DARwIn-OP: A Smart Elderly-Assistant and Companion

The alarming rate in increase of fall in elderly population has motivated the scientific community for the last two decades to develop autonomous assistance, while preserving their self-reliance, privacy and independent activity. The recent advancements in Internet of Things (IoT) and machine learning has improved the autonomous assistance features a lot but still there remains lacuna of research and efficiency, like accurate indoor localization, fall detection and secure human-robot interaction. This paper proposes a complete Robot Operating System (ROS) based architecture on a humanoid robot Dynamic Anthropomorphic Robot with Intelligence-Open Platform (DARwIn-OP) in collaboration with other intelligent entities like wearable tracker(UbiTrack) or Apple watch for localizing and a roomba to map environment for robust obstacle avoidance to respond immediately to a falling SOS by reaching the person autonomously, validating the SOS through robust fall recognition and interacting with the person if possible regarding the requirement of assistance and accordingly take action like calling emergency and next to kin for faster response.