Experimental Evaluation of Robotic Grasping Strategies for On-Orbit Servicing

On-orbit servicing is a key enabler of space sustainability, allowing maintenance, refueling, repairs, and upgrades of satellites to extend their operational life. A typical mission involves rendezvous, relative motion synchronization, and final capture with a robotic arm. This work focuses on the capture phase, where safety and reliability requirements meet the challenges of autonomous navigation and control. The target is uncooperative and slowly tumbling, so residual relative motion remains at capture. To address this, a trajectory planning framework is proposed in free-flying mode, where arm and platform are jointly controlled. An enhanced Rapidly-exploring Random Tree (RRT) planner ensures safe maneuvers, and the guidance algorithm is validated on a high-fidelity model including full 3D manipulator dynamics. Experimental tests are performed on a granite-table facility with a frictionless satellite mock-up (representing the target) and a 6-DOF robotic arm moving on a linear rail (representing the chaser). The facility integrates a metrology system to track the platforms’ position and orientation in real time, ensuring precise measurement of grasping performance. The study demonstrates the effectiveness of the proposed GNC architecture in achieving successful manipulation, while also identifying key limitations related to relative kinematic state critical factors, such as relative velocity, orientation misalignment, and contact dynamics.