[robotics-worldwide] [news] Online Course "Underactuated Robotics" free on MITx starting Oct 1

Russ Tedrake russt at mit.edu
Thu Sep 11 04:04:43 PDT 2014

Dear robotics-worldwide,

I am putting (most of) my graduate/advanced undergrad robotics class online this fall as a 10 week online course starting October 1. You can find more information below and at http://tiny.cc/mitx-underactuated .  The course is free and will have video lectures, online course notes, online problem sets, discussion forums, and an accompanying software distribution (including a semester-long license for MATLAB).   You can register to view all of the content even if you do not have time for the problem sets.  

Let me know if you have any questions.


- Russ

Underactuated Robotics: Algorithms for Walking, Running, Swimming, Flying, and Manipulation

== About this Course ==

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 introduces nonlinear dynamics and control of underactuated mechanical systems, with an emphasis on computational methods. Topics include the nonlinear dynamics of robotic manipulators, applied optimal and robust control and motion planning. Discussions include examples from biology and applications to legged locomotion, compliant manipulation, underwater robots, and flying machines.

== Prerequisites ==

Basic linear algebra and differential equations.

== Syllabus ==

Lecture 1	 Why study robot dynamics?
Motivation case studies; Fully- vs under- actuated systems (vs Non-holonomic); Robotics preliminaries	

Lecture 2	 Nonlinear dynamics
Fixed points, regions of attraction; Graphical analysis; Simple pendula, oscillators	

Lecture 3	 Dynamic programming
Control as an optimization; Dynamic programming (discrete-state discrete-action discrete-time); Finite-horizon and infinite-horizon problems; Approximation for continuous systems	

Lecture 4	 Dynamic Programming II
DP as a Partial Differential Equation (Hamilton-Bellman-Jacobi); Better numerical approximations for continuous systems (basic function approximation); Performance and scaling	

Lecture 5	 Acrobots, Cart-Poles, and Quadrotors
Canonical underactuated systems; Linearization; (Local) controllability analysis; Linear Quadratic Regulator (LQR) design	

Lecture 6	 Acrobots, Cart-Poles, and Quadrotors
Partial feedback linearization; Energy shaping; Differential flatness (or stabilization in SE3)	

Lecture 7	 Algebraic methods for analysis and control design
Lyapunov functions; Regions of attraction; Computational methods	

Lecture 8	 Algebraic methods for analysis and control design
Computational methods (cont); Extensions to robustness analysis (via common Lyapunov functions) and control design	

Lecture 9	 Trajectory optimization
First and second order methods; Shooting vs direct methods	

Lecture 10	 Trajectory stabilization
Time-varying LQR; transverse coordinates; Model-predictive control; Acrobot/cart-pole swingup; quadrotor dodging obstacles; Analysis (finite-time invariance)	

Lecture 11	 Simple walking models
Rimless wheel, Compass gait, Kneed compass gait; Hybrid systems formulation; Limit cycles; Poincare analysis; Finding and stabilizing hybrid limit cycles; Trajectory optimization	

Lecture 12	 Simple running models
Comparative biomechanics; Spring-loaded inverted pendulum; Raibert's hoppers; Orbital stabilization; Region of attraction estimation	

Lecture 13	 Manipulation
Grasp optimization; Complementarity formulations of contact; Planning through contact; Control w/ contact (QP methods)	

Lecture 14	 Humanoid robots
Simple models: Linear inverted pendulum, Zero-moment point, Centroidal moment matrix; Footstep planning; Whole-body planning; Whole-body control	

Lecture 15	 Feasible motion planning
A-star; Rapidly-exploring random trees; Probabilistic roadmaps; Heuristics and scaling (Task-space formulations, etc)	

Lecture 16	 Global policies from local policies
Roadmaps; Feedback motion planning; LQR-Trees; Online feedback motion planning	

Lecture 17	 Systems w/ Stochasticity
Planning under uncertainty; Robust analysis and control; L2-gain optimization; Dissipation inequalities	

Lecture 18	 Fluid locomotion: Swimming and Flying
Extremely brief intro to fluid dynamics; Simple models of swimming and flying; Case study: fixed-wing perching; Flapping flight

More information about the robotics-worldwide mailing list