Estimating Control Barriers from Offline Data
This addresses the challenge of ensuring safety in robot control with limited offline data, which is incremental but improves efficiency and practicality over existing methods.
The paper tackles the problem of learning control barrier functions for safe robot control by proposing a framework that uses a fixed, sparsely-labeled offline dataset, eliminating the need for extensive sampling or online interaction. It achieves state-of-the-art performance in dynamic obstacle avoidance, with statistically safer and less conservative maneuvers.
Learning-based methods for constructing control barrier functions (CBFs) are gaining popularity for ensuring safe robot control. A major limitation of existing methods is their reliance on extensive sampling over the state space or online system interaction in simulation. In this work we propose a novel framework for learning neural CBFs through a fixed, sparsely-labeled dataset collected prior to training. Our approach introduces new annotation techniques based on out-of-distribution analysis, enabling efficient knowledge propagation from the limited labeled data to the unlabeled data. We also eliminate the dependency on a high-performance expert controller, and allow multiple sub-optimal policies or even manual control during data collection. We evaluate the proposed method on real-world platforms. With limited amount of offline data, it achieves state-of-the-art performance for dynamic obstacle avoidance, demonstrating statistically safer and less conservative maneuvers compared to existing methods.