Teaching Robot Motion Planning

Robot motion planning is a fairly intuitive and engaging topic, yet it is difficult to teach. The material is taught in undergraduate and graduate robotics classes in computer science, electrical engineering, mechanical engineering and aeronautical engineering, but at an abstract level. Deep learning could be achieved by having students implement and test different motion planning strategies. However, a full implementation of motion planning algorithms by undergraduates is practically impossible in the context of a single class, even by students proficient in programming. By helping undergraduates grasp motion planning concepts in series of courses designed for increasing advanced levels, we can open the field to young and enthusiastic talent. This cannot be done by asking students to implement motion planning algorithms from scratch or access thousands of lines of code and just figure out how things work. We present an ongoing project to develop microworld software and a modeling curriculum that supports undergraduate acquisition of motion planning knowledge and tool use by computer science and engineering students.