wanxinjin

Pontryagin-Differentiable-Programming

A unified end-to-end learning and control framework that is able to learn a (neural) control objective function, dynamics equation, control policy, or/and optimal trajectory in a control system.

Safe-PDP

Safe Pontryagin Differentiable Programming (Safe PDP) is a new theoretical and algorithmic safe differentiable framework to solve a broad class of safety-critical learning and control tasks.

Learning-from-Sparse-Demonstrations

A framework and method to jointly learn a (neural) control objective function and a time-warping function only from sparse demonstrations or waypoints.

Task-Driven-Hybrid-Reduction

We developed a task-driven hybrid model reduction method for solving dexterous manipulation with 5 minutes of online learning.

IOC-from-Incomplete-Trajectory-Observations

Inverse optimal control from incomplete trajectory observations, proposing the concept of the recovery matrix which provides further insights into objective learning process.

Learning-from-Directional-Corrections

A new method for a robot to learn a control objective from human user's directional corrections.

Learning-LCS

A new learning formulation/method to learn/identify a piecewise linear system, also named as linear complementarity system.

wanxinjin.github.io

asu-robotics

ASU MAE 547: Modeling and Control of Robots

ioc

Inverse optimal control codebase