No abstract

Modern automobiles include an increasing number of assistance systems to increase the driver's safety. This feasibility study investigated unobtrusive capacitive ECG measurements in an automotive environment. Electrodes integrated into the driving seat allowed to measure a reliable ECG in 86% of the drivers; when only (light) cotton clothing was worn by the drivers, this value increased to 95%. Results show that an array of sensors is needed that can adapt to the different drivers and sitting positions. Measurements while driving show that traveling on the highway does not distort the signal any more than with the car engine turned OFF, whereas driving in city traffic results in a lowered detection rate due to the driver's heavier movements. To enable robust and reliable estimation of heart rate, an algorithm is presented (based on principal component analysis) to detect and discard time intervals with artifacts. This, then, allows a reliable estimation of heart rate of up to 61% in city traffic and up to 86% on the highway: as a percentage of the total driving period with at least four consecutive QRS complexes.

Capacitive biopotential measurements suffer from strong motion artifacts, which may result in long time periods during which a reliable measurement is not possible. This study examines contact electrification and triboelectricity as possible reasons for these artifacts and discusses local triboelectric effects on the electrode-body interface as well as global electrostatic effects as common-mode interferences. It will be shown that most probably the triboelectric effects on the electrode-body interface are the main reason for artifacts, and a reduction of artifacts can only be achieved with a proper design of the electrode-body interface. For a deeper understanding of the observed effects, a mathematical model for triboelectric effects in highly isolated capacitive biopotential measurements is presented and verified with experiments. Based on these analyses of the triboelectric effects on the electrode-body interface, different electrode designs are developed and analyzed in order to minimize artifacts due to triboelectricity on the electrode-body interface.

A novel approach is presented for non-constrictive long-term-monitoring of respiration which could particularly become suitable for home care applications. The system is based on textile integrated force sensors which detect expansion of the thorax during respiration and allow wireless data transmission for maximum mobility. Possible applications include long-term monitoring of patients with chronic pulmonary diseases, early recognition of diseases and the performance measurement of athletes during exercise. Results from performance tests under various conditions are presented. The tests show a high correlation with measurements made with a flow meter as gold standard, allowing conclusions about the respiratory volume. Different respiration patterns are reliably detected and can easily be subdivided even during intensive motion of the person. The sensors' textile buildup may allow a future integration into clothing and thus a minimal constriction for the patient. Nevertheless, the respiratory volume estimations are still not accurate enough mainly because of the use of a simplified volume calculation model and the disregard of the abdominal volume changes.