GSB 7.0 Standardlösung


This website uses cookies. These are used for intermediate storage during ordering or registration processes. Data such as frequency of use or behaviour are not recorded. Here you can find out more about data protection and possibilities for contradiction.


Cooperative highly automated driving (Ko-HAF)

Ko-HAF aims at safe, highly automated driving at higher speeds: The driver no longer has to permanently monitor such systems. He can devote himself to other tasks, but must always be in a position to take over the control when requested, with a certain time reserve. To do this, the vehicle's foresight must go further than is possible with its own sensors. This is where Ko-HAF comes in: Vehicles send their environment information via mobile radio to the so-called Safety Server. There, this information is collected and compressed so that vehicles have a highly accurate, up-to-date map at their disposal, which provides the required better forecast in the sense of an artificial horizon.

the picture shows the logo of the Ko-HAF project

The project has been completed. BASt and a further 15 partners from industry, science and the public sector have achieved a lot in the three-and-a-half year project period.

Results at a glance:

  • The results show that collective perception and communication between vehicles and the safety server (backend) can significantly enhance environmental perception. This creates the foresight range necessary for highly automated driving.
  • Algorithms were developed for the aggregation of vehicle fleet data and for the continuous updating of back-end HD maps based on this fleet data.
  • Various manufacturers and suppliers have developed and coordinated joint concepts for the interaction of vehicles with a backend and for the meaningful fusion of sensor data and backend HD maps.
  • Ko-HAF has worked closely with SENSORIS - a consortium developing a standard for the exchange of information between vehicle sensors and backend solutions - and has made valuable contributions to the specification.
  • Highly automated driving functions for driving on the motorway, such as threading, overtaking, shutting down or emergency stop, have been developed. They include robust environment recognition, real-time trajectory planning and automated control of drive, brake and steering.
  • The partners conducted a total of 33 empirical studies with 1,723 participants and 1,750 hours on the role of humans in highly automated driving. More than thirty scientific publications were produced on this topic alone.
  • Tests have shown that sleepiness and fatigue can occur quickly during automated driving and are subject to changeable changes. Drivers showed strong inter- and intraindividual differences in the development of drowsiness and fatigue. Non-driving activities during automated driving were also the focus of the experiments.
  • For planned takeover requests, it was shown that multi-stage HMI concepts accelerate the termination of non-driving activities and thus the takeover time. A preview of planned takeover requests along the route (based on information from the safety server) helps drivers to regulate non-driving activities themselves.
  • Together, methods and tools were developed for efficient experimental testing of highly automated driving functions in simulation, on test sites and on motorways in public road traffic.

The project has a total budget of 36.3 million euros. It was funded by the Federal Ministry of Economics and Energy with 16.9 million euros as part of the "New Vehicle and System Technologies" programme.

The Federal Highway Research Institute is involved in the work packages 1 “Recording and representation of the environment in the back end (safety server)“ and 3 “Cooperative driving and controllable automation“.