Achieving Safe Control Online through Integration of Harmonic Control Lyapunov-Barrier Functions with Unsafe Object-Centric Action Policies

2025-11-18

AI Summary

This work addresses the lack of formal safety guarantees in robotic control policies. We propose a Signal Temporal Logic (STL)-driven online safety enhancement framework. Methodologically, we introduce the first integration of Harmonic Control Lyapunov–Barrier Functions (HCLBFs) with object-centric action policies, enabling real-time filtering of arbitrary base policies (e.g., reinforcement learning policies) via a safety certification mechanism to generate control commands that simultaneously achieve task performance and provably correct collision avoidance. Our contributions are twofold: (1) a systematic mapping from STL-specifiable safety constraints to HCLBFs, ensuring formal verifiability; and (2) co-optimization of safety assurance and policy behavior fidelity. In physical experiments on force-controlled obstacle avoidance with a fixed-base manipulator, the framework successfully transforms an initially unsafe policy into one satisfying STL-defined hard safety constraints, markedly improving robustness and formal verifiability.

Technology Category

• Natural Language Processing: Safety and Robustness
• Intelligent Robots: Behavior Learning & Control
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We propose a method for combining Harmonic Control Lyapunov-Barrier Functions (HCLBFs) derived from Signal Temporal Logic (STL) specifications with any given robot policy to turn an unsafe policy into a safe one with formal guarantees. The two components are combined via HCLBF-derived safety certificates, thus producing commands that preserve both safety and task-driven behavior. We demonstrate with a simple proof-of-concept implementation for an object-centric force-based policy trained through reinforcement learning for a movement task of a stationary robot arm that is able to avoid colliding with obstacles on a table top after combining the policy with the safety constraints. The proposed method can be generalized to more complex specifications and dynamic task settings.

Problem

• Transforming unsafe robot policies into safe ones with formal guarantees
• Preserving both safety constraints and task-driven behavior simultaneously
• Preventing robot collisions with obstacles while executing movement tasks

Innovation

• Integrates harmonic control Lyapunov-barrier functions with robot policies
• Uses safety certificates to enforce formal safety guarantees
• Modifies unsafe policies to maintain safety and task performance