Identification and prediction of stress using existing data processing methodologies are incompetent to predict the stress either in real time or laboratory based experiments. The main aim of this work is to classify the stress. The appropriate signal processing methodology, (i) to analyze the Electrocardiogram (ECG) signal for normal and stress condition and (ii) to derive the optimum features from a set of statistical features over frequency band. This paper represent method to find out

more »

... condition. Electrocardiogram, were recorded continuously while drivers followed a set route through open roads in the greater Boston area. Stress levels are created by driving into different condition. By driving in city, highway and rest condition. We used physionet data base 'Stress Recognition in Automobile Drivers' for ECG signal.ECG signal is used to analysis the stress. Parameters ( RMS, Max ,Std ,kurtosis , min ,mean ,variance)are used for analysis by using discrete wavelet transform by using 'db6' mother wavelet function . Entropy Correlation, Skew ness, median, crest factor are not giving the appropriate changes from normal condition to stressed condition so these parameters are not considered. Artificial neural network was used for decision making you are in stress or not.