A new AI-based system has emerged that helps robots monitor patients and help them with everyday tasks. Such devices will soon appear in hospitals and nursing homes.
Researchers from Monash University, JDQ Systems, and the University of British Columbia have developed a service robot to assist nursing home residents or patients in healthcare facilities. In an article previously published on the arXiv portal, they presented a computing system that allows a robot to identify and track people in its vicinity.
Scientists have provided the robot with the social robotic platform Aether, which will provide daily assistance to the staff of hospitals, nursing homes, and auxiliary institutions. Engineers have identified several important skills for the device – escorting to a canteen for a meal, following it to a location, or even playing with physical activity. All of these skills require the robot to be able to identify and locate people and constantly follow them.
Scientists have developed facial recognition tools that allow robots to recognize people and anonymous human-tracking technology that helps robots track a person’s movements without knowing who they are. But in order to track specific people, the robot will need to simultaneously identify who they are and track their movements.
“Our algorithm is able to deal with defects, very poor or temporary lighting, even if it loses its target, the robot can restore its settings and find the person it needs to monitor. All this creates great difficulties for ordinary robots, ”the researchers noted.
The researchers tested their method in a series of experiments in which the robot had to identify, track, and follow users during five different scenarios. They captured the position of the robot and people in its vicinity using Vicon’s motion capture system. Initial trials by the researchers yielded good results, with the new technique surpassing existing facial recognition and user-tracking tools it was compared to.