Automatic Generation of Mathematical Models of Arthropods

Arthropods are susceptible to abrupt climate changes and modify their organizational behavior, based on ambient humidity and temperature. Miniaturized technologies allow for continuous and detailed observation of animal behavior, as well as the extraction of environmental data. Furthermore, software using Artificial Intelligence techniques such as scene graph generation (SGG) can obtain descriptions of the acquired scenes of animal life. In this research, miniaturized hardware and intelligent software are used to generate a mathematical model of the organizational behavior of bees and ants, based on videos, images, and environmental data extracted in real time from the entrances to these arthropod nests. The nests used for the experiments are located in different geographical positions. For each experiment, a dynamic system is generated, and its accuracy is verified using visual information and data obtained from the nest entrances, which demonstrate the arthropods’ behavior. We measured the similarity between the dynamic system produced and organizational behavior using five evaluation parameters and experimental tests conducted in the field.