On-Line Learning for Planning and Control of Underactuated Robots With Uncertain Dynamics

2022-01-01

IEEE Robotics and Automation Letters

AI Summary

Underactuated robots—such as the Pendubot—exhibit severe dynamic model uncertainty, hindering reliable trajectory planning and high-precision tracking. Method: This paper proposes a co-design framework integrating online learning with motion planning and control. It combines real-time disturbance estimation, optimization-based trajectory planning, and partial feedback linearization of actuated degrees of freedom to enable concurrent model adaptation and iterative controller refinement. Contribution/Results: We introduce the first “planning–control joint online learning” mechanism, achieving dynamically feasible trajectory generation and precise tracking within minimal iterations. Comprehensive simulations and hardware experiments demonstrate rapid convergence and robust performance despite large modeling errors—significantly enhancing system robustness and adaptability to unmodeled dynamics and disturbances.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

We present an iterative approach for planning and controlling motions of underactuated robots with uncertain dynamics. At its core, there is a learning process which estimates the perturbations induced by the model uncertainty on the active and passive degrees of freedom. The generic iteration of the algorithm makes use of the learned data in both the planning phase, which is based on optimization, and the control phase, where partial feedback linearization of the active dofs is performed on the model updated on-line. The performance of the proposed approach is shown by comparative simulations and experiments on a Pendubot executing various types of swing-up maneuvers. Very few iterations are typically needed to generate dynamically feasible trajectories and the tracking control that guarantees their accurate execution, even in the presence of large model uncertainties.

Problem

• Online Learning
• Robot Control
• Model Uncertainty

Innovation

• Online Learning
• Model Uncertainty
• Robust Control