Next level of hospitalisation through smart ICU

Abstract

Automation technology is developing intelligent health applications to facilitate patients by providing smart health solution. However, plethora of research is required in the autonomous robotics industry to provide smart solutions to hospitals.Hence, the proposed study aims to develop an intelligent automated infrastructure for hospitals, capable of performing several smart tasks in an Intensive Care Unit (ICU). The developed system will make food and medicine approachable by using a robotic arm on an autonomous robot. Furthermore, robotic arm can be locally controlled by the patient paramedic’s staff. The proposed intelligent robot will monitor patient’s condition by automatically monitoring the vitals conditions such as sleeping, stress, discomfort etc. This proposed system is very useful for diseases where close proximity can spread the disease, such as the covid-19 situation. Moreover, the automation will aid ICU patients by controllable bed functionality supported with patient’s EEG signals or remotely by staff. This research proposed LSTM based neural network for EEG classification and compare the results with other machine learning algorithms. The proposed LSTM network achieve 94% accuracy on self generated dataset. The achieved results are also compared with other machine learning models like SVM and Multilayer Perceptron (MLP). The intelligent navigation feature is also introduced which enables the robot to move autonomously in ICU. In addition to this, it can also establish video conference set up between the patient, staff and family members. The robot can automatically alert the staff in an emergency and assist the patient through an intelligent chatbot.