Secure data transmission in wireless networking through node deployment and Artificial Bird optimized deep learning network

In the wireless sensor network (WSN), ensuring secure routing in the network is a crucial and critical task. Providing security and sustaining energy is still a difficult problem in the research community, despite the network’s adoption of multiple routing protocols. However, the traditional protocols encountered multiple challenges as the network's adoption is far more susceptible to attacks, tampering, and manipulation due to the deployment scenario, the characteristics of the sensor-equipped nodes, and their communication protocols. Consequently, an optimal solution to ensure secure node deployment for wireless network systems utilizing the Artificial Bird Optimization-based Deep Convolutional Neural Network in conjunction with the Modified Advanced Encryption Standard algorithm (ABO-mAES-DCNN) is developed in this research for effectively encrypting and managing the duty cycles. Specifically, the regional clustering with adaptive region partitioning enabled in this research is used to choose the cluster head. The region-based clustering’s adaptive partitioning feature dynamically modifies the domain's partitions or areas to ensure an even distribution of nodes. The proposed research exploits the deep CNN classifier, which precisely identifies the nodes' states adaptively selects each node’s scheduling mode, and carries out effective Duty cycle management. Further, the Modified Advanced Encryption Standard algorithm (mAES) is adopted to secure the data after identifying the node’s state. Specifically, Artificial Bird Optimization is utilized for optimum path selection and assists in minimizing energy consumption. The proposed ABO-mAES-DCNN model’s performance is reported in terms of alive nodes, delay, energy, and throughput as 14, 0.01 ms, 0.41 J, and 0.55 bps respectively for 100 nodes. With 200 nodes analysis, the ABO-mAES-DCNN model attained 115 alive nodes, a delay of 0.01 ms, energy of 0.35 J, and throughput of 0.47 bps and surpassed other existing techniques.