DEEP LEARNING-BASED INTRUSION DETECTION AND PREVENTION IN WIRELESS COMMUNICATION

Wireless sensor networks (WSNs) are made up of a large number of sensor nodes which collect data and send it to a centralized location. Nevertheless, the WSN has several security difficulties because of resource-constrained nodes, deployment methodologies, and communication channels. So, it is very necessary to identify illegal access in order to strengthen the safety measures of WSN. The use of network intrusion detection systems (IDS) to safeguard the network is now standard procedure for any communication system. While deep learning (DL) methods are often utilized in IDS, their efficacy falls short when faced with imbalanced attacks. An IDS based on a novel transfer deep multicolumn convolution neural network (TDMCNN) technique was presented in this study to address this problem and boost performance. The most significant features of the dataset are chosen using a cross-correlation procedure and then included into the suggested methods for detecting intrusions. The accuracy, precision, sensitivity, and specificity are used to conduct the analysis and comparison. The experimental findings verified the effectiveness of the suggested method over the status quo of deep learning models for attack detection.