SELFI: Autonomous Self-Improvement with Reinforcement Learning for Social Navigation

TL;DR

SELFI is an online reinforcement learning method that enhances pre-trained control policies for social navigation, enabling robots to adapt quickly to real-world environments with minimal human intervention.

Contribution

The paper introduces SELFI, a novel hybrid online learning approach combining model-based pre-training with model-free reinforcement learning for improved robotic social navigation.

Findings

Enhanced collision avoidance in real-world tests

More socially compliant robot behaviors

Rapid policy adaptation with minimal human input

Abstract

Autonomous self-improving robots that interact and improve with experience are key to the real-world deployment of robotic systems. In this paper, we propose an online learning method, SELFI, that leverages online robot experience to rapidly fine-tune pre-trained control policies efficiently. SELFI applies online model-free reinforcement learning on top of offline model-based learning to bring out the best parts of both learning paradigms. Specifically, SELFI stabilizes the online learning process by incorporating the same model-based learning objective from offline pre-training into the Q-values learned with online model-free reinforcement learning. We evaluate SELFI in multiple real-world environments and report improvements in terms of collision avoidance, as well as more socially compliant behavior, measured by a human user study. SELFI enables us to quickly learn useful robotic behaviors with less human interventions such as pre-emptive behavior for the pedestrians, collision avoidance for small and transparent objects, and avoiding travel on uneven floor surfaces. We provide supplementary videos to demonstrate the performance of our fine-tuned policy on our project page.