Real-time Non-intrusive Detection of Driver Drowsiness - FY08 NATSRL

Principal Investigator(s):

Xun Yu, Former Associate Professor, UMD-Mechanical & Industrial Eng

Project summary:

Driver drowsiness is a major cause of serious traffic crashes. Continuous monitoring of drowsiness is therefore a critical factor in the ability to reduce crashes resulting from it. This research project aimed to develop a real-time, non-intrusive driver drowsiness detection system. Biosensors were built on the automobile steering wheel to measure a driver's heartbeat; this enabled the collection of heart rate variability (HRV) data, a physiological signal with established links to waking and sleeping stages, which were then analyzed to detect driver drowsiness. This novel design of measuring heart rate from biosensors on the steering wheel means the system will cause little annoyance for the driver, while the use of physiological signals will ensure the accuracy of the drowsiness detection. To date, the researchers have developed the heart-rate measurement systems both on the steering wheel and on the back of the driver's seat. An adaptive noise cancellation system has also been developed for canceling baseline measurement noise. Both systems have been tested in the lab with several subjects and for different situations, such as the measurement on the seatback with different cloth materials. Researchers are currently working on the HRV analysis and building a driving simulator. In the second phase of this project, human subjects will be recruited for lab tests, and researchers plan to prototype the system and install it on an actual vehicle for real-world data collection.


Project details:

Reports or Products: