A study of data filtering and processing methods in agricultural sensor networks

Abstract

Abstract. The article considers the problem of noise influence on measurement accuracy in sensor systems for precision agriculture. It is shown that the presence of interference, signal instability, and external factors significantly reduces the reliability of data obtained from soil moisture, temperature, and pH sensors. The efficiency of common signal filtering methods, including low-pass filters, median filtering, exponential smoothing, and the Kalman filter, is analyzed. Their advantages, disadvantages, and application features under conditions of varying noise levels are determined. An adaptive approach to signal processing based on automatic parameter estimation is proposed, using the mathematical relationship between the Kalman gain coefficient and the smoothing coefficient of exponential smoothing. To verify the effectiveness of the proposed approach, a model of adaptive sensor signal processing was developed and its operation under noise influence conditions was investigated. The operation of an automatic irrigation system using filtered sensor data was simulated, and a comparative analysis with classical filtering methods was performed. The simulation results demonstrated an increase in the accuracy of sensor data processing, a reduction in measurement error, and an improvement in the efficiency of automatic irrigation system control. The obtained results confirm the feasibility of using adaptive filtering in agricultural sensor networks. The practical value of the work lies in the possibility of applying the proposed method in robotic agricultural systems.