HATO-Deep LSTM: Hybrid Adaptive Trust Aware Optimization Enabled Deep Learning Classifier For Energy Efficient Routing And Attack Mitigation In IoT-WSN

The Internet of Things (IoT) is an emerging technology that gained significant attention in both academia as well as industry. It is a set of heterogeneous devices that are interconnected together for exchanging information between devices without any human intervention. Wireless Sensor Network (WSN) plays a major core part in IoT that enables to generation of seamless data and influences network lifespan. Each device in the WSN is referred to as a sensor node that can detect and process the sensor data between sensor nodes. Routing plays a major role in saving energy, but energy conservation is a major issue in resource-constrained networks. Hence, this research designs a Hybrid Adaptive Trust aware optimization based on Deep Long Short-Term Memory (HATO-Deep LSTM) for selecting the optimal node as cluster head (CH) and optimally performing the routing strategy. The trust factors determine the trustworthiness of all the adjoining nodes through collaboration among nodes and this factor enables the optimization to obtain the global best solution feasibly. The deep learning classifier employed in this research is highly effective in detecting malicious nodes by avoiding false nodes entering the data communication path. The proposed model effectively detects the misbehaving nodes and thereby the network throughput is increased. The performance attained by the proposed model in terms of number of alive nodes, end-to-end delay, throughput, and normalized energy is reported as 76, 0.81ms, 0.29bps, and 0.38J respectively with 100 number of nodes.