Lecture
This lecture introduces the concept of layered control architectures, starting from motor control in vertebrates and extending to robots. It covers the Subsumption architecture, modern behavior architectures, and cognitive control. The instructor explains the deliberative control approach, exemplified by obstacle avoidance, involving analyzing camera images, detecting objects, planning new motion trajectories, and sending motor commands to actuators. The lecture discusses the drawbacks of the deliberative approach, such as dividing tasks into smaller sub-tasks leading to irrelevant sub-problems, lack of robustness and scalability, and difficulty in adding new sub-systems.