1. A hierarchical control structure for the autonomy of next-generation space missions:
Due to safety concerns and high cost of space missions, the existing Guidance, Navigation and Control (GN&C) strategies for robotic operations suffer from a number of drawbacks that impede full autonomy: (i) too conservative control protocols with continuous need for human supervision; (ii) separate control protocols for various operations; (iii) slow (quasi-static) operations when the system is in contact; and (iv) limited workspace due to singularities. To address these problems, it is beneficiary to study the so called off-nominal situations and their recovery plans for space robotic systems, such as, operating in the presence of singularities, contact forces and transition between modes. This study when considered alongside the synergy between subsystems will result in more comprehensive decision-making strategies for long-term control, and finally propose an intelligent hierarchical controller for next generation space missions. This research will be focused on the following topics: (i) different singularity categories, and (ii) interaction between subsystems and the environment.
2. Geometric adaptive and robust control of planetary exploration rovers:
Planetary exploration rovers are systems with long life cycles, over which they must perform many tasks autonomously. Reliable autonomy is particularly important for the GN&C of such systems, since they cannot be operated remotely due to communication latency, high speed for covering vast areas, and facing many environmental disturbances. Control techniques for nonholonomic multi-bodies, developed based on feedback linearization in the reduced phase space, can improve the traction of autonomous rovers through incorporating ideal nonholonomic constraints in the controller to guarantee consistency of control signals with the constraints. However, any controller designed based on feedback linearization is highly sensitive to uncertainties in the kinematic, dynamic and constraint model of the system. An exploration rover also encounters many environmental disturbances, such as, slip/skid of the wheels, oversized obstacles and extreme weather conditions. This research will focus on extending geometric control techniques for nonholonomic systems to design controllers that can minimize the cumulative errors in the system, and hence, help with the long-term autonomy of the exploration rovers by enhancing their traction. Such controllers will be robust to some modelling uncertainties or disturbances, e.g., mass variation, or they will improve localization by detecting wheel slip/skid. This study can be further classified into the following topics: (i) adaptive control of constrained mechanical systems with symmetry, and (ii) robust control of mechanical systems with singular and nonlinear nonholonomic constraints.
3. Observability and control of hyper-flexible light space manipulators:
The trade-off between the mass and stiffness of the space manipulators contributes to their high manufacturing and launching cost. This cost can be considerably reduced if hyper-flexible light manipulators can be reliably controlled. Therefore, this research targets at proposing a high-fidelity model of hyper-flexible multi-bodies not only to perform nonlinear modal analysis, but also to investigate nonlinear observers for the control of the elastic degrees of freedom of the next-generation space manipulators. This study can be broken-down into two topics: (i) studying the observability and designing observers for hyper-flexible space robots, and (ii) introducing novel force/position controllers that guarantee the precision of motion and the desirable system compliance for various operations.
4. Space debris removal using chaser-manipulator system:
Planning for space debris removal missions with a chaser-manipulator system is specifically challenging for large tumbling debris, such as EnviSat (hence the name e.Deorbit for the European Space Agency’s debris removal program). Without achieving a proper relative position and velocity between the manipulator end-effector and a natural feature on the space debris suitable for grappling the mission can ultimately fail due to either miss-capture or high loads during/after the capture. Therefore, high precision of the proximity operations in a space debris removal mission is crucial. Proximity operations by a chaser-manipulator system can be divided into the following sequence: (i) rendezvous and approach; (ii) feature identification on the debris; (iii) pose and motion estimation of the debris; (iv) feature tracking and capture with the arm’s end-effector; (v) rigidization of the arm; and (vi) performing deorbiting maneuvers. This research will focus on the guidance and control algorithms for pre-capture maneuvers of the chaser-manipulator system to guarantee successful feature tracking and autonomous capturing of a tumbling debris in the presence of environmental disturbances. Due to the relatively similar mass properties of chaser and manipulator, the dynamic coupling effect between them is considerable thus rendering the GN&C design complex. For safety reasons the pre-capture maneuvers are normally divided into 2 phases: (i) formation flying of the chaser to synchronize with the tumbling debris while manipulator is in braked mode, and (ii) Controlling the manipulator to approach the debris while the chaser is freely tumbling.