Planning along Differentiable Charts of Constraint Manifolds with the Inverse Function Theorem

@inproceedings{cohn2026planning, title={Planning along Differentiable Charts of Constraint Manifolds with the Inverse Function Theorem}, author={Cohn, Thomas and Shaw, Seiji and Roy, Nicholas and Tedrake, Russ}, booktitle={ICRA Workshop on Geometry in the Age of Data-Driven Robotics}, year={2026} }

Abstract

When planning motions with kinematic equality constraints, analytic inverse kinematics can be used to parameterize the set of feasible configurations. We present a new approach for computing gradients through such parameterizations (a requirement for trajectory optimization), that extends compatibility beyond handwritten analytic IK solutions to general-purpose implementations, including automated symbolic solvers like IKFast.

Poster