Range Characterization of a LoRa-Equipped Decentralized Wireless Sensor Network for Monitoring Forest Areas

Abstract

A decentralized four-node LoRa wireless sensor network was developed and tested to characterize its performance in a forested
environment in terms of RSSI, SNR, and packet error rate. Each ESP32-based node transmitted temperature (BMP280), GPS, and time data, while a
receiver-rotation scheme ensured controlled, non-overlapping transmissions. The system demonstrated full coverage of Arroceros Forest Park, with
performance variations primarily driven by distance, antenna height, and foliage. Regression analysis from controlled two-node experiments showed that
RSSI is strongly predicted by these factors (R² = 0.8468) and SNR to a lesser extent (R² = 0.4602), with all models statistically significant (p < 0.001).
Distance had the strongest influence on RSSI, while foliage introduced the largest degradation in SNR and PER. In the four-node deployment, asymmetric
link behavior and higher PER values were consistently observed on foliage-obstructed paths and on longer links, confirming the regression trends and
highlighting foliage as the dominant source of performance loss. Overall, the measured PER, SNR, and RSSI align with the propagation behavior predicted
by the models, validating the network’s reliability across the test area.