Welcome

This is a set of helpful components to get started with the DARPA SubT Virtual Challenge

What is it

1. A curated set of open source packages which might be useful for the Challenge
   • [Cartographer](https://github.com/googlecartographer): 2D and 3D SLAM across multiple platforms and sensor configurations
   • [Octomap](https://octomap.github.io): an efficient probabilistic 3D mapping framework based on octrees
   • [tensorflow_object_detector](https://github.com/osrf/tensorflow_object_detector.git): OSRF's tensorflow-based object detector. The subt_extended Docker container should be configured to be able to run this on GPU, if the hardware is available.
   • [Darknet](https://github.com/leggedrobotics/darknet_ros): You only look once (YOLO) is a state-of-the-art, real-time object detection system. Capable of being run on either CPU or GPU.
2. Some simple components to aid development
   • [Wall follow]: A robot navigation controller which tries to center itself in corridors or follow the nearest wall in larger spaces. Uses progressive move base goals and the move base controller for interacting with the vehicle.
   • [Frontier exploration](http://wiki.ros.org/explore_lite): An exploration method that attempts to traverse to and eliminate the boundary between known and unknown space.
3. Some launch files to demonstrate how everything goes together in helloworld_subt_launch
   • helloworld.launch : Start wall follow, cartographer, octomap. Assumes robot has been spawned in a valid area of the Ignition simulation with a 3D LIDAR. (i.e. ign launch -v 4 virtual_stix.ign robotName1:=alpha robotConfig1:=X1_SENSOR_CONFIG_4)

Docker Installation

Native Installation

Example Use-Cases

Example CloudSim Submission