models, controllers and objectives for missions is usually intricate. One way to make the design more manageable is through the use of Machine Learning techniques. This talk will deal with two issues related to autonomous systems: the problem of identification of dynamic models for robots; and the problem of finding controllers for the stabilization of such robots when their parameters are unknown or uncertain. Recent results using Learning Based Model Predictive Control, Learning Automata, Reinforcement Learning and Deep Learning will be presented.

Speaker(s): Sidney Givigi,

Location:
1515 St. Catherine W., EV12.163
Montreal, Quebec
H3G 2W1