Skadi: Heterogeneous Human-sensing System for Automotive IoT

Abstract

Over the past years, cars' computing, sensing, and networking capabilities have rapidly increased, and the automotive development aims for autonomous driving. However, the driver is still the focal point for decision making. It has to be alert at all times to avoid traffic accidents due to human factors like tiredness, inattentiveness, and intoxication. Therefore, there is a need for a system that monitors the driver and intervenes before human failure can have a negative impact on traffic. A variety of commercially available wearable IoT devices, such as smartwatches, bracelets, and rings, are capable of monitoring human health conditions. However, those devices come with technological differences and manufacturer-specific implementations. This paper proposes a prototype for a humansensing and health monitoring system based on wearable sensor devices. The aim is to find a solution that ignores the technological heterogeneity of IoT devices and generalises their implementation into the automotive system. Consequently, the data should be available to be analysed together with the data collected from the vehicular sensors. Our solution is compatible with open-source platforms Eclipse Hono and Kuksa.