Research On Security Authentication Method For Internet Of Things

As the technical support of the Fourth-Generation Industrial Revolution,the IoT technology can be applied to a variety of aspects in human's activities,such as military application,smart transportation,smart home,smart healthcare,etc.With the expansion of the IoTs application areas,human's work and life activities are greatly facilitated.However,numerous potential security issues are exposed at the same time.As the first checkpoint in the process of IoTs devices accessing the network,authentication is particularly prominent in the entire IoTs security system.As a result,secure and efficient authentication mechanism in IoTs is a critical issue to be addressed.According to the disadvantages of the existing authentication methods in the IoTs,two new methods are proposed in this thesis.First,most existing authentication methods are not flexible enough to adapt to the actual complex environment.Toward this end,an IoTs authentication strategy based on indoor and outdoor scene recognition is proposed.Second,since most existing centralized authentication mechanisms have the problem of single-point failure,a de-centralized IoTs authentication method based on blockchain technology is proposed.Firstly,the classifier is designed.In this thesis,the Support Vector Machines with strong generalization ability and low limitation of sample size is selected as the classifier for scene recognition.Based on the polynomial kernel function and the RBF kernel function,a construction method of the hybrid kernel function is proposed.And the average classification accuracy,test accuracy and support vector ratio of K-CV are selected as the inputs of the fitness function,the particle swarm optimization algorithm is improved.Then,the improved particle swarm optimization algorithm is used to optimize several parameters of SVM classifier which is designed by using hybrid kernel function,including the polynomial kernel function weight a,the polynomial kernel parameter d,the RBF kernel parameter g and penalty factor C.As shown in the experimental results,compared with existing classifiers without optimization,the classifier designed in this thesis has significant improvements in classification accuracy and generalization capability.Additionally,it can be applied to scene recognition with better recognition accuracy and speed.Then,an IoTs authentication method based on indoor and outdoor scene recognition is proposed,ae well as an authentication framework.Then scene recognition module and scheme decision-making module of the authentication framework are designed and applicable scenarios of the authentication method are given.Finally,an IoTs authentication method based on blockchain technology is proposed.By employing de-centralized idea of blockchain technology,the authentication framework based on hyperledger is designed,including the design of smart contract and data structure.Experiment results show that this method can reduce the risk of single-point failure of IoTs system effectively and bring better robustness and practicability to IoTs system.