No abstract

The wireless Visual sensor network is a highly functional domain of high-potential network generations in unpredictable and dynamic environments that have been deployed from a large number of uniform or non-uniform groups within the desired area, cause the realization of large regulatory applications from the military and industrial domain to hospital and environment. Therefore, security is one of the most important challenges in these networks. In this research, a new method of routing smart cameras with the help of cloud computing technology has been provided. The framework in the cloud computing management layer increases security, routing, inter interaction, and other features required by wireless sensor networks. Systematic attacks are simulated by a series of standard data collected at the CTU University related to the Czech Republic with RapidMiner software. Finally, the accuracy of detection of attacks and error rates with the suggested NN-SVM algorithm, which is a combination of vector machines and neural networks, is provided in the smart cameras based on the visual wireless sensor networks in MATLAB software. The results show that different components of the proposed architecture meet the quality characteristics of visual wireless sensor networks. Detection of attacks in this method is in the range of 99.24% and 99.35% in the worst and best conditions, respectively.

Today detection of new threats has become a need for secured communication to provide complete data confidentiality, integrity and availability. Design and development of such an intrusion detection system in the communication world, should not only be new, accurate and fast but also effective in an environment encompassing the surrounding network. In this paper, a new approach is proposed for network anomaly detection by combining neural network and clustering algorithms. We propose a modified Self Organizing Map algorithm which initially starts with null network and grows with the original data space as initial weight vector, updating neighborhood rules and learning rate dynamically in order to overcome the fixed architecture and random weight vector assignment of simple SOM. New nodes are created using distance threshold parameter and their neighborhood is identified using connection strength and its learning rule and the weight vector updating is carried out for neighborhood nodes. The Fuzzy k-means clustering algorithm is employed for grouping similar nodes of Modified SOM into k clusters using similarity measures. Performance of the new approach is evaluated with standard bench mark dataset. The new approach is evaluated using performance metrics such as detection rate and false alarm rate. The result is compared with other individual neural network methods, which shows considerable increase in the detection rate and 1.5% false alarm rate.