The workgroup Cognitronics and Sensor Systems develops the multi-sensory tracking system CITrack which is used in various projects. The system uses fiducial marker based camera tracking as well as RGB-D based body and Ultra Wideband tracking and can combine them in position and orientation in the laboratory. CITrack enables experiments to be tracked accurately, to provide the measured values at runtime, to visualize them and to perform evaluations.​

CITrack is a laboratory area (6 m x 6 m) which is covered with multiple RGB, grayscale, and RGB-D cameras. All recordings are made available in the local network via the Robot Operating System (ROS) middleware. All detections of the individual sensor systems are synchronized in time, fused and transferred into a common coordinate system. With particular focus on the robustification of fiducial marker recognition, the hypothesis generation is taken into account in particular to be as accurate as possible per camera. Furthermore, the calibration and registration of the multi-sensory system demands for a well designed architecture and system setup. In addition to the software-based approach, real-time and resource-efficient FPGA solutions for marker tracking are also being explored.


Dipl.-Ing. Daniel Wolf