A Subterranean Virtual Cave World for Gazebo based on the DARPA SubT Challenge
This work provides a simulation tool for researchers and developers working on robotic systems for subterranean exploration, which is incremental as it builds on existing platforms like Gazebo.
The authors created an open-source virtual cave world for Gazebo to address the challenge of developing autonomous robots for subterranean environments, providing a test-bed for evaluating algorithms in perception, networking, and mobility.
Subterranean environments with lots of obstacles, including narrow passages, large voids, rock falls and absence of illumination were always challenging for control, navigation, and perception of mobile robots. The limited availability and access to such environments restricts the development pace of capabilities for robotic platforms to autonomously accomplish tasks in such challenging areas. The Subterranean Challenge is a competition focusing on bringing robotic exploration a step closer to real life applications for man-made underground tunnels, urban areas and natural cave networks, envisioning advanced assistance tools for first responders and disaster relief agencies. The challenge offers a software-based virtual part to showcase technologies in autonomy perception, networking and mobility for such areas. Thus, the presented open-source virtual world aims to become a test-bed for evaluating the developed algorithms and software and to foster mobile robotics developments.