Lynkeus is a project granted by the Federal Ministry of Education and Research (16SV2307). The goal of this project is to demonstrate the capability of a new sensor technology. The sensor is a photonic-mixture device (PMD) and is based on time-of-flight principle. Two main fields of robotics research within Lynkeus are human-robot cooperation (HRC) and bin picking (BP). Therefore a demonstrator was built up succesfully in our laboratory that demonstrates the capabilities of our efficient algorithms for both applications.
The main focus of research within HRC is to analyse human motions for markerless tracking and collision avoidance by evaluation of depth image sequences obtained from PMD sensor. Therefore several algortihms have been developed that allow analyzing human motion and mapping the results to a kinematic chain of a human model in real-time. A PMD sensor which is mounted on the ceiling of our robot cell is used for this task. Another challenging task in Lynkeus is high accuracy segmentation of the human body based on efficient mathematical models.
A classical field of research in robotics is bin picking. Combining PMD technology with novel algorithms for scene analysis and pattern recognition, it was demonstrated that we could achieve better solutions than using other existing methods based on classical sensor technologies. In combination with HRC a robot cell was built up and two frameworks for solving the described tasks have been implemented. The communication within the framework is based on a state-of-the-art communication protocol called XIRP.