Path-Loss Model for Wireless Sensor Networks in Air Pollution Environments Leveraging of Drones

Recently, interest in wireless sensor networks (WSNs) and applications has grown significantly. The path loss caused by buildings and tall trees is a major obstacle in the distribution of sensor nodes. Because of absorption, dispersion, and attenuation by the structures or trees, a degradation occurs in the reliability of the communication link. In this paper, four path-loss models (PLMs) were developed for a LoRa (long-range) wireless sensor node mounted on the drone as a part of a WSN for monitoring air pollution over a distance of approximately 1,000 m utilising the received signal strength indicator (RSSI). Four novel PLMs were created using MATLAB’s curve-fitting tools, employing polynomial (POLY), exponential (EXP), Gaussian, and power equations. To improve the accuracy of these equations, their coefficients were fine-tuned using the particle swarm optimisation (PSO) algorithm, resulting in the development of hybrid algorithms—EXP-PSO, POLY-PSO, Gaussian-PSO, and power-PSO. The results demonstrated that PSO-enhanced PLMs achieved an outstanding correlation coefficient (R²) value of 1, indicating their exceptional ability to represent PLMs accurately in the tested air population environment utilising WSNs. Notably, the POLY-PSO algorithm outperformed other algorithms, achieving the lowest root-mean-square error (RMSE) of 1.945 × 10⁻⁴ and sum of squares error (SSE) of 1.853 × 10⁻²⁶. The PLMs based on hybridisation algorithms outperformed existing PLMs regarding the correlation coefficient. These findings hold significant implications for designing and optimising WSNs in air pollution applications, offering valuable insights into radio path-loss modelling in this area.