AA 203: Optimal and Learning-Based Control

Spring 2024

Course Description

Optimal control solution techniques for systems with known and unknown dynamics. Dynamic programming, Hamilton-Jacobi reachability, and direct and indirect methods for trajectory optimization. Introduction to model predictive control. Adaptive control, model-based and model-free reinforcement learning, and connections between modern reinforcement learning and fundamental optimal control ideas.

Course Assistants

Daniel Morton Matt Foutter

Meeting Times

Lectures are held on Mondays and Wednesdays from 1:30pm to 2:50pm in Huang Engineering Center 18.

Lecture recordings will be available on Canvas.

Recitations are held on Fridays from 9:00am to 10:30am for the first four weeks of the quarter. Recitations are hybrid sessions, held both in Durand 251 and on Zoom (link on Canvas). Each recitation will be recorded and available offline on Canvas.

Office Hours

Office hours will begin in the second week of the quarter

Marco Pavone's office hours are on Tuesdays 1:00pm to 2:00pm in Durand 261 and by appointment.

Daniele Gammelli's office hours are on Thursdays 4:00pm to 5:00pm in Durand 217 (project discussion).

Daniel Morton's office hours are on Fridays 3:00pm to 5:00pm in Durand 217.

Matt Foutter's hybrid office hours are on Mondays 3:00pm to 4:00pm and Thursdays 5:00pm to 6:00pm in Durand 251.

Syllabus

The class syllabus can be found here.

Schedule

Subject to change.

Week Topic Lecture Slides
1 Course overview; intro to nonlinear optimization
Optimization theory
Recitation: Automatic differentiation with JAX
Monday: HW0 (ungraded) out
Lecture 1
Lecture 2
Recitation 1 ; Code
2 Calculus of variations
Indirect methods for optimal control
Recitation: Convex optimization with CVXPY
Monday: HW1 released
Lecture 3
Lecture 4
Recitation 2 ; Code
3 Pontryagin's maximum principle and computational methods
Direct methods
Recitation: Regression models
Lecture 5 ; Code
Lecture 6 ; Code
Recitation 3
4 Dynamic programming, discrete LQR
Nonlinear LQR for tracking and trajectory generation
Recitation: Training neural networks with JAX
Monday: HW1 due, HW2 released
Lecture 7
Lecture 8 ; Code
Recitation 4
5 Stochastic DP, value iteration, policy iteration
Introduction to RL, learning settings
Friday: Project proposal due
Lecture 9
Lecture 10
6 HJB, HJI, and reachability analysis
MPC I: Introduction
Monday: HW2 due
Tuesday: HW3 released
Lecture 11 ; Code
Lecture 12
7 MPC II: Feasibility and stability
MPC III: Robustness
Friday: Project midterm report due
Lecture 13
Lecture 14
8 Intro to learning, sys ID, adaptive control
RL I: Model-free RL - Value-based methods
Monday: HW4 released
Tuesday: HW3 due
Lecture 15
Lecture 16
9 No lecture (Memorial Day)
RL II: Model-free RL - Policy optimization

Lecture 17
10 RL III: Model-based RL
RL IV: Model-based policy optimization and Conclusions
Monday: HW4 due
Wednesday: Project video presentation and final report due
Lecture 18
Lecture 19

Follow this link to access the course website for the previous edition of Optimal and Learning-Based Control.