Maintenance Architecture Optimization of a Distributed CubeSat Network Based on Parametric Model

Due to shorter development cycle and lower cost, CubeSats have been widely used in space science, technology, and business missions. CubeSat usually forms a formation/constellation, which boosts the capability of implementing complex space missions. However, the distributed CubeSat network is prone to malfunction. This paper envisions a maintenance architecture that includes launching and replenishing spare CubeSats to replace the faulty one on a regular basis. The major effort is to optimize this architecture in terms of total cost by taking the stochastic failures into consideration. In particular, a parametric model fitted from practical data is used to represent the realistic CubeSat lifetime distribution. A CubeSat lifetime database of 111 CubeSats has been built. The parametric model is obtained via a Bayesian estimation scheme. A cost model that is composed of fixed cost, holding cost, and shortage cost have been proposed. Then, A Monte Carlo simulation-based approach has been adopted to evaluate the cost. Finally, the optimal arrival time and quantity of backup CubeSats corresponding to minimal cost have been obtained by examining all the feasible combinations of arrival time and quantity of backup CubeSats. Results show that the CubeSat network should be replenished in the early stage which agrees with the high infant mortality trend of CubeSats.