Car as Diagnostic Space (CarDS)

点击翻译按钮获取中文摘要

Early detection of symptoms is critical to detect diseases at an early stage. Therefore, continuous health monitoring integrated into private spaces such as the vehicle has the potential to detect diseases earlier. This enables a better treatment, decreases the mortality rate, and reduces costs in the healthcare system. On average, a German spends 43 min per day in a vehicle. Hence, we aim at building an in-vehicle sensor system that integrates a medical check-up into our daily mobility. We want to integrate a health monitoring system with multiple sensors to measure the electrocardiogram (ECG), Photoplethysmogram (PPG), remote PPG (rPPG), and phonocardiogram (PCG). Redundant sensor data increases the reliability of the data analysis. We will integrate the sensor system into the CAN-BUS system and consider the in-build sensors such as the accelerometer, capacitive sensor, and an external camera for the artifact’s detection. We will conduct a study with 20 test persons in different driving scenarios: rest, city, highway, and rural areas. After recording the data, we will develop a sensor fusion approach based on a convolutional neural network (CNN) structure. We will compare the reference heart rate with the calculated heart rate and evaluate the algorithms. Based on our data analytics, we will improve the sensor system and repeat the data recording. As a final step, we will evaluate the artifact detection and sensor fusion algorithm. We will answer open research questions such as, for instance, “What percentage of the driving time is usable for a reliable heart rate analysis?”. Results from this project are the basis to answer other, subject- as well as disease-specific research questions.