Robot-Animal Interaction Using Emotional Perception and Behavioral Generation

Robotics has become a crucial tool for the study of animal behavior, showing promises in the discovery of novel traits in organisms. It is of great scientific significance for utilizing biomimetic robots to enable autonomous behavioral interactions in animal experiments. Thanks to the development of information technology, existing biomimetic robots are able to perform basic behavioral interactions. However, state-of-the-art interactive robots still struggle to convey multi-level, heterogeneous information within biological systems, making it challenging to effectively mediate the complex interaction process. This report focuses on the research progress of the research team at Beijing Institute of Technology (BIT) in the areas of robot-animal interaction, which mainly includes emotion perception and behavior prediction in rats, data-driven robot social behavior generation and planning, object detection and tracking control in robot-rat interaction, and modulating behavioral emotional states of rats during the interaction process. These studies hold great potential for applications in the field of neuroscience (e.g., animal behavior research).