BeBOP -- Combining Reactive Planning and Bayesian Optimization to Solve Robotic Manipulation Tasks

TL;DR

BeBOP is a novel method that integrates reactive planning with Bayesian optimization to efficiently learn robotic manipulation behaviors, outperforming existing reinforcement learning methods in speed and robustness.

Contribution

The paper introduces BeBOP, a new approach combining reactive planning and Bayesian optimization for behavior tree generation in robotic manipulation tasks.

Findings

BeBOP outperforms state-of-the-art reinforcement learning algorithms by up to 46 times in speed.

The method is less dependent on reward shaping, simplifying the learning process.

A modified uncertainty estimate for random forest models significantly improves performance.

Abstract

Robotic systems for manipulation tasks are increasingly expected to be easy to configure for new tasks. While in the past, robot programs were often written statically and tuned manually, the current, faster transition times call for robust, modular and interpretable solutions that also allow a robotic system to learn how to perform a task. We propose the method Behavior-based Bayesian Optimization and Planning (BeBOP) that combines two approaches for generating behavior trees: we build the structure using a reactive planner and learn specific parameters with Bayesian optimization. The method is evaluated on a set of robotic manipulation benchmarks and is shown to outperform state-of-the-art reinforcement learning algorithms by being up to 46 times faster while simultaneously being less dependent on reward shaping. We also propose a modification to the uncertainty estimate for the random forest surrogate models that drastically improves the results.