A Study To Develop a Security Algorithm for Sybil Attack for Environment in Internet of Things Establishment

Abstract

This manuscript contributes a brief description of Internet of Things (IoT) system development with wireless sensor network (WSN) to prevent the network from the intruders. Safety and privacy are the main issues for IoT applications and the researchers still have to face some big challenges in the design of IoT system. In order to make possible this emerging domain, an IoT security system is designed to secure the network from the external attack. In this manuscript an IoT network is designed that comprises of two sections; Wireless Sensor network and Authentication system using Biometrics. Biometric is a pattern recognition system which can be either an identification or verification system. WSN mainly comprises of small sensor devices that are interconnected through wireless network. Small, stable, cheap, low powered Wireless Sensors can make the smallest devices installed in any environment, at a reasonable cost. The combination of these devices with IoT can be considered as an essential development of WSN. For authenticating the network, fusion of biometric authentication system is used. Biometric fusion is the best method used in the IoT system to secure the network from the various types of attacks and to secure the important data during the transmission. In this research, a biometric image fusion based on fingerprint and face is used. Initially, pre-processing is applied on the uploaded test images, which is used to remove the unwanted areas and detect the face. The feature vectors of the face images are extracted by using Scale Invariant Feature Transform (SIFT). Minutia feature extraction is used for fingerprint images. After this, optimization algorithm named as Genetic Algorithm (GA) will be applied on both the face and fingerprint images that will reduce the unwanted features from the feature extracted image. Then training of uploaded images will be done by using Artificial Neural Network (ANN). The proposed biometric image fusion recognition system will overcome the limitations of individual biometric systems in terms of response time and accuracy. After training and testing, next step will be of decision making i.e. to matched the test image (face and finger print) with the save training database. Then the parameters like FAR, FRR, Accuracy, Throughput, PDR, Energy Consumption and Delay will be used. newline newline newline newline