Media Summary: Underactuated explores dynamic programming and value iteration by applying these concepts to the pendulum control problem. The lecture examines the relationships between cost-to-go functions, feedback policies, and discretization techniques in lower-dimensional systems compared to reinforcement learning methods. Um like right now if you want them or anytime but we're going to actually use them again uh throughout this Underactuated explores the transition from discrete to continuous dynamic programming in optimal control. Through the lens of the Hamilton-Jacobi-Bellman equation, the lecture examines how these mathematical frameworks provide tools for designing optimal feedback controllers, using the double integrator as a primary case study for understanding cost-to-go functions and policy optimization.
6 8210 Spring 2023 Lecture - Detailed Analysis & Overview
Underactuated explores dynamic programming and value iteration by applying these concepts to the pendulum control problem. The lecture examines the relationships between cost-to-go functions, feedback policies, and discretization techniques in lower-dimensional systems compared to reinforcement learning methods. Um like right now if you want them or anytime but we're going to actually use them again uh throughout this Underactuated explores the transition from discrete to continuous dynamic programming in optimal control. Through the lens of the Hamilton-Jacobi-Bellman equation, the lecture examines how these mathematical frameworks provide tools for designing optimal feedback controllers, using the double integrator as a primary case study for understanding cost-to-go functions and policy optimization. ... equations the trampoline is represented as a 6.8210 Spring 2023 Final Project Presentation
