Study of Algorithms of Real-Time Scheduling with Fault-Tolerance in Small Satellite On-Board Computer Systems

Scheduling small satellite's mission must meet with two acquirements: one is that all real-time tasks must be complemented in time limit, the other is that it must have fault-tolerant capability. In this paper, an integration mission model of both periodical mission and non-periodical mission is built based on the characteristic of small satellite mission and a fault-tolerant scheduling algorithm of static scheduling and dynamic adjustment for small satellite is presented combined with mission allocate algorithm and partial dynamic scheduling algorithm of embedded processor. This algorithm combines both the technique of primary/backup copies and the Hopfield algorithm and sum up static assignment to planning and operation problem held in resource and time. Neural network algorithm is used in partial dynamic adjustment of non-periodical mission and backup mission copies that assures the reliability and temporal effectiveness of mission execution. Simulation results show that it achieves higher schedulability compared to some other methods. Further, in terms of scheduling length and load balancing, it is superior to that of conventional graphic algorithm