剛体力学の自動微分と連続感度解析
Automatic Differentiation and Continuous Sensitivity Analysis of Rigid Body Dynamics
剛体力学の微分可能な物理シミュレータを提案し、パラメータ推定手法の比較と、経験からシミュレーションパラメータを推定しつつコスト最小化を達成する閉ループモデル予測制御アルゴリズムを導入した。
著者: David Millard, Eric Heiden, Shubham Agrawal, Gaurav S. Sukhatme
分類: cs.RO, cs.LG, cs.SY, eess.SY
原文アブストラクト
A key ingredient to achieving intelligent behavior is physical understanding that equips robots with the ability to reason about the effects of their actions in a dynamic environment. Several methods have been proposed to learn dynamics models from data that inform model-based control algorithms. While such learning-based approaches can model locally observed behaviors, they fail to generalize to more complex dynamics and under long time horizons. In this work, we introduce a differentiable physics simulator for rigid body dynamics. Leveraging various techniques for differential equation integration and gradient calculation, we compare different methods for parameter estimation that allow us to infer the simulation parameters that are relevant to estimation and control of physical systems. In the context of trajectory optimization, we introduce a closed-loop model-predictive control algorithm that infers the simulation parameters through experience while achieving cost-minimizing performance.
