A Novel SDN-Based Security Framework for Wireless Sensor Networks Using TDCNN and PGF-ECC

Wireless sensor networks (WSN) comprise a huge number of stationary or mobile sensor nodes that considerably support multi-hop communication. Intruder attacks such as wormholes, spoofing, etc., make it more susceptible in turn influence WSN’s confidentiality. Therefore, utilizing tanh activation function centered deep convolution neural network (TDCNN) and polynomial Galois field elliptical curve cryptography (PGF-ECC), a novel approach named malicious node detection and privacy preservation of software-defined WSN is conceptualized in this work. Many measures have been taken in the work for assuring security. Utilizing taxicab metric centered KMedoid (TM-KMedoid) algorithm, the sensor nodes of the network are initially sorted into clusters. After that, using the Horse optimization algorithm (HOA), the cluster head (CH) selection is performed where the best node in the cluster is selected as the CH. Next, the data packets are transferred by the CH to the base station (BS) via the software-defined networking (SDN) controller. The SDN controller is applied with a TDCNN classifier that predicts the node’s trustworthiness for detecting whether the node is malicious or not. The data is transferred to BS if the node is a trusted one else, the node gets rejected. The PGF-ECC algorithm is employed for ensuring the data’s secure transmission to the BS. The proposed method’s performance is contrasted with the existent systems in an experimental evaluation. It exhibits superior performance when analogized to top-notch techniques, and highly assures the network’s authenticity. © 2023, Human-centric Computing and Information Sciences. All Rights Reserved.