TLoRa: On a Feedback-Driven Segmented Turbo Coding Scheme for LoRa Sensor Network Communication in Agriculture

In smart agriculture, agricultural sensor networks rely on reliable data transmission, but traditional LoRaWANs face challenges in low signal-to-noise environments. The primary reason for this is the weak error correction capability of the Hamming coding in the LoRa physical layer, which makes it difficult to correct multibit errors. To address these issues, we propose TLoRa, which integrates enhanced turbo codes into LoRa networks and combines them with LoRaWAN’s feedback mechanism to improve transmission reliability and energy efficiency. TLoRa addresses the following three challenges: 1) TLoRa employs a simplified recursive systematic convolution (RSC) coding scheme, ensuring its compatibility with the requirements of sensor nodes; 2) although RSC coding is simpler, it does not have the same error correction capability as turbo coding. Therefore, we use the feedback mechanism of LoRaWAN to implement joint decoding, which enhances the error correction capability and improves the transmission reliability; and 3) in addressing the limitations of conventional decoding algorithms (e.g., maximum a posteriori (MAP) and Viterbi), TLoRa employs the soft output path selection (SOPS) algorithm, which offers enhanced decoding performance while concurrently reducing computational complexity. Outdoor experimental results show that TLoRa performs well in harsh environments, especially under heavy rain and occlusion conditions, with a packet delivery ratio (PDR) of more than 87%, proving its transmission reliability.