ReLMoGen: Leveraging Motion Generation in Reinforcement Learning for Mobile Manipulation

TL;DR

ReLMoGen introduces a hierarchical reinforcement learning framework that leverages motion generation and subgoal prediction to efficiently solve complex, long-horizon mobile manipulation and navigation tasks in simulation, outperforming existing methods.

Contribution

The paper presents ReLMoGen, a novel framework combining learned subgoal prediction with motion planning, enabling RL to tackle complex tasks beyond traditional action spaces.

Findings

ReLMoGen outperforms state-of-the-art RL baselines on seven robotics tasks.

The method demonstrates strong transferability across different motion generators.

ReLMoGen effectively handles long-horizon, interactive navigation and manipulation tasks.

Abstract

Many Reinforcement Learning (RL) approaches use joint control signals (positions, velocities, torques) as action space for continuous control tasks. We propose to lift the action space to a higher level in the form of subgoals for a motion generator (a combination of motion planner and trajectory executor). We argue that, by lifting the action space and by leveraging sampling-based motion planners, we can efficiently use RL to solve complex, long-horizon tasks that could not be solved with existing RL methods in the original action space. We propose ReLMoGen -- a framework that combines a learned policy to predict subgoals and a motion generator to plan and execute the motion needed to reach these subgoals. To validate our method, we apply ReLMoGen to two types of tasks: 1) Interactive Navigation tasks, navigation problems where interactions with the environment are required to reach the destination, and 2) Mobile Manipulation tasks, manipulation tasks that require moving the robot base. These problems are challenging because they are usually long-horizon, hard to explore during training, and comprise alternating phases of navigation and interaction. Our method is benchmarked on a diverse set of seven robotics tasks in photo-realistic simulation environments. In all settings, ReLMoGen outperforms state-of-the-art Reinforcement Learning and Hierarchical Reinforcement Learning baselines. ReLMoGen also shows outstanding transferability between different motion generators at test time, indicating a great potential to transfer to real robots.