Viability-Preserving Passive Torque Control

2025-10-03

AI Summary

To address the vulnerability of conventional passive torque controllers to safety constraint violations under external disturbances, this paper proposes a passivity-preserving torque control method grounded in viability theory. The approach integrates data-driven learning with analytical modeling to construct a joint feasible set encompassing self-collision, environment collision, and joint physical limits—transformed equivalently into hard constraints on joint accelerations. Crucially, viability theory is incorporated into the passive control framework for the first time, ensuring strict passivity while guaranteeing that system states remain within the safety set for infinite time horizons. The method combines an accurate robot dynamics model with real-time quadratic programming optimization. Experimental validation on a Franka Emika 7-DOF manipulator demonstrates a 32% increase in control frequency, significantly improved trajectory smoothness, and provably guaranteed long-term safety.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Natural Language Processing: Safety and Robustness
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

Conventional passivity-based torque controllers for manipulators are typically unconstrained, which can lead to safety violations under external perturbations. In this paper, we employ viability theory to pre-compute safe sets in the state-space of joint positions and velocities. These viable sets, constructed via data-driven and analytical methods for self-collision avoidance, external object collision avoidance and joint-position and joint-velocity limits, provide constraints on joint accelerations and thus joint torques via the robot dynamics. A quadratic programming-based control framework enforces these constraints on a passive controller tracking a dynamical system, ensuring the robot states remain within the safe set in an infinite time horizon. We validate the proposed approach through simulations and hardware experiments on a 7-DoF Franka Emika manipulator. In comparison to a baseline constrained passive controller, our method operates at higher control-loop rates and yields smoother trajectories.

Problem

• Ensuring robot safety under external perturbations using viability theory
• Pre-computing safe state-space sets for collision and joint limits avoidance
• Enforcing viability constraints via quadratic programming on passive controllers

Innovation

• Viability theory pre-computes safe state-space sets
• Quadratic programming enforces constraints on passive control
• Data-driven viable sets ensure infinite-time horizon safety