Safe Active Navigation and Exploration for Planetary Environments Using Proprioceptive Measurements

TL;DR

This paper introduces SAEGT, a framework enabling legged robots to safely explore unknown granular terrains using proprioceptive sensing, especially where visual data is unreliable, by estimating safe regions online for navigation.

Contribution

SAEGT is the first framework to use proprioceptive sensing and Gaussian Process regression for real-time safe exploration of granular terrains in robotic navigation.

Findings

SAEGT successfully navigates towards goals in simulation.

The framework accurately estimates safe regions using proprioception.

SAEGT outperforms traditional visual-based methods in deformable terrains.

Abstract

Legged robots can sense terrain through force interactions during locomotion, offering more reliable traversability estimates than remote sensing and serving as scouts for guiding wheeled rovers in challenging environments. However, even legged scouts face challenges when traversing highly deformable or unstable terrain. We present Safe Active Exploration for Granular Terrain (SAEGT), a navigation framework that enables legged robots to safely explore unknown granular environments using proprioceptive sensing, particularly where visual input fails to capture terrain deformability. SAEGT estimates the safe region and frontier region online from leg-terrain interactions using Gaussian Process regression for traversability assessment, with a reactive controller for real-time safe exploration and navigation. SAEGT demonstrated its ability to safely explore and navigate toward a specified goal using only proprioceptively estimated traversability in simulation.