Learn method generates optimal robot poses much faster than before
LOInK: Learned Optimal Inverse Kinematics via Structured Neural Surrogate Models
Robotics
Summary
Inverse kinematics means figuring out how a robot should move its joints to reach a desired position. The authors developed LOInK, a method that learns to quickly find near-best robot joint settings by training on many examples. LOInK creates a special space where the best solutions are easy to find, speeding up the search process greatly. They tested this on simple robot arms, a climbing robot, and a soft actuator, showing it finds good answers much faster than traditional methods.
What this means in practice
- •For robotics engineers: Generate near-optimal robot joint configurations quickly to improve robot motion planning and control.
- •For automation system designers: Speed up configuration calculations for robots in complex environments like climbing or soft actuation.
Authors
Michael Somerfield, Damian Abood, Ruigang Wang, Ian R. Manchester
Abstract
We introduce Learned Optimal Inverse Kinematics (LOInK), a method to generate approximately optimal solutions to an inverse kinematics problem. When trained on data consisting of sampled configurations and associated task variables and a given cost function, LOInK learns a bi-Lipschitz invertible mapping from configuration space to a decoupled task/latent space, and moreover, the latent space is structured so as to place cost-minimizing solutions at the origin. This enables efficient sampling of cost-minimizing solutions via a network-inversion algorithm based on operator splitting. We demonstrate the proposed approach on three problems: an illustrative three degree-of-freedom manipulator problem; a quadrupedal climbing robot for which LOInK can generate near-optimal solutions on average 31 times faster and up to 100 times faster than a constrained optimization approach; and a simulated soft actuator as a purely data-driven example, in which LOInK can explicitly generate high-quality solutions, unlike existing generative approaches that require diverse sampling and evaluation of candidate solutions.