Abstract
Nowadays, robots are playing a vital role in healthcare applications to provide patients support and assistance in critical situations. The robots are trained by artificial intelligence systems which help to learn the robot according to their patient needs. However, the robots require the medical staff while diagnosing diseases with maximum accuracy, remote treatment, paralyzed patient treatment and so on. For these precise and accurate issues, an intelligent learning process is applied to train the robot to support the patient's mental health and related task in this work. Initially, the patient health details are collected along with their simple daily routine, medical checkup information and other healthcare details. The gathered details are processed with the help of a deep reinforcement learning process used to get important information. The learning approach uses the state and action process to determine every patient's needs and the respective assistance. Based on the information, robots are trained continuously to keep patient positive attitudes in their mental health problems. The excellence of the system is evaluated using experimental analysis in which the deep reinforcement system ensures a 0.083 error rate and 98.42% accuracy.