Self-Augmented Robot Trajectory: Efficient Imitation Learning via Safe Self-augmentation with Demonstrator-annotated Precision

2025-09-11

AI Summary

In imitation learning, large-scale demonstration data and unsafe random exploration—particularly hazardous in confined-space tasks like peg-in-hole assembly—pose significant challenges. To address these issues, this paper proposes the Safety-Aware Robust Trajectory (SART) framework, enabling safe autonomous data augmentation from a single human demonstration. SART leverages human-annotated key waypoints and their associated spherical accuracy bounds, integrating collision-avoidance trajectory generation with trajectory rerouting to expand training data diversity while strictly preserving physical safety constraints. Experiments in both simulation and real-robot settings demonstrate that SART substantially improves policy success rates and data efficiency compared to pure human demonstration baselines. Notably, it reduces human intervention frequency by orders of magnitude and establishes, for the first time, a one-shot imitation learning paradigm that simultaneously achieves high safety, strong generalization, and sample efficiency.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Intelligent Robots: Manipulation

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Imitation learning is a promising paradigm for training robot agents; however, standard approaches typically require substantial data acquisition -- via numerous demonstrations or random exploration -- to ensure reliable performance. Although exploration reduces human effort, it lacks safety guarantees and often results in frequent collisions -- particularly in clearance-limited tasks (e.g., peg-in-hole) -- thereby, necessitating manual environmental resets and imposing additional human burden. This study proposes Self-Augmented Robot Trajectory (SART), a framework that enables policy learning from a single human demonstration, while safely expanding the dataset through autonomous augmentation. SART consists of two stages: (1) human teaching only once, where a single demonstration is provided and precision boundaries -- represented as spheres around key waypoints -- are annotated, followed by one environment reset; (2) robot self-augmentation, where the robot generates diverse, collision-free trajectories within these boundaries and reconnects to the original demonstration. This design improves the data collection efficiency by minimizing human effort while ensuring safety. Extensive evaluations in simulation and real-world manipulation tasks show that SART achieves substantially higher success rates than policies trained solely on human-collected demonstrations. Video results available at https://sites.google.com/view/sart-il .

Problem

• Reduces imitation learning data requirements
• Ensures safety in autonomous robot exploration
• Minimizes human effort in robot training

Innovation

• Self-augmentation with annotated precision boundaries
• Single demonstration learning with autonomous expansion
• Collision-free trajectory generation within safety spheres