Learning of Parameters in Behavior Trees for Movement Skills
This addresses the adoption barrier of RL in industrial robotics by providing a safer, more interpretable, and efficient learning method, though it is incremental as it builds on existing BT frameworks.
The paper tackles the problem of slow convergence, unsafe exploration, and lack of interpretability in reinforcement learning for robotics by learning parameters of Behavior Trees in simulation, achieving generalization to a physical robot without additional training and outperforming baselines in a peg-in-hole task with obstacle avoidance.
Reinforcement Learning (RL) is a powerful mathematical framework that allows robots to learn complex skills by trial-and-error. Despite numerous successes in many applications, RL algorithms still require thousands of trials to converge to high-performing policies, can produce dangerous behaviors while learning, and the optimized policies (usually modeled as neural networks) give almost zero explanation when they fail to perform the task. For these reasons, the adoption of RL in industrial settings is not common. Behavior Trees (BTs), on the other hand, can provide a policy representation that a) supports modular and composable skills, b) allows for easy interpretation of the robot actions, and c) provides an advantageous low-dimensional parameter space. In this paper, we present a novel algorithm that can learn the parameters of a BT policy in simulation and then generalize to the physical robot without any additional training. We leverage a physical simulator with a digital twin of our workstation, and optimize the relevant parameters with a black-box optimizer. We showcase the efficacy of our method with a 7-DOF KUKA-iiwa manipulator in a task that includes obstacle avoidance and a contact-rich insertion (peg-in-hole), in which our method outperforms the baselines.