Fault-Tolerance Technique Research For Parallel Processing On-Board Computer Systems

Small satellite has become a hot research in the area of aerospace for its great advantages such as short period of development, agility of launch mode, low cost, poor risk, and ect. PFT-OBCS ( Parallel Fault-Tolerance On-Board Computer System ) which is a parallel processing On-Board computer system of four-CPU based on COTS technique. This system using RTEMS and processor with SPARC V8 architecture as its software-hardware platform, have the functions of parallel computing based on message passing, state monitoring and self-configuration, data exchange and distribute, software re-injection, function reconstruction in orbit, and ect.In this thesis, we study the key technologies of PFT-OBCS, analysed the Fautl-Tolerance architecture and sourcecode of RTEMS carefully. Fault-Tolerance policies based on these research have been established and hardware support for these policies was put forward. we focus on some characteristic practical Fault-Tolerance technique, such as software re-injection, slave reboot etc. Combined the architecture with policies, reliability modeling for PFT-OBCS has been done based on combined reliability model and markov reliability model, pointed out the key components which affect the reliability of system most. Then, we test the function of Fault-Tolerance module and performance of PFT-OBCS. Last, summarized and analysed the problems during development, we looked forward to the prospect.