Font Size: a A A

Research On FPGA Online Evolution Self-repair Method

Posted on:2021-04-22Degree:MasterType:Thesis
Country:ChinaCandidate:S YiFull Text:PDF
GTID:2428330611998869Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
With the increasing integration of electronic technology and the gradual progress of science and technology,electronic information systems working in the environment of space,deep sea,etc.also have more functions.Technological development has enabled humans to have a far-reaching influence on every frontier of the Earth and beyond.However,in such extreme environments,the system will inevitably encounter various types of faults and errors.How to eliminate and repair these faults has always been a difficult problem when it comes to designing the system.The traditional repair method has great limitations.On the one hand,it is too wasteful of resources.On the other hand,the effect cannot meet the requirements.Therefore,an efficient and feasible self-repair method is needed to apply to extreme conditions of electronic information systems.Evolutionary hardware,as a developing technology,has great potential in dealing with fault repair.The combination of a microprocessor and a programmable logic area allows the system to be functionally robust.The use of dynamic partial reconfiguration technology(DPR)in the programmable logic area makes the evolution more efficient and less energyconsuming.This dissertation studies the use of genetic algorithms to repair programmable logic hardware resources based on DPR technology.The outline of this research is described as follows:First,the author examines the similarity between hardware and biological tissues in structure,analyzes the repair of tissues in biological systems,and then makes an analogy between that process and the self-repair of hardware.The selfrepair method at the hardware organization level is proposed,and the feasibility analysis of the repair of programmable logic resources is carried out through the use of DPR technology.Secondly,the dissertation discusses the use of Zynq-7000 series So C as the system's experimental verification platform to design the self-healing evolutionary hardware system.As an indispensable part of the self-healing system design,static and dynamic resources on the PL side of the system are allocated,and the method of the configuration file is selected.At the same time,the way of using storage modules such as DDR and SD card is designed,and the implementation processes of the fault injection module and the repair evolution calculation module are elaborated upon.There are solutions to various faults caused by possible radiation in space.Finally,the verification part of the experiment.First,the feasibility of the selfhealing system was verified by simulation;then the reliability and portability of the self-healing evolutionary hardware system were verified by board-level experiments.The data analysis proves that the system can achieve the expected goal and complete the expected function.
Keywords/Search Tags:evolvable hardware, genetic algorithm, error self-repair, dynamic partial reconfiguration, field programmable gate array
PDF Full Text Request
Related items