Does automated driving improve safety?

Automated driving is regarded as the future of mobility. It is expected to make traffic flow more efficiently and reduce the number of road accident victims as well as emissions and traffic jams. This will be more of a multidimensional, gradual transition than a rapid change. What we can say today is that vehicles with different levels of automation will be sharing the roads with manually driven vehicles in the foreseeable future. This development will affect both cars and commercial vehicles.

In the coming years, systems will become available that will be able to drive autonomously for certain periods of time but will only be able to handle selected situations. This is referred to as conditional automation. The driver no longer has to monitor the vehicle continuously but does have to take control on request when the system limits are reached. The fundamental problem described here in connection with constant monitoring coupled with intervention in the event of critical situations is based on a human characteristic investigated by psychologists over 100 years ago. The resulting Yerkes-Dodson law describes the general relationship between a person’s ability to perform well and their state of physiological and mental arousal. When a person has a low level of arousal, their performance remains at a minimum level. As the person becomes more aroused, their performance increases up to a maximum level. If arousal increases beyond that, performance starts to drop again until it reaches a similar minimum level to the level at low arousal (figure 4). Put simply, this means that people perform demanding tasks best with a moderate level of arousal. Driving a car is such a task. Monotonous tasks, like driving down a perfectly straight road with no traffic, can result in a low level of performance or failure. Monitoring a Level 2 autobahn or freeway pilot is one such task. Equally, if a driver is overtaxed, the result will be poor performance and even failure. Suddenly being requested to take over control from a Level 3 autobahn or freeway pilot would be an example of this.

Conclusion:

Highly automated and autonomous vehicles could bring great benefits in terms of road safety if they functioned flawlessly under all conditions within their intended scope. Until such time as these systems come onto the market, drivers of manually controlled vehicles should benefit in terms of road safety from continual improvements in assistance systems. In this transitional phase, automated functions will nevertheless be offered that require the support of the driver. If these are not to represent a threat to road safety, they should have the key characteristics presented here. As a general principle, vehicles with automated driving functions should only be used on public roads provided they are not less safe than corresponding vehicles equipped with advanced driver assistance systems.