Automated real-time stress monitoring of police horses using wearable technology

Mounted police horses and riders are repeatedly subjected to demanding and stressful situations. Intensive selection and training of police horses is required to ensure performance, safety and welfare of the horses and their riders. At the mounted police in Brussels, Belgium, the selection of police horses is mainly based on intuition built upon previous experience which makes it a subjective decision. Although this decision mostly leads to good results, sometimes horses are purchased that turn out to be unsuited for the mounted police. Including an objective measure when evaluating the longer term suitability of candidate police horses could help the mounted police to further improve their selection procedure for police horses. This study investigated the potential of real-time stress detection as a measure to evaluate the longer term suitability of police horses. Four experimental protocols were developed in consultation with the mounted police and were performed by 17 horse-rider pairs. The horses were divided into four categories according to their experience and suitability as police horses: good beginner, bad beginner, good experienced and bad experienced. The relative stress of the horses was monitored with wearable technology during every protocol. For one protocol, the time percentage spent over 20% relative stress by the horse was found to be significantly lower for good compared to bad beginner horses (p-value = 0.0277). In conclusion this study demonstrated that real-time stress detection with wearable technology in mounted police horses provides information on the longer term suitability of police horses.