Heterogeneous Ground and Air Platforms, Homogeneous Sensing: Team CSIRO Data61’s Approach to the DARPA Subterranean Challenge

Abstract

Heterogeneous teams of robots, leveraging a balance between autonomy and human interaction, bring powerful capabilities to the problem of exploring dangerous, unstructured subterranean environments. Here we describe the solution developed by Team CSIRO Data61, consisting of CSIRO, Emesent, and Georgia Tech, during the DARPA Subterranean Challenge. These presented systems were fielded in the Tunnel Circuit in August 2019, the Urban Circuit in February 2020, and in our own Cave event, conducted in September 2020. A unique capability of the fielded team is the homogeneous sensing of the platforms utilized, which is used to obtain a decentralized multi-agent SLAM solution on each platform (both ground agents and UAVs) using peer-to-peer communications. This approach enabled a shift in focus from constructing a pervasive communications network to relying on multi-agent autonomy, motivated by experiences in early circuit events. These experiences also showed the surprising capability of rugged tracked platforms for challenging terrain, which in turn led to the heterogeneous team structure based on a BIA5 OzBot Titan ground robot and an Emesent Hovermap UAV, supplemented by smaller tracked or legged ground robots. The ground agents use a common CatPack perception module, which allowed reuse of the perception and autonomy stack across all ground agents with minimal adaptation.

OHM: GPU Based Occupancy Map Generation

“…All data was collected with a Velodyne VLP-16 Puck lidar (either flat or spinning on a 30 • inclined pedestal), with dual return mode enabled. Sensor trajectory information is provided from the Wildcat SLAM pipeline described in [27]. This sensor produces around 300,000 rays/s.…”

“…The OHM implementation presented in this paper has been used widely in UGV navigation applications on wheeled, tracked, quadruped and hexapod platforms [27], as well as UAV navigation and planning and offline point cloud processing and filtering.…”

Wildcat: Online Continuous-Time 3D Lidar-Inertial SLAM

“…In multi-agent collaborative SLAM scenarios, each agent synchronises its database of submaps (containing submaps generated by itself and others) with other agents (i.e., other robots or the base station) within its communication range via peer-to-peer communication. We refer the reader to [6] for additional information about our ROS-based data sharing system, Mule. The maximum size of each submap with a lidar range of 100 m is about 500 KB, whereas the average submap size in underground SubT events was about 100-170 KB.…”

“…The maximum size of each submap with a lidar range of 100 m is about 500 KB, whereas the average submap size in underground SubT events was about 100-170 KB. [6]. Therefore, Wildcat can easily share submaps between the agents with a modest communication bandwidth.…”

“…• We present Wildcat, a highly robust and versatile state-ofthe-art lidar-inertial SLAM system. This paper provides a detailed technical description of Wildcat beyond the broad non-technical overview previously presented in [6].…”