These Maps Are Made For Walking: Real-Time Terrain Property Estimation for Mobile Robots

TL;DR

This paper introduces a real-time Bayesian semantic mapping framework that accurately estimates terrain properties for mobile robots using a single RGB-D camera, improving navigation safety.

Contribution

It presents a novel recursive Bayesian inference method for real-time terrain property estimation and semantic mapping with a ROS interface, outperforming existing methods in simulation and real-world tests.

Findings

Outperforms state-of-the-art semantic mapping methods in simulation.

Accurately predicts terrain properties in indoor and outdoor robot deployments.

Operates in real-time with a ROS interface for easy integration.

Abstract

The equations of motion governing mobile robots are dependent on terrain properties such as the coefficient of friction, and contact model parameters. Estimating these properties is thus essential for robotic navigation. Ideally any map estimating terrain properties should run in real time, mitigate sensor noise, and provide probability distributions of the aforementioned properties, thus enabling risk-mitigating navigation and planning. This paper addresses these needs and proposes a Bayesian inference framework for semantic mapping which recursively estimates both the terrain surface profile and a probability distribution for terrain properties using data from a single RGB-D camera. The proposed framework is evaluated in simulation against other semantic mapping methods and is shown to outperform these state-of-the-art methods in terms of correctly estimating simulated ground-truth terrain properties when evaluated using a precision-recall curve and the Kullback-Leibler divergence test. Additionally, the proposed method is deployed on a physical legged robotic platform in both indoor and outdoor environments, and we show our method correctly predicts terrain properties in both cases. The proposed framework runs in real-time and includes a ROS interface for easy integration.