Adaptive wavelet packets for minimization of error in computation of mental stress

Abstract

This paper describes a study on evaluation of mental stress using adaptive wavelet packets filtering techniques. For mental stress to become chronic or if it remains for longer periods of time, then it can cause or trigger diseases such as hypertension, diabetes, insomnia, depression etc. There is a need for objective assessment of stress and develop counter conditions to relieve the body from stress. We have designed a wireless wearable sensor platform that can be used to capture a vast range of biomedical signals including heart rate, skin conductance and respiration rates. As stress is a reactionary phenomenon and its response vary from person to person, so an adaptive wavelet filtering packet is proposed to minimize intra subjective variations. The filtering technique has two folded effects; it reduces the classification error and it provides tuning to adapt it to users requirements. The wearable sensor platform is based on ease of use and a subject can be monitored for days and changes in his/her physiological conditions are recorded. The main advantage in monitoring stress is that if required in needs than necessary actions can be taken to remove stressful conditions.