Augmenting Reinforcement Learning with Behavior Primitives for Diverse Manipulation Tasks
This work addresses the exploration burden in deep reinforcement learning for robotic manipulation, offering an incremental improvement by integrating existing primitives into a hierarchical framework.
The paper tackles the challenge of long-horizon manipulation tasks in robotics by augmenting reinforcement learning with pre-defined behavior primitives, resulting in MAPLE outperforming baseline approaches significantly in simulated tasks and enabling policy transfer to new variants and physical hardware.
Realistic manipulation tasks require a robot to interact with an environment with a prolonged sequence of motor actions. While deep reinforcement learning methods have recently emerged as a promising paradigm for automating manipulation behaviors, they usually fall short in long-horizon tasks due to the exploration burden. This work introduces Manipulation Primitive-augmented reinforcement Learning (MAPLE), a learning framework that augments standard reinforcement learning algorithms with a pre-defined library of behavior primitives. These behavior primitives are robust functional modules specialized in achieving manipulation goals, such as grasping and pushing. To use these heterogeneous primitives, we develop a hierarchical policy that involves the primitives and instantiates their executions with input parameters. We demonstrate that MAPLE outperforms baseline approaches by a significant margin on a suite of simulated manipulation tasks. We also quantify the compositional structure of the learned behaviors and highlight our method's ability to transfer policies to new task variants and to physical hardware. Videos and code are available at https://ut-austin-rpl.github.io/maple