When today's robots move around, the motion that you observe is the result of two software stages: planning and control. Planning is the part that computes a trajectory from the initial state of the system to some goal state. Control is the part that deals with perturbations or modelling errors, and stabilizes the system at best while it performs the trajectory output by the planner.
In this talk, we are going to explore the questions of planning and control for humanoid robots. We will see that straightforward formulations of the trajectory generation problem yield spaces of both high dimension and complex structure. We will then describe a number of solutions to "unwind" this structure into smaller sub-problems, which can be solved using a combination of stochastic and optimal-control algorithms.
|Slides (opens in new window/tab)|