Adaptive and Responsive Design Under Uncertainty for Resource-Constrained Small Satellites

Abstract

The space environment that satellites face is uncertain and challenging for survival and operation. Traditionally, satellite design methods mitigate the effects of uncertainty through the use of ample margin. However, robust designs often sacrifice significant nominal performance in exchange for this reduced sensitivity to uncertainty. Small satellites in particular are limited in size, weight, and power (SWaP) and do not have the luxury of resources for ample design margin. They can ill-afford the performance sacrifice of robust design. As SmallSats continue to decrease in size - even down to the hundreds of grams - the need grows for design techniques that offer resilience under uncertainty without the inevitable sacrifice of performance that comes with robust design.

In this thesis, a methodology is presented to mitigate the effects of uncertainty in the space environment with the ability to adapt the satellite’s behavior during operation. Compensation for uncertainty on-orbit allows for dynamic allocation of margin on an as-needed basis, reducing the performance loss while improving the ability to maintain operation. The methodology covers two phases. First, design prior to operation enables provisioning of resources to plan for and provide the capability for passive and active dynamic mitigation of predicted uncertainty. Second, reprogramming of dynamic behavior in operation allows for optimal mitigation actual uncertainty. The resultant designs balance improved resilience in the face of uncertainty with minimal overdesign and sacrifice of performance.

The methodology is applied to a novel SmallSat concept, WaferSat - a SWaPconstrained satellite etched on a 300 g silicon wafer using microelectromechanical systems (MEMS) production. Optimization of active and passive dynamic compensation is utilized to mitigate effects of thermal uncertainty with limited sacrifice of payload power. Multiple design families - with the same available payload power (isoperforming) and confidence of operating temperature constraint satisfaction (isofeasible) are identified utilizing different combinations of responsive and adaptive mitigation techniques.

Application of the methodology is expanded to a second system, DiskSat, a similar larger, more thermally complex system. A detailed comparison of the continuum between responsiveness and adaptability is made, demonstrating the Pareto-set of isoperforming and isofeasible designs between two mitigation methods. Design for the balance of active and passive uncertainty mitigation over multiple constraints is explored, highlighting implementation considerations.