MISC: Minimal Intervention Shared Control with Guaranteed Safety under Non-Convex Constraints

2025-07-03

AI Summary

Existing shared control approaches struggle to simultaneously ensure safety, feasibility, and minimal deviation from user intent—particularly under unpredictable user inputs and non-convex constraints—lacking rigorous theoretical guarantees. This paper proposes a minimal-intervention shared control framework that integrates control barrier functions with invariant set theory. It precomputes a control-invariant set offline and solves a constrained optimal control problem online, thereby guaranteeing strict satisfaction of non-convex constraints and enabling real-time decision-making. Crucially, the framework minimizes intervention in user commands while preserving safety and intent adherence. In a user study with 66 participants, the system significantly improved task performance, perceived controllability, trust, and cognitive load—without compromising safety or fidelity to user intent. This work establishes the first provably safe, real-time shared control architecture for non-convex constraint domains.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Economics, Online Markets and Human Computation: The sharing economy
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

Shared control combines human intention with autonomous decision-making, from low-level safety overrides to high-level task guidance, enabling systems that adapt to users while ensuring safety and performance. This enhances task effectiveness and user experience across domains such as assistive robotics, teleoperation, and autonomous driving. However, existing shared control methods, based on e.g. Model Predictive Control, Control Barrier Functions, or learning-based control, struggle with feasibility, scalability, or safety guarantees, particularly since the user input is unpredictable. To address these challenges, we propose an assistive controller framework based on Constrained Optimal Control Problem that incorporates an offline-computed Control Invariant Set, enabling online computation of control actions that ensure feasibility, strict constraint satisfaction, and minimal override of user intent. Moreover, the framework can accommodate structured class of non-convex constraints, which are common in real-world scenarios. We validate the approach through a large-scale user study with 66 participants--one of the most extensive in shared control research--using a computer game environment to assess task load, trust, and perceived control, in addition to performance. The results show consistent improvements across all these aspects without compromising safety and user intent.

Problem

• Ensures safety and feasibility in shared control systems
• Handles non-convex constraints in real-world applications
• Minimizes user intent override while maintaining performance

Innovation

• Constrained Optimal Control Problem framework
• Offline-computed Control Invariant Set
• Accommodates non-convex constraints