Giant Perfumer Armadillo Optimization‐Driven Deep Learning Model for Mobile Target Tracking in Wireless Sensor Networks

ABSTRACT

Wireless Sensor Networks (WSNs) play an important role in adaptive target tracking in different areas like surveillance, environment monitoring, and disaster management. Nevertheless, the realization of precise positioning and energy‐efficient wireless communication is still difficult to achieve. The problem of target tracking is addressed using a Hybrid Giant Perfumer Armadillo Optimization Algorithm (HGPAOA), which combines the two distinct behaviors of armadillos: digging for global exploration and following scent trails for local exploitation. Additionally, received signal strength indicator values are used to guide the solutions toward the target's potential location. Fitness is evaluated using the dominance of the signal, resulting in swift optimization of the global solution, along with learning in space and time to ensure accuracy, robustness, and energy conservation. The energy efficiency is improved through the use of a cluster‐based Hyperbolic Reinforced GhostNet Affinity Attention Network (HR‐GAAN) for an adaptive sleep/awake schedule that activates only the nodes that are contextually necessary. This way redundant activity is avoided, thereby reducing communication cost and maximizing network lifetime without compromising on tracking performance. To ensure security, the Covertext Encryption Model (CEM) is used to embed encrypted tracking data within benign covertext, mitigating interception risks. Extensive MATLAB simulations demonstrate tracking accuracy of 99.5%, energy consumption as low as 0.2091 J, and low delays. Compared with existing methods, the proposed approach achieves superior accuracy, efficiency, and security, making it a robust and scalable solution for real‐time WSN tracking in complex environments.