System-based Micro-nano Satellite On-board Computer Modeling And Fault Injection Research

With the development of space exploration and the diverse of its formation,the micro-satellites become a hot spot among the projects of universities and research institutes because of its low-cost,flexibility and reduced development cycle.The OBC(On-Board-Computer)is the command center in the micro-satellite,it is essential to validate its reliability under space radiation environment.OBC modeling and fault injection,as a new method of validation,has the advantages of fast simulation and compatibility,so it becomes a region which is worth of research.This research is based on space radiation environment and the features of OBC.Firstly,the radiation source and radiation effects are analyzed in the low-earth-orbit,and the OBC's radiation effects caused by energetic particles are introduced,based on which fault models are established.Secondly,after research of the chips modeling method both domestic and abroad,a fast modeling method which is based on SystemC and OVPsim is proposed.A TLM(Transaction Level Model)is established based on the structure of a typical OBC,which contains CPU,SRAM,Flash and I2C.Moreover,in order to simulate the usual OBC's fault-tolerant design,the model of dual-OBC-redundancy is established in order to test its resistance against SEU(Single Event Upset).In the aspect of fault injection,a fault injection mechanism is built to implement the function.The fault injection mechanism includes the fault injector,analyzer and observer which is responsible for fault generation,control of fault injection process and running status monitoring.Several algorithms which are usually used in the micro-satellite are compiled to run on the OBC model,and the faults are injected into the OBC model at the same time,the influence caused by the SEU to the OBC is analyzed,and the correctness of OBC model and simulation frame are assessed.Finally the fault injection based on OBC model and based on OBC hardware is carried out and the output of both OBCs are compared.The confidence level of model is assessed and the feasibility of using OBC model to evaluate its reliability is verified.