Optimization of Multi-arm Robot Locomotion to Reduce Satellite Disturbances for In-orbit Assembly

Kurzfassung

Traditionally, manufacturing and assembly of space assets is performed on ground before sending them into orbit. However, this monolithic approach involves high launch costs due to increasing asset sizes, e.g., large telescopes for space observation. Alternatively, in-orbit assembly of space structures after launching the raw materials to orbit opens wider possibilities at a reduced cost. Mobile robotics, such as walking manipulators or multi-arm robots, are a critical component for this approach due to their mobility in orbit. However, unlike terrestrial assembly tasks, the continuous motion of the robot and materials, coupled with the change of inertial properties of the structure, results in a rotational deviation of the platform due to conservation of angular momentum in orbit. This might violate the tolerance limits of the platform antennas cone angle for communication with the ground stations. Although exploiting the attitude control system of the platform is a straightforward solution, it might lead to issues related to the associated actuators like reaction wheels saturation, high-frequency vibration, or high fuel consumption. To deal with this problem, in this paper we formulate the attitude disturbance problem as a minimization of the effects created by the gait of the walking manipulator. Investigating the dynamic coupling between the robot system and the space structure gives a deeper understanding of the spacecrafts behavior depending on the robot gaits. The paper proposes a controller that optimizes the forces that the robotic arm applies to the structure, hence minimizing the base rotation. As an application, we use a space structure composed of identical elements, namely the mirrors of a segmented telescope, endowed with standard interfaces to allow the robot locomotion. We show the effects of optimizing these interaction forces in various scenarios and positions on the structure through multiple dynamic simulations.