Physically-Consistent Parameter Identification of Robots in Contact

2024-09-15

arXiv.org

AI Summary

This work addresses the challenge of identifying robot inertial parameters without contact force sensing. We propose a physically consistent identification method leveraging only joint current/torque measurements. By projecting the full-body dynamics model onto the nullspace of contact constraints, we formulate a linear matrix inequality (LMI) optimization problem explicitly embedding physical consistency—enabling inertial parameter identification solely from standard joint sensors, even during dynamic contact interactions. This approach eliminates reliance on external contact force measurements, significantly improving sample efficiency and generalization across diverse gaits. We validate the method experimentally on the Boston Dynamics Spot quadruped robot across multiple locomotion patterns. Results demonstrate superior parameter accuracy and task generalization compared to purely data-driven approaches such as deep neural networks (DNNs), confirming its efficacy for real-world complex contact tasks.

Technology Category

• Intelligent Robots: State Estimation
• Search and Optimization: Sampling/Simulation-based Search
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate inertial parameter identification is crucial for the simulation and control of robots encountering intermittent contact with the environment. Classically, robots' inertial parameters are obtained from CAD models that are not precise (and sometimes not available, e.g., Spot from Boston Dynamics), hence requiring identification. To do that, existing methods require access to contact force measurement, a modality not present in modern quadruped and humanoid robots. This paper presents an alternative technique that utilizes joint current/torque measurements -- a standard sensing modality in modern robots -- to identify inertial parameters without requiring direct contact force measurements. By projecting the whole-body dynamics into the null space of contact constraints, we eliminate the dependency on contact forces and reformulate the identification problem as a linear matrix inequality that can handle physical and geometrical constraints. We compare our proposed method against a common black-box identification mrethod using a deep neural network and show that incorporating physical consistency significantly improves the sample efficiency and generalizability of the model. Finally, we validate our method on the Spot quadruped robot across various locomotion tasks, showcasing its accuracy and generalizability in real-world scenarios over different gaits.

Problem

• Identifies robot inertial parameters without contact force measurements.
• Uses joint current/torque measurements for parameter identification.
• Improves sample efficiency and generalizability with physical consistency.

Innovation

• Uses joint current/torque measurements for parameter identification
• Projects dynamics into null space of contact constraints
• Reformulates identification as linear matrix inequality