Our smart nursing cart uses 2D LiDAR sensors and general RGB cameras for following, autonomous navigation and environmental recognition. Our smart nursing cart does not need to rely on any Bluetooth or beacon to operate, and during the ward rounds, if an abnormal event is detected, the monitoring platform will issue an alarm to inform the nursing staff. In terms of health education communication, we have a customized health education information system.

This project develops a posture recognition method to detect whether an abnormal event has occurred. The feature extraction part will use skeleton features with lower dimensions to improve the efficiency of subsequent classification processing. The feature classification part plans to design a merged fuzzy neural network, which not only provides relatively fast calculation speed but also can achieve good training results on small data sets.

The smart nursing cart is designed to assist nurses in completing multiple tasks, including autonomously making rounds or patrols and transmitting real-time images back to the nursing station. In addition, through the voice command translation system, nurses can combine with AI voice assistants to complete nursing data input operations and easily complete shift handovers.