Can we manage the aggressive behaviour of autonomous vehicles?

The connected and automated vehicles will have to cope with other automated or not vehicles and connected or not road users (drivers, riders, pedestrians) and a priority negotiation will always be necessary, especially at junctions. The behaviour of both the drivers and of the autonomous vehicles is inevitably dependent to all the other road users. Traffic behaviour can never be continuously stable, it should always be interactive and changing depending on the traffic behaviour of the other road users.

Consequently, as we have more polite or more aggressive drivers we will inevitably have more “polite” or more “aggressive” machines (autonomous vehicles) and this will be reflected though their programming.  As in non-autonomous traffic, in situations with one “polite” and one “aggressive” vehicle the aggressive gains priority, in situations with two “polite” vehicles conflicts are minimised and in situations with two “aggressive” vehicles, potential conflicts are increased. Similarly, the design and the management of the road and traffic system can be more or less aggressive, guiding road users and autonomous vehicles to respective more or less aggressive behaviours. The related repercussions on traffic safety are very important and focussed advanced research in the field of managing the “aggressiveness” of the road and vehicle systems is necessary.

Within this framework, there is need to “train” and “educate” the connected and automated vehicles as well as the respective road infrastructure and traffic, in a way equivalent to the one that we train and educate the drivers, the riders and the pedestrians. Data driven driver behaviour monitoring will be necessary to support the development of data driven machine behaviour. The development and programming of the machines’ behaviour entails an optimisation of traffic safety and efficiency, and the best field to bring in experience is current non autonomous traffic through detailed monitoring of all driver and traffic behaviour data, through extensive and targeted research.

Contribution at the ERTRAC Workshop on Future Research for Road Transport Safety & Security, Brussels, May 2017