Performance Evaluation of LoRa Communication in IoT-Based Agricultural Systems

The Internet of Things (IoT) is transforming agriculture with automation, real-time monitoring, and data-driven decision making that improves productivity and efficiency of resources. Among all wireless technologies, Long Range (LoRa) is one of the best Low Power Wide Area Network (LPWAN) solutions due to its long-range connectivity, low power consumption, and affordable deployment, making it most suitable for IoT-based agricultural applications in rural areas. This paper offers simulation-based performance evaluation of LoRa communications in several real-world wireless scenarios through multiple channel models such as Additive White Gaussian Noise (AWGN), Rayleigh, and Rician Fading models in the Indian ISM band (865–867 MHz). Important communication performance measures such as Bit Error Rate (BER), Signal-to-Noise Ratio (SNR), Received Signal Strength Indicator (RSSI), and Packet Delivery Ratio (PDR) are compared at different spreading factors and channel degradation. In addition, Forward Error Correction (FEC) improved PDR by 15–20% and signal strength across distances, Machine Learning-Adaptive Data Rate (ML-ADR) protocol based on random forest algorithm is proposed to adaptively adjust LoRa transmission parameters dynamically according to real-time channel conditions. Simulation results show that the suggested ML-ADR method reduces BER by up to 65% when SNR is low, increases energy efficiency by about 30%, and provides over 97% PDR at a distance of 500 m, which is better than traditional methods under all investigated channel models. The results confirm the scalability and reliability of LoRa for smart agriculture and support its use for large-scale deployment with reliability in rural IoT environments.