Developing an Early Warning System for Pig Resilience Using Sensor Data

The welfare of pigs in intensive farming environments is increasingly becoming a concern due to their exposure to various environmental and physiological stressors. In this study, we aim to enhance the resilience of pigs by implementing an early warning system. By utilizing non-invasive sensors, we continuously monitor several critical parameters, including water consumption, climate conditions (CO2, NH3, humidity, temperature, illuminance), audio indicators of respiratory issues, RFID tags and behavioral cues like drinking, eating and abnormal activity.
Our study was conducted in six compartments, each containing two pens with thirty weaned piglets. These piglets were observed over a six-week period, from post-weaning to when they reached an average weight of 25 kg. Data from multiple production cycles were collected. In addition to automated sensor data, manual observations were made to record key indicators of reduced resilience, such as tail and ear damage, post-weaning diarrhea, and respiratory distress.
The central aim of this research is to identify patterns in the collected data that signal early signs of diminished resilience. By integrating behavioral and environmental data, we aim to detect potential problems before they escalate. Anomaly detection techniques will be applied to improve the system’s predictive capabilities. The outcomes of this research will support informed decision-making, leading to improved welfare for pigs and more effective farm management in intensive farming operations.