The revolution of transportation is a lot closer than we think. In the foreseeable future, mixed human-driven and autonomous public road transport can be expected. Safety, comfort, and ecology of road traffic will depend on the interaction between automated and human transporters. New methods and competencies are needed by which the control of automated vehicles in interaction with other road users avoids or manages critical situations. The gradual introduction of automated driving functions into existing traffic requires a deep understanding of human road users' interactions, which is a significant challenge.

We should generate and gather much data representing a broad range of status and behavior features on vehicles and road infrastructure. CCAM Living Labs are needed, equipped with sensors for data collection, and vehicles with onboard data collection devices. A critical task is properly organizing this data for having the necessary, secured information where and when it is required.

That is why Mobility Platform decided to lead up the complete functional and logical plans of a large scale Traffic Data Platform of public road transport data. The functional requirements and architecture plan of an IT platform for gathering, managing, and analyzing such data was developed. Traffic Data Platform is aimed to ensure the availability of the data needed for transport within the framework of various projects, as well as the IT frame of the data exchange for the organizations and vehicles related to the projects.

‍

‍

Tasks of the Data Platform:

1. Collecting data in GIS, developing data processing models and algorithms that support efficient intelligent traffic management.
2. Support the preparation and expansion of autonomous transport.
3. Support real traffic management tasks in dedicated projects.
4. Contribution in local and international intelligent transport R&D projects.
5. IT environment for the development of new methods and algorithms in scientific co-operation.
6. Support of synergistic cooperation of related data providers.

The Data Platform can thus contribute to preparatory work for regulatory tasks and standardization, testing in simulation and real-world environments, efficient traffic management (e.g. better utilization of roads, optimization of transport investments, parking management), enhancement of knowledge base in Hungary and broaden international co-operation in this field. It can improve road safety by reducing the number of incidents, accidents, and increasing public confidence and acceptance of highly automated transport. The data platform supports the R&D of self-driving vehicles. In different projects, individual road users can share data with the platform about their position and the status of each of their sensors. The system may be linked to other traffic databases.

Data Platform incorporates a so-called Data Lake layer, which ensures the collection and storage of data assets originate from players in the transport arena. The analytical layer, connected to this Data Lake, hosts the development of AI models and algorithms based on the collected data. (Fig. 2.)

‍

‍

International and domestic data management best practices were reviewed to lay the foundation for the design. Data science and Artificial Intelligence (AI) solutions, for example, can investigate the consequences of blocking a roadway for other road loads, predict traffic on a particular roadway during holidays or bad weather, if the necessary data is available, delays in traffic schedules can be estimated.

Data analysis solutions can also save costs by predicting expected vehicle malfunction, optimum fuel consumption. Road quality control and deterioration forecasting can be made developing visual diagnostic tools. Linking relevant data from public and private companies (even including car-sharing and bike-sharing) opens up additional opportunities.

Data Platform can handle large amounts of data from multiple data sources to serve future projects' needs. A priority is to build new AI or data science applications as quickly and efficiently as possible over the collected data assets. The analytical layer should be an on-premise solution that provides maximum data security.  The AI based Data Asset Management System can serve data scientists and machine learning algorithms in the long run. IT should incorporate optimal AI runtime environment (like GPUs and FPGAs), to run the latest deep learning algorithms. GPS tracking, road, and traffic data can be analyzed with high efficiency, valuable and useful patterns can be searched for in visual and video signals, to mention a few options. GDPR requirements can be also met with IT security and personal data management solutions incorporated into the system. (Fig. 3.)

‍

‍

The concept allows the simultaneous work of several independent analytical or data scientist teams. The completed proposal also covers sophisticated system management services.

The Data Platform makes the collected data more valuable, as the artificial intelligence algorithms run on them will result in more efficient operation and better control methods.

‍