Resuming Manual Control or Not?: Modeling Choices of Control Transitions in Full-Range Adaptive Cruise Control

Automated vehicles and driving assistance systems such as adaptive cruise control (ACC) are expected to reduce traffic congestion, accidents, and levels of emissions. Field operational tests have found that drivers may prefer to deactivate ACC in dense traffic flow conditions and before changing lanes. Despite the potential effects of these control transitions on traffic flow efficiency and safety, most mathematical models evaluating the impact of ACC do not adequately represent that process. This research aimed to identify the main factors influencing drivers’ choice to resume manual control. A mixed logit model that predicted the choice to deactivate the system or overrule it by pressing the gas pedal was estimated. The data set was collected in an on-road experiment in which 23 participants drove a research vehicle equipped with full-range ACC on a 35.5-km freeway in Munich, Germany, during peak hours. The results reveal that drivers were more likely to deactivate the ACC and resume manual control when approaching a slower leader, when expecting vehicles cutting in, when driving above the ACC target speed, and before exiting the freeway. Drivers were more likely to overrule the ACC system by pressing the gas pedal a few seconds after the system had been activated and when the vehicle decelerated. Everything else being equal, some drivers had higher probabilities to resume manual control. This study concludes that a novel conceptual framework linking ACC system settings, driver behavior characteristics, driver characteristics, and environmental factors is needed to model driver behavior in control transitions between ACC and manual driving.