| dc.description.abstract | The use of highly autonomous free-flying spacecraft has been investigated for potential utility in future human spaceflight endeavors. In general, ’free-flyer’ robots are small, self-sufficient spacecraft that operate near the exterior of larger space structures, such as the International Space Station, and are designed to provide support during various operations. Free-flyer designs and concepts often include a large degree of autonomy to provide mission support with little operational overhead.
One of the building blocks of autonomous free-flyer behavior is safe and reliable point-to-point maneuvering during proximity operations. This thesis explores the development and simulation of an integrated guidance and control (G&C) system to enable safe free-flyer point-to-point maneuvering in proximity to larger space structures, including the avoidance of collision and jet plume impingement. The foundation for this system is an existing trajectory planning method introduced by Roger [1]. This method represents the free-space as a discrete harmonic potential field and uses the resulting field gradient as a condition for generating collision-free trajectories that efficiently account for natural dynamics. A real-time guidance process is built around this method that can manage an internal model of the environment based upon obstacle mapping data and react quickly to dynamic obstacles. A linear-programming jet selection technique is implemented to fulfill six DOF velocity impulse commands using a free-flyer’s RCS propulsion system, and additionally is augmented to include jet plume impingement avoidance functionality. Finally, an attitude con-troller process was developed and implemented to enable the free-flyer to reach and track a desired attitude during translational maneuvers.
To verify the system’s capabilities, a test-bed simulation was developed using the SpaceCRAFT platform, specifically utilizing it’s modular, asynchronous architecture. In a set of four maneuver-ing tests set in distinct obstacle environments, the G&C system demonstrated the ability to maneuver the free-flyer to the goal state along collision-free trajectories. In three of the test cases, the plume avoidance strategy results in a large reduction in the accumulated plume cost (36-54%). Overall, these simulation results demonstrate that the system enabled point-to-point maneuvering for a reference free-flyer design, and support its feasibility and practicality.
This work represents a somewhat unique approach to free-flyer autonomous maneuvering in that it departs from a trajectory planning, or offline-online paradigm. Instead, this approach relies on the refinement of an internal model of the environment and the resulting potential field to per-form reactive path-finding. This approach results in better overall flexibility when the free-flyer lacks knowledge of the position, geometry, and motion of nearby obstacles. In the future, the integration of Simultaneous Localization and Mapping (SLAM) and 3D mapping algorithms will help to further verify the feasibility of this approach. Other future improvements include integrating more robust methods of dynamic obstacle avoidance and automated positioning and scaling of the potential field grid. | |
