Learning from Demonstration with Failure Awareness for Safe Robot Navigation

2026-04-25

AI Summary

Traditional imitation learning struggles to ensure robotic safety in unseen environments due to its inability to model failure states. This work proposes a failure-aware offline reinforcement learning framework that innovatively decouples the roles of successful and failed demonstrations: successful trajectories are used to learn the task policy, while failure experiences—such as collisions—are explicitly leveraged to shape value estimates in hazardous regions, thereby constructing safety boundaries without interfering with policy optimization. Evaluated in both simulation and real-world settings, the method significantly reduces collision rates while maintaining high task success rates and demonstrates strong generalization across diverse environments and robotic platforms.

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Intelligent Robots: Learning & Optimization for ROB
• Natural Language Processing: Safety and Robustness

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Learning from demonstration is widely used for robot navigation, yet it suffers from a fundamental limitation: demonstrations consist predominantly of successful behaviors and provide limited coverage of unsafe states. This limitation leads to poor safety when the robot encounters scenarios beyond the demonstration distribution. Failure experiences, such as collisions, contain essential information about unsafe regions, but remain underutilized. The key difficulty lies in the fact that failure data do not provide valid guidance for action imitation, and their naive incorporation into policy learning often degrades performance. We address this challenge by proposing a failure-aware learning framework that explicitly decouples the roles of success and failure data. In this framework, failure experiences are used to shape value estimation in hazardous regions, while policy learning is restricted to successful demonstrations. This separation enables the effective use of failure data without corrupting policy behavior. We implement this design within an offline reinforcement learning (RL) setting and evaluate it in both simulation and real-world environments. The results show that our framework consistently reduces collision rates while preserving the task success rate, and demonstrate strong generalization across different environments and robot platforms.

Problem

• Learning from Demonstration
• Robot Navigation
• Failure Awareness
• Safety
• Offline Reinforcement Learning

Innovation

• failure-aware learning
• learning from demonstration
• offline reinforcement learning
• safe robot navigation
• value estimation