Research On An Improved NHPP Software Reliability Evaluation Method Of IoT Intelligent Sensing Node

There are more and more applications of the Internet of Things,and the functions that can be realized are getting larger and larger.Today,the application of the Internet of Things can be seen everywhere,in every aspect of daily life.The Internet of Things has expanded the meaning of the Internet and has gone beyond the scope of communication between people.The Internet of Things embeds a variety of sensors in different entities to achieve the purpose of connecting objects to objects,transmitting messages in real time,and processing by the receiver.The more functions realized by the Internet of Things,the larger its scale and the higher its complexity.If a software failure occurs in the Internet of Things,it may cause property losses at a slight level,and may pose a threat to people's lives.The perception layer of the Internet of Things contains countless intelligent perception nodes,which are the foundation of the entire Internet of Things.Therefore,it is necessary to study the reliability of the software layer of the sensing node.This thesis puts forward an NHPP reliability growth model suitable for embedded software through in-depth research on the knowledge system of the Internet of Things architecture and software reliability growth model.First of all,this thesis analyzes the necessity and research methods of the random generalized multi-level software reliability growth model of the intelligent sensing node of the Internet of Things through in-depth research on the research background and significance of the subject.Secondly,this thesis introduces the basic modeling theory of software reliability growth model of Io T intelligent sensing node,analyzes the characteristics of the Io T architecture,and introduces the basic theory of SRGM.Next,on the basis of introducing the basic theory of the non-homogeneous Poisson process SRGM,this thesis gives the basic framework and assumptions of the NHPP model,introduces the classic G-O model and the model of Hoang Pham et al.,and establishes a new model: Weibull FDR SRGM with the Uncertainty of Operating Environments,abbreviated as W-E-SRGM.On this basis,this thesis combines W-ESRGM and Failure Mode and Effect Analysis Method(FMEA),to propose a new embedded software reliability analysis method: FMEA-WE.Finally,this thesis carries on parameter estimation and model comparison to the model,and uses LSE to estimate the parameters of the SRG(Software Reliability Growth)model proposed in this thesis.Using goodness-of-fit criteria such as MSE(Mean Squared Error),SAE(Sum Absolute Error),PRR(Predictive Ratio Risk),and PP(Predictive Power),the W-E-SRGM is compared with several other models.Experiments show that W-E-SRGM has more suitable results.