The ability to acquire vital parameters and classify cognitive conditions opens doors to new technologies in diverse areas such as medical technology, automation, aerospace, fitness/wellness and security. TROUT gained considerable expertise in biometric data processing during automotive and medical technology developments which focused on machine learning and AI (Artificial Intelligence). With variations in the heartbeat, the organism can respond optimally to changing endogenous and exogenous influences and thus adapt to the current needs of the blood supply. Heart rate variability (HRV) provides not only information on the degree of stress on the cardiovascular system, but also on the quality of cardiovascular regulation and has also become established in other areas in recent years, due to ever smaller measuring instruments and lower costs, as well as applications in clinical research.

Drive test in Maischbergers TV Show

Project Team

TV Hessenschau

TV Maischberger

The first processing step is to extract features from the RR intervals of the ECG signal using a sliding window. The RR intervals within such a window are used to calculate several features. Sliding Windows are characterized by two parameters: First the length of the sliding Windows and second the shift of the sliding Windows. The first parameter determines the number of values ​​or RR intervals that are taken into account when calculating a characteristic; the second parameter determines after how many values ​​or RR intervals a renewed feature calculation takes place. For each of the sliding windows ten features are calculated.

For feature selection, filter or wrapper approaches can be used in principle. Filter approaches evaluate features based solely on characteristics that can be calculated from the data. Wrapper approaches use a classification algorithm and evaluate features according to their contribution to its performance. The former approaches have the advantage that they are easy to calculate, while the latter have the advantage that they give very problem-specific results, but consume much more calculation resources.

Based on the protocols, data has been assigned to the different phases of the study and also to various levels of mental and physical stress. These classes are: neck exercise, relaxation and then on a driving simulator:  adaption phase, aggressive traffic during rush hour (Fig.3), math test and night time drive.

Heart rate based on the resting heart rate (green), ratio of pulse increases to pulse decreases (blue), HRV – Heart Rate Variability (light green), Event Marker (red – in German language), Calculated Stress Level (yellow) in a relative scale from 0 (no stress) to 7 (maximum stress) on the ordinate, the abscissa shows the time in seconds.