Adaptive Manipulation using Behavior Trees

2024-06-20

arXiv.org

AI Summary

Robots struggle with real-time adaptation to unforeseen changes and lack experience reuse in dynamic manipulation tasks relying on non-visual feedback (e.g., valve tightening), where force/tactile sensing is critical. Method: This paper proposes an adaptive Behavior Tree (BT) architecture that unifies online policy selection, multimodal (vision + force/tactile) observation fusion, and incremental policy performance prediction learning within a BT framework. It leverages BT extension modeling, online state estimation, and closed-loop feedback-driven policy optimization to ensure safe and robust autonomous adaptation. Contribution/Results: Evaluated on industrial benchmark tasks, the system achieves 100% task success rate, reduces completion time by up to 36%, and significantly improves operational safety and execution robustness—demonstrating the first integration of these capabilities in a BT-based robotic control architecture.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Many manipulation tasks pose a challenge since they depend on non-visual environmental information that can only be determined after sustained physical interaction has already begun. This is particularly relevant for effort-sensitive, dynamics-dependent tasks such as tightening a valve. To perform these tasks safely and reliably, robots must be able to quickly adapt in response to unexpected changes during task execution, and should also learn from past experience to better inform future decisions. Humans can intuitively respond and adapt their manipulation strategy to suit such problems, but representing and implementing such behaviors for robots remains a challenge. In this work we show how this can be achieved within the framework of behavior trees. We present the adaptive behavior tree, a scalable and generalizable behavior tree design that enables a robot to quickly adapt to and learn from both visual and non-visual observations during task execution, preempting task failure or switching to a different manipulation strategy. The adaptive behavior tree selects the manipulation strategy that is predicted to optimize task performance, and learns from past experience to improve these predictions for future attempts. We test our approach on a variety of tasks commonly found in industry; the adaptive behavior tree demonstrates safety, robustness (100% success rate) and efficiency in task completion (up to 36% task speedup from the baseline).

Problem

• Adapting to non-visual environmental changes during manipulation tasks
• Learning from past experiences to improve future task performance
• Ensuring safety and efficiency in dynamic, effort-sensitive robotic tasks

Innovation

• Adaptive behavior trees for dynamic task adaptation
• Learning from visual and non-visual observations
• Optimizing task performance through predictive strategy selection