KITT - Artificial Intelligence for the improvement of the safety of tunnels and tunnel control centers

Whether through mountains, rivers or oceans - in Germany alone, traffic runs through more than 420 road tunnels. They connect countries and facilitate travel by eliminating tiresome detours. As traffic increases and the digitalization of modern vehicles continues , safety requirements for road tunnels are changing as well. New technologies such as connected and autonomous vehicles are becoming more sophisticated, thus requiring a more extensive data analysis. With regard to tunnel safety, this results in numerous new opportunities, but risks as well.

Aims and approach

KITT stands for "Artificial Intelligence for the Improvement of Safety in Tunnels and Tunnel Control Centers." Safety in road tunnels is achieved by using a monitoring and control system provided by tunnel control centers and accompanying information and communication technologies (ICT). In the wake of increasingly complex tunnel and IT security requirements , Germany and Austria have joined forces for the KITT project to provide better and more innovative security solutions for road tunnels by means of artificial intelligence (AI) . Using AI vast amounts of data are analyzed in real time. Since the AI is not intended to replace the human operator in the tunnel control center, but to support him, a weak artificial intelligence is used.

The aim of the project is to be able to react to incidents and dangerous situations in tunnels early on. The main focus is on rescuing people, preserving the structure and maintaining the traffic flow.

Safe tunnels through artificial intelligence

Road tunnels are equipped with sensors and video cameras that supply data to tunnel control centers. In the future, it will be possible to transmit information from vehicles to the tunnel infrastructure via interfaces. This is referred to as car-to-infrastructure communication (C2I). The large volumes of data generated are to be evaluated by an AI, particularly with regard to possible anomalies, and recommendations provided to the human decision-maker. In the future, communication between infrastructure and car (I2C) may also be possible.

Data security requirements

If the analysis of data by AI primarily focuses on the assessment of the security situation (traffic and tunnel security, IT security, civil security), questions concerning the use of personal data arise. The aim of Professor Georg Borges' chair is to study data and IT security. Emphasis is placed on the possibilities of compliance with data protection law through anonymization, pseudonymization, privacy by design and deletion obligations. Furthermore, it will be determined who is entitled to the rights of the data used and to what extent this must be taken into account during data processing. Finally, the handling of possible errors is examined. This ranges from the requirements of error control to the question of which of the actors involved in the complex infrastructure is liable in the event of damage. To approach these issues, the legal framework is identified from the perspective of public authorities.
Since this is a transnational project, the differences between German and Austrian legislation will also be taken into account.