Optimization assisted deep learning based intrusion detection system in wireless sensor network with two-tier trust evaluation

Nowadays, owing to the openness of transmission medium, wireless sensor networks (WSNs) suffer from a variety of attacks, together with DoS attacks, tampering attacks, sinkhole attacks, and so on. Therefore, an effectual system is necessary for recognizing the intrusions in WSN. This paper aims to set up a novel intrusion detection system (IDS) via a deep learning model. Initially, optimal cluster head (CH) is selected among the sensor nodes, from which the sensor nodes that have high energy will be prioritized to act as CH. In this proposed work, the CH selection is evaluated optimally by not only considering the energy parameter, further under the constraints like delay and distance. For optimal selection, a novel approach named as self-improved sea lion optimization (SI-SLnO) model is introduced in this work. As per the proposed strategy, the trust of CH and nodes is evaluated based on a multidimensional two-tier hierarchical trust model by considering content trust, honesty trust, and interactive trust. Finally, the deep learning-based intrusion detection takes place via optimized neural network (NN), where the training is done by the proposed SI-SLnO algorithm via the optimal weight tuning process. At last, the supremacy of the developed approach is examined via evaluation over numerous extant techniques.