Collaborative Exploration of Planetary Analog Terrain with Rovers and Stationary Sensors

Kurzfassung

When planetary rovers land on extraterrestrial surfaces, it is not taken for granted that the terrain is known a priori, nor is there a Global Navigation Satellite System GNSS available. This is an important issue if not one but more rovers collaborate on the same area. On the other hand, two or more rovers can explore a much wider area at the same time and achieve a more comprehensive understanding of the environment. And as long as the rovers operate close to the landing site, sensors on the lander can further enhance the output. This scenario has been put into practice at the German Aerospace Center DLR with two very different mobile robots, the Lightweight Rover Unit 1 LRU1 and the Scout Rover, and a stationary Sensor Pole acting as lander. During one of the measurement runs, the systems were expanded with an Ultra-Wide Band UWB localization network to provide a second external tracking of the mobile robots. Operating these systems together, but independently from each other, already results in a good map of the site more rapidly than would be done by just one system. Fusion of the individual measurements in a second step results in a better map of the site than would have been measured by just one system. Therefore, an important part of the setup, and of this publication, is the common network and data exchange between the systems. The Robot Operating System 2 ROS2 is the middleware for LRU1, Scout Rover and the sensor pole. The three systems further are in the same network that also has a Network Time Protocol NTP server for time synchronization. This permits online data fusion. How to further enhance the outcome in post processing is shown with the additional UWB distance measurement. An important advantage is that the systems need not know each other prior network setup as long as they run the same ROS version or provide bridges for compatibility over different versions. This makes the advantages of a heterogeneous team of robots (complementary sensors, different mobility capabilities, co-working, etc.) as easy to put into practice as a team of the same systems. The activity shows that collaborative exploration and Simultaneous Localization And Mapping SLAM of an unknown off-road terrain as expected, for example on the Moon, is possible with rovers and stationary sensors using relatively simple systems and setup and without excessively large computational power. The text also shows the pitfalls and limitations of the systems and gives hints on how to overcome them. This can be a good inspiration for future exploration of the Solar System with heterogeneous robotic systems, as is foreseeable with the current development in national and international Moon programs but not yet reality. The setup can likewise be applied in terrestrial situations where human beings would be set to risk, e.g. collapsed buildings or areas with toxic atmosphere, or where common solutions based on GNSS and a priori knowledge of the terrain are not possible.