Dynamics and control of robotic systems for on-orbit objects manipulation

Multi-body systems (MSs) are assemblies composed of multiple bodies (either rigid or structurally flexible) connected among each other by means of mechanical joints. In many engineering fields (such as aerospace, aeronautics, robotics, machinery, military weapons and bio-mechanics) a large number of systems (e.g. space robots, aircraft, terrestrial vehicles, industrial machinery, launching systems) can be included in this category. The dynamic characteristics and performance of such complex systems need to be accurately and rapidly analyzed and predicted. Taking this engineering background into consideration, a new branch of study, named as Multi-body Systems Dynamics (MSD), emerged in the 1960s and has become an important research and development area in modern mechanics; it mainly addresses the theoretical modeling, numerical analysis, design optimization and control for complex MSs. The research on dynamics modeling and numerical solving techniques for rigid multi-body systems has relatively matured and perfected through the developments over the past half century. However, for many engineering problems, the rigid multi-body system model cannot meet the requirements in terms of precision. It is then necessary to consider the coupling between the large rigid motions of the MS components and their elastic displacements; thus the study of the dynamics of flexible MSs has gained increasing relevance. The flexible MSD involves many theories and methods, such as continuum mechanics, computational mechanics and nonlinear dynamics, thus implying a higher requirement on the theoretical basis. Robotic on-orbit operations for servicing, repairing or de-orbiting existing satellites are among space mission concepts expected to have a relevant role in a close future. In particular, many studies have been focused on removing significant debris objects from their orbit. While mission designs involving tethers, nets, harpoons or glues are among options studied and analyzed by the scientific and industrial community, the debris removal by means of robotic manipulators seems to be the solution with the longest space experience. In fact, robotic manipulators are now a well-established technology in space applications as they are routinely used for handling and assembling large space modules and for reducing human extravehicular activities on the International Space Station. The operations are generally performed in a tele-operated approach, where the slow motion of the robotic manipulator is controlled by specialized operators on board of the space station or at the ground control center. Grasped objects are usually cooperative, meaning they are capable to re-orient themselves or have appropriate mechanisms for engagement with the end-effectors of the manipulator (i.e. its terminal parts). On the other hand, debris removal missions would target objects which are often non-controlled and lacking specific hooking points. Moreover, there would be a distinctive advantage in terms of cost and reliability to conduct this type of mission profile in a fully autonomous manner, as issues like obstacle avoidance could be more easily managed locally than from a far away control center. Space Manipulator Systems (SMSs) are satellites made of a base platform equipped with one or more robotic arms. A SMS is a floating system because its base is not fixed to the ground like in terrestrial manipulators; therefore, the motion of the robotic arms affects the attitude and position of the base platform and vice versa. This reciprocal influence is denoted as "dynamic coupling" and makes the dynamics modeling and motion planning of a space robot much more complicated than those of fixed-base manipulators. Indeed, SMSs are complex systems whose dynamics modeling requires appropriate theoretical and mathematical tools. The growing importance SMSs are acquiring is due to their operational ductility as they are able to perform complicated tasks such as repairing, refueling, re-orbiting spacecraft, assembling articulated space structures and cleaning up the increasing amount of space debris. SMSs have also been employed in several rendezvous and docking missions. They have also been the object of many studies which verified the possibility to extend the operational life of commercial and scientific satellites by using an automated servicing spacecraft dedicated to repair, refuel and/or manage their failures (e.g. DARPA's Orbital Express and JAXA's ETS VII). Furthermore, Active Debris Removal (ADR) via robotic systems is one of the main concerns governments and space agencies have been facing in the last years. As a result, the grasping and post-grasping operations on non-cooperative objects are still open research areas facing many technical challenges: the target object identification by means of passive or active optical techniques, the estimation of its kinematic state, the design of dexterous robotic manipulators and end-effectors, the multi-body dynamics analysis, the selection of approaching and grasping maneuvers and the post-grasping mission planning are the main open research challenges in this field. The missions involving the use of SMSs are usually characterized by the following typical phases: 1. Orbital approach; 2. Rendez-vous; 3. Robotic arm(s) deployment; 4. Pre-grasping; 5. Grasping and post-grasping operations. This thesis project will focus on the last three. The manuscript is structured as follows: Chapter 1 presents the derivation of a multi-body system dynamics equations further developing them to reach their Kane's formulation; Chapter 2 investigates two different approaches (Particle Swarm Optimization and Machine Learning) dealing with a space manipulator deployment maneuver; Chapter 3 addresses the design of a combined Impedance+PD controller capable of accomplishing the pre-grasping phase goals and Chapter 4 is dedicated to the dynamic modeling of the closed-loop kinematic chain formed by the manipulator and the grasped target object and to the synthesis of a Jacobian Transpose+PD controller for a post-grasping docking maneuver. Finally, the concluding remarks summarize the overall thesis contribution.