Robots adapting to dogs: a new frontier?

Abstract

While Human-Robot Interaction (HRI) has seen extensive exploration, Animal-Robot Interaction (ARI) remains a less mature field.This paper presents a first AI-based prototype designed to enable a humanoid robot to recognize emotional and postural states in dogs and adapt its behavior accordingly.Using a deep learning-based pipeline for real-time detection and classification, the robot could adapt its movements to better accommodate canine responses.We propose that such an adaptive approach paves the way for more natural coexistence between robots and animals in domestic settings, raising new challenges in perception, behavior design, and ethics within ARI.