312: Drowsiness Detection Mirror

Hundreds of thousands of deaths occur every year due to drowsy driving. With assisted driving technology improving each year, reducing the tragedies of drowsy driving is becoming more likely. Our drowsiness-sensing rear-view mirror will enable car manufacturers to predict when the driver is sleepy and enable assistive driving features. 

We used a radar system inside the mirror to record data from the driver. The radar emits a wave and calculates the driver’s breathing rate and head tilt from the reflected wave. A slower breathing rate and a driver’s head tilting down for an extended period suggests they are getting drowsy. The device uses signal processing to extract these physical signals. Next, the data is fed into an algorithm which predicts if the driver is falling asleep due to changes in their body. Once the algorithm estimates the driver is falling asleep, it alerts the vehicle’s operating system to activate safe driving measures. We conducted tests at times when there is a high risk of feeling sleepy, such as at night. We used a chest belt to record breathing rate, while head angle was observed directly. Our mirror was able to detect drowsiness about a third of time. Since every person has unique sleep patterns, personal calibration could be introduced in a future design to adjust for this. Also, other factors that indicate drowsiness, like heart rate, could improve the accuracy of the model.