A Novel Biometric Key Security System with Clustering and Convolutional Neural Network for WSN

Abstract

Development in Wireless Communication technologies paves a way for the expansion of application and enhancement of security in Wireless Sensor Network using sensor nodes for communicating within the same or different clusters. In this work, a novel biometric key based security system is proposed with Optimized Convolutional Neural Network to differentiate authorized users from intruders to access network data and resources. Texture features are extracted from biometrics like Fingerprint, Retina and Facial expression to produce a biometric key, which is combined with pseudo random function for producing the secured private key for each user. Individually Adaptive Possibilistic C-Means Clustering and Kernel based Fuzzy C-Means Clustering are applied to the sensor nodes for grouping them into clusters based on the distance between the Cluster head and Cluster members. Group key obtained from fuzzy membership function of prime numbers is employed for packet transfer among groups. The three key security schemes proposed are Fingerprint Key based Security System, Retina Key based Security System, and Multibiometric Key based Security System with neural network for Wireless Sensor Networks. The results obtained from MATLAB Simulator indicates that the Multibiometric system with kernel clustering is highly secured and achieves simulation time less by 9%, energy consumption diminished by 20%, delay is reduced by 2%, Attack Detection Rate is improved by 5%, Packet Delivery Ratio increases by 6%, Packet Loss Ratio is decreased by 27%, Accuracy enhanced by 2%, and achieves 1% better precision compared to other methods.

Discrete Chaotic Fuzzy Neural Network (DC-FNN) Based Smart Watch Embedded Devices for Sports and Health Monitoring

“…The existing methods such as Deep generative Bayesian optimization (D-BO) [26], Internet of Things (IoT) [27] and Smartphone Sensors (S-S) are used to assess the proposed method (DC-FNN). The parameters such as cost efficiency, processing time, encryption time, decryption time, energy consumption and precision are obtained.…”

Unsupervised Learning-Based Exploration of Urban Rail Transit Passenger Flow and Travel Pattern Mining

“…Unsupervised learning, a subset of machine learning techniques that operates without labeled outcomes, is increasingly applied across various domains to extract meaningful patterns from data [9,10]. This section synthesizes insights from recent studies [11,12,13,14,15] to demonstrate the versatility and impact of unsupervised learning methodologies.…”

“…The data sources are extensive, including usage data from smart transit cards and video surveillance data from stations, all of which form the basis for analysis. By employing advanced data mining technologies and algorithms such as K-Means++, DBSCAN, etc., these massive datasets can be effectively processed to extract passengers' travel characteristics, thereby offering more precise operational decision-making support [9,10,11,12].…”

A Named Entity Recognition Method Enhanced with Lexicon Information and Text Local Feature

“…There are many special cases in TCM ancient texts, for example, "牛黄解毒丸" and "牛黄" belong to the category of prescriptions and material medical respectively, so Local feature modelling plays a role in the task of TCM named entity recognition. CNN [17,18] is widely used in image processing because of their efficient local modelling performance. Therefore, we used CNN to model the local features of the sentences.…”