Three into five will go

Blog post posted on 12/06/19 |

I was really interested to read recently that experts at the German equivalent of the AA (ADAC) have suggested a new classification scheme to help drivers understand the different types of driverless car. 

This is important as in many of the surveys we see about attitudes to driverless cars the ‘don’t knows’ often outnumber the ‘pro and anti’ views. Consumers have some idea of what’s coming but for many it is a complex and difficult to follow debate. 

Among car engineers and software developers a five (or six depending on who you talk to) point scale has been in use for some time.  That’s all well and good, but a simple three-point summary could be much more effective. 

Understanding what your car can and cannot do on its own is vital for safety as we move towards growing levels of autonomy.

This is not helped by the names that some car makers may call their systems e.g. ‘auto pilot’, but in truth there are no truly driverless cars on our roads today or likely to be for several years yet.  The outcome is that some drivers can be fooled into thinking their car can do it all for them, or some early adopters may deliberately push the boundaries to see how autonomous their car can be on its own.  We have already seen where that can lead to with several fatal crashes in the USA and elsewhere. 

So the new system boils it down to the three A’s; assisted, automated and autonomous.  Most new cars are now ‘assisted’ in some form right now with new systems that can be a valuable aid to safe driving. 

Growing numbers of top end ‘automated’ cars are starting to come on stream with adaptive cruise control, cornering assist etc but which ultimately still require full human attention.

It is only when you get to the third level ‘autonomous’, that you can really switch of and let the vehicle do it all. Autonomous is still a long way off and sometime seems to be drifting ever further away as car makers and software experts struggle with the real world.

Below are the details of this less complex model and its modes of operation;

  • First operation mode: ‘assisted’ driving. Drivers are responsible for driving at all times, are liable for traffic offences and accidents. To some extent, the vehicle keeps in lane, brakes and accelerates, but there is no mandatory feedback, and the assist feature can be aborted at any time
  • Second operation mode: ‘automated’ driving. The vehicle drives autonomously only in a situation defined by the manufacturer, e.g. stop-and-go driving in traffic jams. Drivers may temporarily take their attention off driving and the road, but at the system’s request, they must take control without delay. Drivers are liable only if they ignored the request to assume control
  • Third operation mode: ‘autonomous driving’. By deciding to relinquish full control of the vehicle, the driver becomes a passenger. It is conceivable that the vehicle travels with no occupants at all. The system autonomously handles critical driving situations and it is the operator rather than the driver that must permanently monitor the vehicle to be able to respond to any disruption. There is no liability on the part of passengers.

In our view at IAM RoadSmart it is vital that we plan for extra training to ensure understanding of the capabilities of ‘assisted’ and automated’ cars and the differences between them. 

Without such training the systems full potential to save lives will never be realised and we will still see avoidable accidents happening. 

When we get to the ‘autonomous’ phase training will also be needed for ‘operators’ to ensure they keep these new vehicle types in top condition and functioning at their peak. 

It is good to see that expert bodies such as Thatcham Research agree that education is needed to help develop understanding of the plethora of acronyms and technologies that are being thrown at today’s driving consumer.  Check out their website for more information on

By Neil Greig, IAM RoadSmart director of policy and research