摘要
Unique mission architectures become possible when formations of close-proximity spacecraft have the ability to reconfigure on orbit to adapt to different mission requirements. A necessary precursor to the realization of these missions, however, is the development of control architectures that accommodate the different geometries of a reconfigurable space system at different mission phases. This paper investigates the closed-loop control of a reconfigurable set of two CubeSat-scale spacecraft that are connected via noncontacting flux-pinned interfaces. A design concept known as location-scheduled control is used both to tune a Proportional-Integral-Derivative control law and to choose an optimal configuration for the spacecraft formation during a rest-to-rest reorientation maneuver. The location-scheduled control approach makes use of a genetic algorithm to search the design space of configurations and controller gains. The simulated performance of the controller is compared to experiments performed on Cornell University's Float Cube testbed, a planar air-bearing environment for validating nanosatellite-scale spacecraft architectures. The system's performance using the location-scheduled control-derived gains shows the same trends in the simulated and experimental data while providing insight into the behavior of the flux-pinned interfaces. Both simulation and experimental results also suggest that the location-scheduled control algorithm can select a combination of controller gains and system configuration that outperforms a human engineer's informed trial-and-error search.
- 出版日期2013-12