Experts' opinion on how to validate an individual welfare monitoring system for fattening pigs

The pig production sector today is a fast-developing and challenging sector. Pig farmers try to improve production while keeping animal health and welfare standards high. For the IOF2020 project, the Pig Farm Management use case focuses on the use of Internet of Things (IoT) technology in pig production. Automated, individual monitoring of the health and production of fattening pigs is an example of IoT technology which can help pig farmers to optimize their time management and resources by helping them to identify pigs with health problems. Additionally, health, welfare and also productivity can be increased via early treatment of individual pigs where problems otherwise might be detected late or not at all. An additional aim of the use case is to develop an individual pig monitoring dashboard. To develop this, feeding and drinking data of individual pigs have been collected using RFID eartags to identify each individual pig in combination with RFID antennas at the feeders and the drinkers. Three fattening rounds have so far taken place, each with 120 pigs, divided into eight pens. Algorithms were designed to detect abnormal deviations in individual pig’s behaviors, which then result in an alert. The process of designing, testing and validating the detection system is a very challenging task which includes many decisions to be made. During an interactive session which will be held at a PLF workshop organised by the University of Copenhagen coming April, expert opinions from the participants will be collected. During this interactive session, using example cases, different steps in this process will be visualised and discussed such as: what problems should this system detect, how to categorize health problems whilst also considering severity, how to handle the observers’ bias, time difference between alerts and observations, etc. These expert opinions will be used to identify challenges, investigate the different approaches to validation and identify possible needs of the potential users of a detection system. The results of this workshop will be presented at the conference.

ISBN:978-90-8686-339-6