Data-driven Interpretable Hybrid Robot Dynamics

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Mower, Christopher E., Zong, Rui, Bou-Ammar, Haitham
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908708299079680
author Mower, Christopher E.
Zong, Rui
Bou-Ammar, Haitham
author_facet Mower, Christopher E.
Zong, Rui
Bou-Ammar, Haitham
contents We study data-driven identification of interpretable hybrid robot dynamics, where an analytical rigid-body dynamics model is complemented by a learned residual torque term. Using symbolic regression and sparse identification of nonlinear dynamics (SINDy), we recover compact closed-form expressions for this residual from joint-space data. In simulation on a 7-DoF Franka arm with known dynamics, these interpretable models accurately recover inertial, Coriolis, gravity, and viscous effects with very small relative error and outperform neural-network baselines in both accuracy and generalization. On real data from a 7-DoF WAM arm, symbolic-regression residuals generalize substantially better than SINDy and neural networks, which tend to overfit, and suggest candidate new closed-form formulations that extend the nominal dynamics model for this robot. Overall, the results indicate that interpretable residual dynamics models provide compact, accurate, and physically meaningful alternatives to black-box function approximators for torque prediction.
format Preprint
id arxiv_https___arxiv_org_abs_2512_11900
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Data-driven Interpretable Hybrid Robot Dynamics
Mower, Christopher E.
Zong, Rui
Bou-Ammar, Haitham
Robotics
We study data-driven identification of interpretable hybrid robot dynamics, where an analytical rigid-body dynamics model is complemented by a learned residual torque term. Using symbolic regression and sparse identification of nonlinear dynamics (SINDy), we recover compact closed-form expressions for this residual from joint-space data. In simulation on a 7-DoF Franka arm with known dynamics, these interpretable models accurately recover inertial, Coriolis, gravity, and viscous effects with very small relative error and outperform neural-network baselines in both accuracy and generalization. On real data from a 7-DoF WAM arm, symbolic-regression residuals generalize substantially better than SINDy and neural networks, which tend to overfit, and suggest candidate new closed-form formulations that extend the nominal dynamics model for this robot. Overall, the results indicate that interpretable residual dynamics models provide compact, accurate, and physically meaningful alternatives to black-box function approximators for torque prediction.
title Data-driven Interpretable Hybrid Robot Dynamics
topic Robotics
url https://arxiv.org/abs/2512.11900