Underactuated Robotics free videos and free material uploaded by Massachusetts Institute of Technology Staff .
Fully- vs. under-actuated systems
Preliminaries
Nonlinear dynamics of the simple pendulum
Introduction to optimal control
Double-integrator examples
Double integrator (cont.)
Quadratic regulator (Hamilton-Jacobi-Bellman (HJB) sufficiency),
min-time control (Pontryagin)
Dynamic programming and value interation: grid world, double integrator,
and pendulum examples
Acrobot and cart-pole: controllability, partial feedback linearization
(PFL), and energy shaping
Acrobot and cart-pole (cont.)
Policy search: open-loop optimal control, direct methods, and indirect
methods
Policy search (cont.): trajectory stabilization, iterative linear
quadratic regulator (iLQR), differential dynamic programming (DDP)
Simple walking models: rimless wheel, compass gait, kneed compass gait
Feedback control for simple walking models
Simple running models: spring-loaded inverted pendulum (SLIP), Raibert
hoppers
Midterm
Motion planning: Dijkstra's, A-star
Randomized motion planning: rapidly-exploring randomized trees and
probabilistic road maps
Feedback motion planning: planning with funnels, linear quadratic
regulator (LQR) trees
Function approximation and system identification
Model systems with uncertainty: state distribution dynamics and state
estimation
Stochastic optimal control
Aircraft
Swimming and flapping flight
Randomized policy gradient
Randomized policy gradient (cont.)
Model-free value methods: temporal difference learning and Q-learning
Actor-critic methods
Robots today move far too conservatively, using control systems that attempt to maintain full control authority at all times. Humans and animals move much more aggressively by routinely executing motions which involve a loss of instantaneous control authority. Controlling nonlinear systems without complete control authority requires methods that can reason about and exploit the natural dynamics of our machines.
This course discusses nonlinear dynamics and control of underactuated mechanical systems, with an emphasis on machine learning methods. Topics include nonlinear dynamics of passive robots (walkers, swimmers, flyers), motion planning, partial feedback linearization, energy-shaping control, analytical optimal control, reinforcement learning/approximate optimal control, and the influence of mechanical design on control. Discussions include examples from biology and applications to legged locomotion, compliant manipulation, underwater robots, and flying machines.
- 0 Reviews
- 1 Students
- 179 Courses
Write a public review