance of reference directions and metrics in ensuring stable and reactive obstacle avoidance.