Smart Connected Glasses for Drowsiness Detection: a System-Level Modeling Approach

Road safety applications based on embedded systems are becoming increasingly reliable for the assistance of vehicle drivers. Cars start to be equipped with applications such as the detection of falling asleep at the wheel. The Ellcie-Healthy start-up is developing smart connected glasses, a wearable device designed for e-Health and driver safety applications that embeds a driver drowsiness detection application. As size and cost are critical, the capacity of the battery on this wearable device is very limited. In consequence, finding the best compromise between the autonomy of the system and its performance is a challenging task. Therefore, estimating both the power consumption of the device and the QoS of the application for a particular system configuration has to be performed early in the design flow. In this paper, we propose a system-level modeling approach based on analytical power consumption models to estimate the autonomy of the connected glasses while considering the quality of the resulting sensor measurements. Obtained results demonstrate the benefits of our system-level power modeling approach to find the best trade-off between QoS and autonomy.