Learning Robotic Navigation from Experience: Principles, Methods, and Recent Results

TL;DR

This paper reviews how machine learning enables robots to learn navigation from experience, surpassing traditional geometric methods by reasoning about real-world traversability and improving with more data.

Contribution

It presents a unified toolkit for experiential learning in robotic navigation, summarizing recent approaches, principles, and experimental results, and discusses future research directions.

Findings

Experiential learning improves navigation robustness.

Data-driven methods outperform geometric planning in complex environments.

The toolkit unifies multiple recent approaches.

Abstract

Navigation is one of the most heavily studied problems in robotics, and is conventionally approached as a geometric mapping and planning problem. However, real-world navigation presents a complex set of physical challenges that defies simple geometric abstractions. Machine learning offers a promising way to go beyond geometry and conventional planning, allowing for navigational systems that make decisions based on actual prior experience. Such systems can reason about traversability in ways that go beyond geometry, accounting for the physical outcomes of their actions and exploiting patterns in real-world environments. They can also improve as more data is collected, potentially providing a powerful network effect. In this article, we present a general toolkit for experiential learning of robotic navigation skills that unifies several recent approaches, describe the underlying design principles, summarize experimental results from several of our recent papers, and discuss open problems and directions for future work.