A Robocentric Paradigm for Enhanced Social Navigation in Autonomous Robotic: a use case for an autonomous Wheelchair

The rise of autonomous technologies with unparalleled accuracy is revolutionizing computing and robotics by integrating machine learning techniques. This study focuses on advancing social navigation in autonomous robotics by improving object detection methods. We have refined the classification of objects within social environments into four distinct categories: living dynamic objects, non-living dynamic objects, living non-dynamic objects, and non-living non-dynamic objects. This differentiation in social navigation enables robots to process and respond to social cues, fostering a harmonious coexistence between humans and machines in shared spaces. Furthermore, we have introduced an adaptive proxemic zone surrounding these objects to define the boundaries for interaction. This concept, borrowed from human sociology, is instrumental in developing socially aware robots that respect personal space and societal norms. The proxemic zone is a buffer that helps mitigate potential conflicts or uncomfortable situations during humanrobot interactions. The efficacy of our approach is validated through results presented herein, which lay the groundwork for the development of socially intelligent robots that can seamlessly integrate into human environments and interact with people in a more natural and empathetic manner.