Autonomous Off-road Navigation over Extreme Terrains with Perceptually-challenging Conditions

TL;DR

This paper presents a resilient autonomous navigation framework for extreme, perceptually-challenging off-road environments, enabling real-time on-board decision-making without prior maps or communication, demonstrated through successful deployment in DARPA challenges.

Contribution

It introduces a multi-sensor resilient architecture and a fast algorithm for real-time traversability estimation in complex terrains, advancing autonomous off-road navigation capabilities.

Findings

Successfully deployed on diverse robots in challenging terrains

Achieved top placements in DARPA Subterranean Challenge

Demonstrated robustness in GPS-denied, perceptually-degraded environments

Abstract

We propose a framework for resilient autonomous navigation in perceptually challenging unknown environments with mobility-stressing elements such as uneven surfaces with rocks and boulders, steep slopes, negative obstacles like cliffs and holes, and narrow passages. Environments are GPS-denied and perceptually-degraded with variable lighting from dark to lit and obscurants (dust, fog, smoke). Lack of prior maps and degraded communication eliminates the possibility of prior or off-board computation or operator intervention. This necessitates real-time on-board computation using noisy sensor data. To address these challenges, we propose a resilient architecture that exploits redundancy and heterogeneity in sensing modalities. Further resilience is achieved by triggering recovery behaviors upon failure. We propose a fast settling algorithm to generate robust multi-fidelity traversability estimates in real-time. The proposed approach was deployed on multiple physical systems including skid-steer and tracked robots, a high-speed RC car and legged robots, as a part of Team CoSTAR's effort to the DARPA Subterranean Challenge, where the team won 2nd and 1st place in the Tunnel and Urban Circuits, respectively.