Deep learning-based energy prediction in wireless sensor networks

A wireless sensor network (WSN) plays a major role in network protection. In WSN, the data is routed towards the sink node and attack detection is the key process of WSN. In this research, the Taylor Walrus optimisation algorithm (Taylor WaOA) enabled LeNet model (Taylor WaOA-LeNet) is developed for attack detection. Here, the WSN is simulated and the deep Q network (DQN) is used for predicting the energy. The proposed Taylor WaOA is utilised in routing using fitness factors like energy, throughput, trust and distance. In base station (BS), the feature selection and data augmentation are carried out in attack detection; the proposed Taylor WaOA-LeNet model is employed. Furthermore, the energy, throughput, distance, trust, accuracy, sensitivity, and detection rate metrics are used to evaluate the model performance, which offered the finest values of 0.984 J, 0.073 Mbps, 67.08 m, 0.157, 0.918, 0.947, and 0.958.