Evaluation of Takeover Time Performance of Drivers in Partially Autonomous Vehicles Using a Wearable Sensor

Abstract

Advanced driver assistance systems and conditional automation (SAE level L3) bring full of challenges, central to the human role in such automated driving and are the transitions from automated to manual driving mode: driver monitoring systems should be able to adapt to the detected human state adjusting time given for takeover request. In this study, the human driver’s control authority during the transition from the automated system is evaluated by using a large set of sensors, including the galvanic skin response sensor, the speed, and the current sensor of the electric vehicle. The Local Outlier Factor (LOF) method detects the driver’s readiness to take control during the control authority transition transient stage. Eighteen drivers (8 females and 10 males) between 24 and 65 years of age participate in the experimental evaluation and data collection campaign. Takeover request time analysis evaluates each driver and three driver categories, such as experienced, semiexperienced, and inexperienced, and the questionnaire validates the takeover request time analysis. Galvanic skin response measurements and the LOF approach illustrate successful detection of the driver’s status about his/her readiness to take over and regulate the driving task when the transition of driving authority from the automated system is requested in a real road driving scenario. The takeover request time is personalized for each driver, which may improve the current conditional automated driving technologies’ penetration and acceptance.