A Soil pH Sensor and a Based on Time-Series Prediction IoT System for Agriculture

• Saulo Aguiar UFPE
• Edna Barros UFPE

Resumo

Agriculture 4.0, which combines precise sensors, The Internet of Things (IoT) and Artificial Intelligence (AI) revolutionize precision agriculture. This paper addresses the union of technological solutions from new sensors, AI, IoT, and Edge Computing to deliver an innovative solution. The pH is one of the most critical parameters of the soil, presenting a diagnosis of the soil. Depending on the pH value, plantations can develop or not. In this work, an accurate and robust soil pH sensor was developed, as well as a deep learning model for time series prediction operating at the edge of the IoT system. The proposed sensor showed good performance and a low error percentage compared to commercial sensors and pH measurement by traditional laboratory methods. To take advantage of the richness of data generated by the pH sensor, a neural network LSTM(long-term memory architecture) was implemented for time series prediction. The LSTM model integrated into the edge computing system demonstrated accuracy in periodic pH predictions. This implementation provides valuable insights for agriculture and smart practices, empowering farmers and agronomists to maximize crop productivity.

Palavras-chave: Soil pH Sensor, Long Short-Term Memory( LSTM), Edge Computing, Time Series Prediction, Agriculture 4.0, Internet of Things(IoT), Deep Learning