Automated vehicles in which the driver can allow the vehicle to do all the work on certain parts of the trip and no longer has to monitor anything are currently being developed by many automotive manufacturers. When there are driving tasks that these highly automated vehicles can no longer handle, control must be returned to the driver. The driver must have sufficient time to take over manual control of the vehicle safely and easily. An empirical study was designed and conducted to find out when full physical and cognitive control over a vehicle was reestablished after a phase of highly automated driving.

For that reason takeover scenarios and secondary tasks of varying complexity were developed and tested in a static driving simulator with 60 subjects aged from 20 to 76. The effect of a driver being “out of the loop” was analyzed, in particular. In some experimental conditions, the driver was thus completely uninvolved in the task of driving, and distracted by a secondary task, at the time of the takeover request.

After a drive with a high level of distraction, 90 percent of the drivers looked at the road again for the first time after 3-4 seconds, had their hands on the steering wheel and their feet on the pedals after 6-7 seconds and had switched off the automated system after 7-8 seconds. However, if you look at the first glance at the mirror and the glance at the speedometer as indicators of awareness of the driving situation, you see that 12-15 seconds are required. These reactions, which are required in order to understand the current traffic situation, are thus delayed by up to 5 seconds compared to the reactions of drivers in manual control in the same situation. When the reactions of monitoring drivers and drivers with the maximum level of distraction are compared with those of drivers in manual control, there are generally delays in all reactions. This was clearest in the case of distracted drivers. But even drivers who were monitoring the situation (partially automated driving) had delayed reactions compared to drivers in manual control.

It should be noted here that the times were ascertained in a driving simulator and thus can only be understood as an approximation of the time required for a takeover in a real vehicle. However, more recent studies do indicate that times ascertained in a simulator correlate well with those in a real vehicle. It should also be emphasized that the takeover time alone cannot be an adequate measure of the quality of a takeover. The times must always be seen in connection with other measures of the quality of the takeover. This include the quality of the safety of the takeover and the fullness of the driver's awareness of the situation during the takeover.

The takeover process is one of the key challenges in automated driving. It is possible to create a safe takeover when certain findings will be considered.