PREDICTION AND MONITORING OF ANIMAL HEALTH USING ARTIFICIAL INTELLIGENCE

Farm animals, which number more than 70 billion worldwide, are increasingly kept on large, intensive farms. Currently, there are no scientifically based “benchmarks” for quantifying transient physiological (affective) states in farm animals, and there are no established indicators, only indicators of poor condition, such as injury, pain, and fear. Traditional approaches to monitoring animal health are time-consuming, interrupt agricultural processes, and involve subjective judgment. Biometric sensor data supported by artificial intelligence is a smart new solution for covert animal monitoring, but its potential for quantifying physiological states and innovative solutions in its application have not yet been realized. This review presents innovative methods for collecting large amounts of data on the condition of farm animals that can be used to train artificial intelligence models to classify, quantify, and predict the physiological states of farm animals. By extending this to the group level, artificial intelligence can be applied to model emotional dynamics and contagion among animals. Digital twin models open up new data-driven approaches to modelling animal behaviour and physiological state. By displaying the information and predictions of the digital twin in an intuitive way on an interactive dashboard, caregivers will be better informed about their animals, enabling them to respond quickly and effectively to reduce pain and ensure animal welfare. Further research in this area can be aimed at developing and improving data analysis algorithms and expanding the range of measured parameters.