Efficient Differentiable Simulation of Articulated Bodies

TL;DR

This paper introduces a fast, memory-efficient method for differentiable simulation of articulated bodies, facilitating integration with deep learning and significantly accelerating optimization and reinforcement learning tasks.

Contribution

The authors develop a novel, efficient approach for differentiable simulation of articulated bodies using spatial algebra and the adjoint method, outperforming autodiff tools in speed and memory usage.

Findings

Simulation is an order of magnitude faster than autodiff tools.

Memory requirements are reduced by two orders of magnitude.

Gradient-based optimization accelerates convergence in control and inverse problems.

Abstract

We present a method for efficient differentiable simulation of articulated bodies. This enables integration of articulated body dynamics into deep learning frameworks, and gradient-based optimization of neural networks that operate on articulated bodies. We derive the gradients of the forward dynamics using spatial algebra and the adjoint method. Our approach is an order of magnitude faster than autodiff tools. By only saving the initial states throughout the simulation process, our method reduces memory requirements by two orders of magnitude. We demonstrate the utility of efficient differentiable dynamics for articulated bodies in a variety of applications. We show that reinforcement learning with articulated systems can be accelerated using gradients provided by our method. In applications to control and inverse problems, gradient-based optimization enabled by our work accelerates convergence by more than an order of magnitude.