Research And Implementation Of Distributed Intelligent Intrusion Detection System For Internet Of Vehicles

With the continuous development of networking technology,the network security problems of the Internet of vehicles are gradually exposed,and the research on the security of the Internet of vehicles is also in-depth.Intrusion detection technology as an active means of security protection has been applied to the Internet of vehicles,the traditional intrusion detection technology is not suitable for the complex multi-dimensional environment of the Internet of vehicles,the existing research work of the Internet of vehicles security pays more attention to vehicle security and functional security,lack of intrusion detection research,but more examples prove that the intrusion behavior is constantly occurring and increasing,research suitable for the Internet of vehicles intrusion detection technology is very necessary.Therefore,in view of the complex multi-dimensional conditions of the data of the Internet of vehicles,this thesis carries out the feature processing of the data of the Internet of vehicles,and uses convolution neural network algorithm and distributed big data technology to realize the design and implementation of the intrusion detection system,and makes systematic research on the four aspects of the data dimension reduction method of the Internet of vehicles,convolution neural network algorithm,distributed big data technology and the design and implementation of the intrusion detection system.In view of the problems of high data dimension and various types of features of Internet of vehicles data,this thesis proposes an intrusion detection data dimension reduction method based on improved principal component analysis.Firstly,the local linear blocks are processed,and then the coordinates are aligned by using the coordinate arrangement method.Finally,the overall low dimensional data is obtained.The experimental results show that the accuracy of this method is 98.54% and 96.91% respectively,compared with 89.45% and 87.27% of the traditional principal component analysis method.In view of the poor generalization ability and detection accuracy of convolutional neural network,this thesis proposes an intrusion detection method based on improved convolutional neural network.First,the end-to-end semi supervised training classifier and hierarchical learning method are used to detect the network data,then the sample features are learned,and the process is simplified.Experimental results show that the accuracy of the improved method is 97.01% and 96.92%,which is higher than that of convolution neural network algorithm 93.34%and 92.54%.Aiming at the problem of high detection accuracy and real-time performance in the Internet of vehicles,this thesis proposes a distributed intelligent intrusion detection method based on spark framework.The improved convolutional neural network is combined with longterm and short-term memory to build a big data distributed framework,and ICNN-LSTM combined deep learning method is run on spark platform.The experimental results show that the detection accuracy of the distributed combined deep learning method is 99.7% and 98.6%respectively,which is significantly higher than that of the convolution neural network 94.9%and 95.6%,and the detection time is also better.On the basis of the above research work,this thesis designs and implements a distributed intelligent intrusion detection system for the Internet of vehicles,which provides reliable protection for the security of the Internet of vehicles through data collection,feature processing and distributed intrusion detection.