Ground vehicle approaching negative obstacles

ROS pointcloud and odometry streams recorded during various scenarios where a robotic ground vehicle approaches negative obstacles. These values were generated by the Wildcat SLAM software based on measurements by a Velodyne lidar and an IMU. Often negative obstacles cannot be observed directly by a ground vehicle as they are obscured by the upper surface between the vehicle and the obstacle. This data contains examples that can be used to demonstrate methods of indirectly inferring negative obstacles from typical observations made during an approach. Screenshot image previews of the data can be found in the `screenshots` directory. ROS topics are described in the `README.md` file. Models of the simulation worlds have been provided as Blender files in the `worlds` directory.\nLineage: This data was collected by one of the robots built by the CSIRO Data61 DARPA SubT Challenge team. Some of this data was recorded while the robot was moving autonomously and some of it was recorded while the robot was being teleoperated by a human.