Secure Communication Using Multi-Layer Perceptron Neural Network and the Adaptive-Network-Based Fuzzy Inference System in Wireless Network

A Wireless Sensor Network (WSN) consists of many sensor nodes that collect data from various environmental conditions using the Internet of Things (IoT) and are often used to monitor and tune network environments. In this case, the presence of malicious nodes in the network leads to transmission security challenges, as it is believed to be a significant problem for successfully delivering captured data. Therefore, it is essential to protect network communication from security threats by detecting dangerous behavior for each sensor node and separating malicious nodes. This can be achieved by deploying an Intrusion Detection System (IDS) at the sensor nodes. However, limitations exist when dealing with high-dimensional data with complex underlying distributions. To tackle this issues, we introduce the Multi-layer Perceptron Neural Network (MLPNN) algorithm and the Adaptive-Network-Based Fuzzy Inference System (ANFIS) algorithm for secure communication in WSN. Initially, we gathered a dataset called Darknet Internet Traffic from an online source. This dataset was organized using the Min–Max Scaling (MMS) technique. Afterwards, our proposed method identifies the network traffic using Traffic Intensive Cumulative Rate (TICR) method. Based on the network traffic, we analyze the transmission delay and optimal route using Trust factor Evaluation Rate (TFER). Next, it picks the best features of malicious activity using the ANFIS algorithm. Lately, our proposed MLPNN classifier with ReLU activation function has been used to categorize malicious activity and improve security in the network. Therefore, the proposed classifier's significant advantages in this paper include increased classification accuracy, precision, recall, and F1-score. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.