Beyond Predefined Actions: Integrating Behavior Trees and Dynamic Movement Primitives for Robot Learning from Demonstration

TL;DR

This paper presents a novel method that integrates Behavior Trees and Dynamic Motion Primitives to enable robots to learn complex tasks from single demonstrations, improving interpretability, modularity, and adaptability without predefined actions.

Contribution

The authors propose a joint learning framework that combines BT structure and DMP actions from a single demonstration, removing the need for predefined low-level actions and enhancing policy flexibility.

Findings

Successfully learns robot policies from single demonstrations

Improves interpretability and modularity of robot control policies

Enables combining partial demonstrations into coherent behaviors

Abstract

Interpretable policy representations like Behavior Trees (BTs) and Dynamic Motion Primitives (DMPs) enable robot skill transfer from human demonstrations, but each faces limitations: BTs require expert-crafted low-level actions, while DMPs lack high-level task logic. We address these limitations by integrating DMP controllers into a BT framework, jointly learning the BT structure and DMP actions from single demonstrations, thereby removing the need for predefined actions. Additionally, by combining BT decision logic with DMP motion generation, our method enhances policy interpretability, modularity, and adaptability for autonomous systems. Our approach readily affords both learning to replicate low-level motions and combining partial demonstrations into a coherent and easy-to-modify overall policy.