GNSS for Closed Loop Control of Space Manipulators

Future in-situ servicing space mission will be increasingly autonomous and exploited by robotic manipulators to save risks and costs of human presence onboard spacecraft. The accuracy and the bandwidth requested in such a kind of manoeuvres are very high and a completely automatic control is likely to be mandatory. Feedback control loops need the knowledge of the state of the plant, usually gained in terrestrial applications by means of encoders located at the joints of the manipulator. Due to specific characteristics of space structures, i.e. the flexibility resulting from the limits on the mass to be launched, encoders could not be really effective. The paper proposes the possible use of GNSS signal – and specifically of its phase - to define the kinematic state of the robotic arms during in orbit (rendez-vous, docking, grasping) operations. Limited dimensions of typical GNSS antennas allow for their distribution along the span of each manipulator’s arm, to take into account flexibility effects. A really accurate determination of the state of the system can be obtained, a result which should be a quite complicate achievement by other possible techniques like RF ranging or image analysis. The bandwidth granted for by GNSS receivers is suitable for orbital manoeuvres, typically slow. As a drawback, a substantial amount of computation is requested in order to perform a RTK process; also, possible limitations due to multipath have still to be considered. Preliminary experimental tests, performed with two-link robotic arms moving on a limited friction table in order to validate the substantial amount of modelling and the numerical simulation work already carried out, are presented and discussed.