A Novel Behavioral Monitoring based Trust Model for enhancing Edge Security using Adaptive Neuro-fuzzy Inference System

The Internet of Things (IoT) is in a recent state of instability due to the flooding of virtual data. It is believed that IoT and cloud computing have met their maximum thresholds and loading them with data after this point will only deteriorate their performance. Hence, edge computing has been introduced to mitigate the processing burden of IoT. To meet the security demands of edge computing, we intend to combine the method of blockchain along with edge computing for a better solution. Accordingly, this paper proposes the introduction of a novel blockchain model that is based on artificial neural networks and trust estimation called the behavioral monitoring trust estimation model. Performance metrics such as accuracy, precision, recall, and F-measure are calculated under normal conditions and under the injection of attacks like false data injection, booting attack, and node capturing. The proposed behavioral monitoring trust classification model is compared with existing classifiers like Naive Bayes, K-nearest neighbor, Auto Encoder, Random Forest, and Support Vector Machine, and is found to have improved performance. Additional evaluation parameters like execution time, encryption time, storage cost, computational overhead, energy efficiency, and packet drop possibility are also calculated for the proposed model and compared with existing blockchain techniques of Bitcoin, Ethereum, Hyperledger, Direct and indirect trust model, and mutual trust chain based blockchain model. The proposed model achieved an accuracy of 95%, a precision score of 90%, a recall score of 94%, and an F-measure of 94% indicating superior performance.