ValueNetQP: Learned one-step optimal control for legged locomotion

TL;DR

ValueNetQP introduces a learned approach to predict value function derivatives, enabling real-time optimal control for legged robots with complex constraints, demonstrated on a quadruped robot in dynamic motions.

Contribution

The paper presents a novel method to learn value function gradients and Hessians for fast, constrained optimal control in legged locomotion, bridging the gap between accuracy and real-time performance.

Findings

Enables real-time control with quadratic programming.

Successfully applied to a quadruped robot performing dynamic gaits.

Handles complex constraints like friction cones effectively.

Abstract

Optimal control is a successful approach to generate motions for complex robots, in particular for legged locomotion. However, these techniques are often too slow to run in real time for model predictive control or one needs to drastically simplify the dynamics model. In this work, we present a method to learn to predict the gradient and hessian of the problem value function, enabling fast resolution of the predictive control problem with a one-step quadratic program. In addition, our method is able to satisfy constraints like friction cones and unilateral constraints, which are important for high dynamics locomotion tasks. We demonstrate the capability of our method in simulation and on a real quadruped robot performing trotting and bounding motions.