Learning from Demonstration via Spatiotemporal Tubes for Unknown Euler-Lagrange Systems

2026-07-01

AI Summary

This work addresses the challenge of robustly learning tasks from demonstrations for unknown Euler–Lagrange systems without prior system identification. It proposes the STT-LfD framework, which uniquely models demonstrations as spatio-temporal tubes (STTs) that serve as safety constraints. By employing heteroscedastic Gaussian processes in a data-driven manner, the approach captures the time-varying precision requirements inherent in the demonstrated task. A closed-form feedback controller is then designed to track this intention envelope while respecting actuator limits. Integrating motion learning with control, the method preserves the temporal structure of demonstrations and operates without an explicit dynamics model. Experimental results on both mobile robots and a 7-degree-of-freedom manipulator demonstrate significant improvements over baseline methods in terms of disturbance robustness and computational efficiency.

Abstract

We present STT-LfD, a unified Learning from Demonstration (LfD) framework that integrates motion learning with control for unknown Euler-Lagrange systems. Unlike traditional decoupled approaches that track a fixed reference, the proposed method treats demonstrations as a data-driven safety specification. Using heteroscedastic Gaussian Processes, STT-LfD learns Spatiotemporal Tubes (STTs) as an intent envelope that capture time-varying precision requirements of a task. A closed-form feedback controller then enforces these learned constraints while respecting actuator limits, without requiring explicit system identification. The approach preserves the temporal structure of demonstrations, remains computationally efficient, and avoids explicit system identification. Hardware experiments on a mobile robot and a 7-DOF manipulator show that it outperforms baselines in robustness to disturbances and computational speed.

Problem

• Learning from Demonstration
• Euler-Lagrange systems
• Spatiotemporal Tubes
• Unknown dynamics
• Motion control

Innovation

• Spatiotemporal Tubes
• Learning from Demonstration
• Euler-Lagrange systems
• Heteroscedastic Gaussian Processes
• Feedback Control