Assessment of Driver Situation for Control Authority Transition from Conditionally Automated Vehicles using Chassis and Galvanic Skin Response Sensors

Abstract

The authority transitions are important when human interact with automated driving. The monitoring systems should be able to smartly adapt to the detected driver state, adjusting the time given for take-over requests (TOR). The proposed system in the study obtains drivers’ ideal driving authority takeover times by analyzing wearable sensor and other sensor data. The driver’s authority during the transition is evaluated in this study using the sensors including the chassis velocity sensor, galvanic skin response (GSR) sensor, and current (torque) sensor subjected to longitudinal quality metrics. Three different traffic situations are analyzed to compare four different takeover times (0s, 2s, 4s, and 6s) and the ideal TOR times of drivers are detected for the authority transition. Here, TOR 6s is close to a critical/dangerous situation and TOR 0s time is a sudden transition. According to the results, ideal TOR times are mostly TOR 2s and TOR 4s. TOR 6s is not considered as an ideal TOR time for any driver in the study.