Testing of a Docking Mechanism with a 6 DOF Manipulator and a Force-Torque Sensor

During the period I spent at Thales Alenia Space (TAS) to complete my Master's Thesis my job was to implement a suitable control strategy, using the Schunk 6DOF Powerball Light Weight Arm 4P, controlled with a custom framework developed by Università di Genova for TAS, and the Schunk FTM115 force-torque sensor, to test a docking mechanism previously designed by TAS in collaboration with Politecnico di Torino. The mechanism is composed by two parts, the male active one is controlled separately with an Arduino Mega microcontroller to change its position in order to recover for possible misalignments, reducing the contact forces between its telescopic probe and a cone-shaped passive half. The docking maneuver between two spacecrafts, called target and servicer, is composed by four phases: Approach and Deployment: the GNC systems of the two spacecrafts reduce the relative velocities, gradually moving the satellites closer and closer. At the end of this phase the passive half (on the target) and the active one (on the servicer) of the docking mechanism could be misaligned. Alignment: once the two parts are in contact, the active one moves in order to recover for the misalingments and reduce the contact forces. Soft docking: when the two parts are aligned three spring loaded petals connect them, still allowing small relative movements. Hard docking: the active part moves to reach the center of the mechanism and then the telescopic probe drags the target towards the servicer, until the passive half touches the support plate of the active one. Three radial hooks secure the two elements completing the docking maneuver. To test the mechanism the four phases have to be reproduced in the laboratory, maintaing fixed the active element, mounted above the arm, while the passive half is mounted on the arm. During the first two phases the arm simply has to move rigidly towards the active element, but in the other two it has to follow the movements imposed by the docking mechanism. The original idea was to implement an impedance control algorithm but, due to impediments of practical nature, mostly related to the lack of documentation about the arm and the framework and the impossibility to have direct access to the CAN bus network dedicated to the motors of the manipulator, a simpler but still effective strategy was chosen: the black-box approach. It consists in using the data, coming from the CAN bus network dedicated to the force-torque sensor, to produce a velocity reference to be given to the framework, that can control the arm in the Cartesian space. In this way the end effector moves in the same direction of the sensed forces, with a velocity proportional to them. Since the goals is to test only the functionalities of the mechanism and not its structural resistance, this fully compliant approach proved to be sufficient. Finally this control strategy, after being successfully implemented and tested, was also incapsulated in a terminal command, in order to make any user able to quickly use it.