Obstacles modeling and collision detection of space robots for performing on-orbit services

A space robotic system will play important role in on-orbital servicing missions, such as target capturing, ORU(Orbital Replacement Unit) replacement, payload transfer, and so on. To manipulate the space robot safely, we addressed feasible obstacles modeling and collision detection methods. In-depth investigation and analysis showed that most spacecraft are composed of modular and standardized devices, whose contour can be approximated by regular geometry shapes-sphere, cylinder, cuboid, cone, etc. Then, a super-quadric function was used to describe their spatial surfaces, establishing a unified frame to model the typical obstacles. On the other hand, we realized three effective collision detection methods to supply early warning before the danger comes. These methods are the line group method based on OSG(Open scene Graph), bounding box method based on OBB(Oriented Bounding Box) and cylinder method based on the dual vector. Finally, we developed a simulation system and verified the proposed methods on it.