A Black-Box Physics-Informed Estimator based on Gaussian Process Regression for Robot Inverse Dynamics Identification

Introduction The Robot Inverse Dynamics Dataset is a collection of trajectories and joint torque measurements of two robotic manipulators, a 7 DoF Franka Emika Panda, and a 6 DOF MELFA RV4FL. Additionally, the dataset contains the inverse dynamical models and other useful quantities learned to reproduce the results reported on our reference paper "A Black-Box Physics-Informed Estimator based on Gaussian Process Regression for Robot Inverse Dynamics Identification". The proposed model relies on a novel multidimensional kernel, called Lagrangian Inspired Polynomial (LIP) kernel. At a Glance The size of the unzipped dataset is ~700MB. The dataset contains collections of joint trajectories and joint torque measurements of two robot manipulators: a 7 DoF Franka Emika Panda, and a 6 DOF MELFA RV4FL. models of the inverse dynamics of the two manipulators learned on the datasets The main directories are Simulated_PANDA/ contains the trajectories, models and results obtained on different configurations of a Franka Emika PANDA robot, simulated in sympybotics. Robots/ contains the data, models and results obtained on two real robots, a Franka Emika PANDA and a Mitsubishi Electric MELFA RV4FRL See the README.md file for a detailed description of the directories. Other Resources Python code to train the models and reproduce the results in the paper are available here. Citation If you use the Robot Inverse Dynamics dataset in your research, please cite our contribution: @InProceedings{ title={A Black-Box Physics-Informed Estimator based on Gaussian Process Regression for Robot Inverse Dynamics Identification}, author={Giacomuzzo, G., Dalla Libera, A., Romeres, D.,}, booktitle={IEEE Transaction on Robotics}, year={2024} } License The Robot Inverse Dynamics dataset is released under CC-BY-SA-4.0 license. All data: Created by Mitsubishi Electric Research Laboratories (MERL), 2024 SPDX-License-Identifier: CC-BY-SA-4.0