Prediction of Path Loss in Wireless Communication Networks based on Swarm Optimized Gradient Dual Layer Graph Neural Network

The ability to make an accurate estimate of path loss (PL) is required to estimate the coverage that a transmitter will provide and enhance the functionality of a wireless network. It may be challenging for traditional PL models to accommodate the ongoing trend in diversity, time-varying, and mass wireless channel development. This study presents a novel technique, the swarm-optimized gradient dual layer graph neural network (SGDLGNN), which can effectively predict PL. The low-dimensional environmental characteristics are generated using linear discriminant analysis (LDA), which also removes irrelevant data from environments of the same kind. Moreover, measurements are used to acquire information on the base station (BS) and the receiver (Rx), including 3D positions, frequency, the BS's broadcast energy, antenna information, the feeder loss, and the received power of all locations. Seven datasets for PL prediction techniques are built using the data from BS and Rx and other environmental factors. The effectiveness of the suggested prediction model is shown by contrasting it with specific conventional approaches. Simulation findings show that the proposed prediction approach is greater than the traditional methods.