Safe Learning for Contact-Rich Robot Tasks: A Survey from Classical Learning-Based Methods to Safe Foundation Models

2025-12-10

AI Summary

This study addresses safety assurance throughout the entire learning lifecycle—exploration and execution—in contact-intensive robotic manipulation, particularly under high uncertainty, strong dynamic coupling, and high-risk physical interactions. Methodologically, it systematically reviews two safety paradigms—safe exploration and safe execution—and integrates key techniques including constrained reinforcement learning, control barrier functions, model-predictive safety filtering, and uncertainty-aware modeling. It proposes, for the first time, a unified safety-learning taxonomy tailored to contact-rich tasks. Innovatively, it introduces language-level safety constraint specification and multimodal safety signal alignment as novel foundation-model adaptation paradigms, revealing the dual nature of vision-language-action (VLA) models: simultaneous safety gains and emergent risks. The project delivers the first comprehensive safety-learning landscape for contact-intensive tasks, releases the open-source knowledge repository *Awesome-Learning4Safe-Contact-rich-tasks*, and distills six fundamental challenges alongside five actionable technical pathways.

Technology Category

• Natural Language Processing: Safety and Robustness
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Security and Privacy: Security and privacy of machine learning and AI applications
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Contact-rich tasks pose significant challenges for robotic systems due to inherent uncertainty, complex dynamics, and the high risk of damage during interaction. Recent advances in learning-based control have shown great potential in enabling robots to acquire and generalize complex manipulation skills in such environments, but ensuring safety, both during exploration and execution, remains a critical bottleneck for reliable real-world deployment. This survey provides a comprehensive overview of safe learning-based methods for robot contact-rich tasks. We categorize existing approaches into two main domains: safe exploration and safe execution. We review key techniques, including constrained reinforcement learning, risk-sensitive optimization, uncertainty-aware modeling, control barrier functions, and model predictive safety shields, and highlight how these methods incorporate prior knowledge, task structure, and online adaptation to balance safety and efficiency. A particular emphasis of this survey is on how these safe learning principles extend to and interact with emerging robotic foundation models, especially vision-language models (VLMs) and vision-language-action models (VLAs), which unify perception, language, and control for contact-rich manipulation. We discuss both the new safety opportunities enabled by VLM/VLA-based methods, such as language-level specification of constraints and multimodal grounding of safety signals, and the amplified risks and evaluation challenges they introduce. Finally, we outline current limitations and promising future directions toward deploying reliable, safety-aligned, and foundation-model-enabled robots in complex contact-rich environments. More details and materials are available at our href{ https://github.com/jack-sherman01/Awesome-Learning4Safe-Contact-rich-tasks}{Project GitHub Repository}.

Problem

• Ensuring robot safety during exploration and execution in contact-rich tasks
• Balancing safety and efficiency in learning-based control for complex manipulation
• Addressing safety challenges when integrating foundation models with robotic systems

Innovation

• Survey categorizes safe learning methods into exploration and execution domains
• Reviews techniques like constrained RL and control barrier functions for safety
• Extends safety principles to vision-language foundation models for manipulation