AUTUKAR - Highly Automated Tunnel Monitoring for Disaster Management and Regular Operations

AUTUKAR, a project funded by the BMBF, aimed to develop an assistance system for operators at tunnel control centres. Previously, operators had to scan the video images of all cameras in the tunnel sequentially in order to assess the situation in the tunnel. By classifying and tracking vehicles in the tunnel with the help of image analysis methods involving several cameras and supplementary sensors at the tunnel entrance, we created a virtual representation of the traffic situation. This encompassed the development of an innovative, interactive user interface as a 3D visualisation that the operators could employ. The tracking data could also be used to automatically identify a variety of risk situations (traffic jams, broken-down vehicles, accidents, conspicuous driving behaviour) that were displayed accordingly via the user interface. This helped to reduce the workload of the operators and thus increased the safety of the overall system.

Following the approach of the user-centred development process, workshops and interviews were first conducted with operators and operators of tunnel control centres to analyse the requirements. The results were then used to a large extent in the planning and realisation of the user interface. User test were then conducted with the operators to evaluate the new user interface at the Tunnel Rennsteig in Germany.

In order to be able to test a large number of situations and, above all, incidents with the system without actually having to cause them in the tunnel, a traffic simulation was implemented as an alternative for the real-time tracking data. At the same time, this simulation provided the basis for a prediction model that enabled the tunnel situation to be displayed for a short time even if the video technology failed (e.g. in the event of smoke development).