Performance Analysis and Detection of Stress using IoMT

The stress is the common thing that a person has to deal with it every day which may arise according to their daily activities. The stress level increases when the things are not occurring according to the particular person's plan, which creates tension it may affect mentally. If this is frequently noticed within a person may lead to abnormality and affect the person physically and mentally. To avoid this various stress detection mechanisms and systems were introduced. Since there are n number of frameworks that were existing each with specific methodologies. This paper proposes some new sort of techniques and mechanisms which has higher probability in detection and user friendly than compared with all other existing systems. In this paper, to determine the stress value, the person is observed under three conditions. (1) The stress value is observed at normal conditions. (2) The stress value is observed by inducing the person with creating an illusion through illusion clip. (3) The stress value is observed by playing an aesthetic clip. The parameters that were considered during these three conditions are heart rate and working condition. The heart rate is calculated by the common method Electrocardiogram (ECG) and the working condition is directly fetched from the personal details of the person. Age is also considered as an input since it is also a common factor for stress affection and from various researches, it has been concluded that the person between the age of 25 and 45 has more stress. Some of the assessments namely CAGE questionnaire and FTND assessment also taken regarding the person's alcoholic and smoking habits respectively. After noting all these parameters, it is processed with a machine learning technique known as Residual Neural Networks. The enhanced algorithm used in RNN is efficient for obtaining accurate information from the input. Then it is compared with the general values. Hence the stress value is identified as low, medium, or high according to the input parameters. The observed readings were uploaded into a Arduino cloud platform so that it will be helpful to compare the particular's previous stress values and easier for handling the data for the same.