Gradient-Based Trajectory Optimization With Learned Dynamics

TL;DR

This paper demonstrates that neural network-based learned dynamics models can enable effective gradient-based trajectory optimization for complex, highly dynamic robotic systems without relying on precise analytical models.

Contribution

The authors introduce a method combining learned neural network dynamics with gradient-based trajectory optimization for robots, effective over large time horizons with minimal data.

Findings

Neural networks accurately model complex nonlinear dynamics.

The approach works with only 25 minutes of data per robot.

Successful hardware experiments on Boston Dynamics Spot and RC car.

Abstract

Trajectory optimization methods have achieved an exceptional level of performance on real-world robots in recent years. These methods heavily rely on accurate analytical models of the dynamics, yet some aspects of the physical world can only be captured to a limited extent. An alternative approach is to leverage machine learning techniques to learn a differentiable dynamics model of the system from data. In this work, we use trajectory optimization and model learning for performing highly dynamic and complex tasks with robotic systems in absence of accurate analytical models of the dynamics. We show that a neural network can model highly nonlinear behaviors accurately for large time horizons, from data collected in only 25 minutes of interactions on two distinct robots: (i) the Boston Dynamics Spot and an (ii) RC car. Furthermore, we use the gradients of the neural network to perform gradient-based trajectory optimization. In our hardware experiments, we demonstrate that our learned model can represent complex dynamics for both the Spot and Radio-controlled (RC) car, and gives good performance in combination with trajectory optimization methods.