WellBEEing: IoT monitoring of bee colonies in the presence of external stressors

Beekeeping has been popular in Slovenia for centuries, with multitudes of practices and traditions that have been inscribed to the UNESCOs list of intangible cultural heritage in 2022. With pollination, bees provide crucial ecosystem services and make the highest contribution to agriculture worldwide. Bee products such as honey, wax, pollen, propolis, and royal jelly have always been important to humans. Careful management of a beehive is crucial for successful apiculture, however, each unnecessary inspection induces stress to the colony and may spread pathogens. Nowadays, beekeepers are mainly aiming to optimize the gains at the expense of colony fitness. Precision beekeeping, with the use of digital technologies, is a new and emerging topic, increasingly gaining prominence. Professional beekeepers who choose to use technology mainly rely only on one metric, the weight of the beehive, which is a good indicator of honey flow evolution. However, the main breakthrough is expected from the use of artificial intelligence (AI) for analysis of multi-sensor data (including temperature, humidity, etc.), with the AI algorithms being able to recognize patterns in the data and either identify different types of events or predict their occurrence. Here, the concepts of trustworthy AI are becoming increasingly important, in particular in view of explainability, where methods of explainable AI can assist the decision support and are used for knowledge discovery.    Today, domesticated bees are increasingly suffering from various stress factors which cause the colonies to die or become less productive. These factors include physical (beekeeper interventions, crowding colonies, weather, light, noise, and electromagnetic pollution), biological (parasites and pathogens, such as varroa mite, nosema fungus, numerous viruses, as well as absence of queen), and chemical stressors (toxic metals, acaricide and pesticide residue, microplastic).  In this highly interdisciplinary project, we will bring together the experts from the fields of computer science, physics, chemistry, veterinary science, beekeeping, and electronics. The main objective is to design and apply a multi-sensor intelligent system to monitor the behavior of 20 bee colonies at four locations in the presence of external stressors. Specifically, we will focus on manipulation of the bee colony by a beekeeper, infestation with the common bee parasite varroa, and the effect of pesticides and toxic metals. We will also measure the biomarkers for oxidative stress in bees. We will advance beyond the currently available passive monitoring systems with data analysis using explainable artificial intelligence (XAI) to recognize different events in a beehive remotely. We will build on this knowledge to assess the impact of external stressors on the colony. Finally, the major fundamental scientific novelty of this project will be the new apiary knowledge connecting the impact of external stressors with the behavior of a bee colony, including health, stress, and productivity.   The project will create a novel multi-sensor platform that will be of major importance to both the researchers (allowing breakthrough studies in particular aspects of beekeeping, such as queen rearing, bee medicines development and testing, ensuring equality of colonies during research projects) and at a later stage to beekeepers, by providing insight into the colony without manual and stressful intervention in the colony and more efficient beekeeping, thus improving animal welfare. Some of the results of this basic project will be possible to be exploited commercially at a later stage, by integrating them into consumer products. Potential impact of the project includes a new approach of tracking the quality of bee products and using bees as a model organism to assess the burden of chemical impact on the environment.