Multi-sensor Data Fusion Based on Fuzzy Neural Network and Its Application in Piggery Environmental Control Strategies

摘要

In intensive and large-scale farming, piggery environmental monitoring is an important method for improving productivity and preventing disease in pig herds. The piggery environment mainly includes air temperature, air humidity, and ammonia concentration. Piggery environmental monitoring mostly concentrates on a single environmental factor and data processing. Few studies focus on multiple environmental factors and information fusion from a variety of monitoring sensors. Therefore, in this paper, the main environmental factors that affect herd health, such as temperature, humidity, and ammonia concentration, were monitored. According to the analysis of piggery environmental adaptability indicators, a multi-sensor data fusion model with five layers based on the adaptive fuzzy neural network of T-S inference was established. The structure of the model is 3-8-18-18-1. To check the performance of the model, 432 groups of data were selected by monitoring a piggery environment for three consecutive days. A total of 288 sets of data were obtained for training, and 144 sets of data were obtained for checking. The result shows that the model effectively fuses multi-sensor data. After 600 training cycles, a Root Mean Squared Error (RMSE) of 0.14 × 10~(-3) between the actual output of the system and the desired output can be achieved. The system output has high accuracy. The RMSE of the checking data and fuzzy system output is 0.18 × 10~(-2). Moreover, the operation efficiency of the system is high. Thus, the system is suitable for analyzing data in real time and for the processing of piggery environments. This system provides a good data processing method for monitoring and controlling piggery environments.