The exploration of other planets is an important domain for mobile robotic systems. On Earth, mobile robots can help for search and rescue tasks, by quickly delivering images and maps for assessing critical situations without putting humans into danger. Common in both domains is the need to move and navigate the systems in unknown, rough environments without relying on external infrastructure. In space exploration, some level of local autonomy is required for coping with high communication latencies. In disaster scenarios, local autonomy relieves rescue workers from low level control. Furthermore, using several cooperating systems supports efficiency and fast response.
We are working on multicopers, rovers and crawlers for these scenarios. All systems use stereo vision and Inertial Measurement Units (IMUs) as sensors, create 2.5D or 3D maps and perform on-board path planning and navigation. Thus, our flying as well as ground based systems are capable of autonomous navigation to user defined target points, through unknown, rough terrain and without using GPS.
Ongoing work includes the cooperation of multiple autonomous robots. The aims are to complement skills in a heterogeneous team (e.g. driving and flying) and to increase robustness and reduce mission time using a homogeneous team. In any case, each mobile system is able to fulfill its job on its own, without depending on other systems, a central station, communication link, etc.
Sensors are crucial in robotics. The sensors developed at the RMC are taylored to the rising demands of complex robotic applications. Different tasks like exploration, autonomous 3D-modelling or navigation need versatile and robust visual perception. Therefore the multisensory 3D-Modeller (DLR 3DMo) with stereo cameras, laser scanner and laser stripe profiler as well as later the novel laser scanner VR-SCAN were developed.