Interpretable Robotic Friction Learning via Symbolic Regression

2025-05-19

AI Summary

Robot joint friction modeling has long suffered from a trade-off among accuracy, interpretability, and generalizability: physics-based models rely heavily on expert knowledge and lack adaptability, whereas neural networks offer flexibility but suffer from poor robustness and limited interpretability. This paper introduces symbolic regression (SR) to robot friction modeling for the first time, synergistically integrating physical priors with data-driven learning to automatically discover interpretable, closed-form analytical expressions from real-world torque–velocity–load measurements of a KUKA LWR-IV+ manipulator. The resulting model achieves complexity comparable to classical models (e.g., LuGre or Dahl), yet delivers significantly improved prediction accuracy. Crucially, it natively supports extensions to load-dependent behavior and coupled dynamic effects—addressing longstanding limitations in both interpretability and adaptability of conventional friction models.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Neuro-Symbolic Learning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Accurately modeling the friction torque in robotic joints has long been challenging due to the request for a robust mathematical description. Traditional model-based approaches are often labor-intensive, requiring extensive experiments and expert knowledge, and they are difficult to adapt to new scenarios and dependencies. On the other hand, data-driven methods based on neural networks are easier to implement but often lack robustness, interpretability, and trustworthiness--key considerations for robotic hardware and safety-critical applications such as human-robot interaction. To address the limitations of both approaches, we propose the use of symbolic regression (SR) to estimate the friction torque. SR generates interpretable symbolic formulas similar to those produced by model-based methods while being flexible to accommodate various dynamic effects and dependencies. In this work, we apply SR algorithms to approximate the friction torque using collected data from a KUKA LWR-IV+ robot. Our results show that SR not only yields formulas with comparable complexity to model-based approaches but also achieves higher accuracy. Moreover, SR-derived formulas can be seamlessly extended to include load dependencies and other dynamic factors.

Problem

• Challenges in modeling robotic joint friction accurately
• Limitations of traditional and data-driven friction modeling methods
• Proposing symbolic regression for interpretable, adaptable friction estimation

Innovation

• Symbolic regression for friction torque estimation
• Generates interpretable formulas from robot data
• Extends formulas to include dynamic factors