Current trends in vehicle technology foresee a change in the transport system of the future. The expected spread of autonomous vehicles will create the opportunity to influence the transport system more effectively. Costs related to transport processes can be reduced, traffic is expected to become safer, and a number of research tasks have become and become necessary.
In the course of our work, we sought the answer to how to characterize and improve the flow and safety of vehicle processes in an autonomous transportation system. Considering the goals and tools of traffic management problem of autonomous transportation systems, we developed a time and space discrete, binary variable optimization model (Pauer & Török, 2021). The safety indicators and the effect of the graph structures of the environment on the relevant parameters were determined based on the developed model (Pauer & Török, 2022).
Thanks to the results of our research work, our colleague has received outstanding recognition. Gábor Pauer, a colleague of the Safety & Security Research Group of the Department of Automotive Technologies, won the BME e-grant for his outstanding achievements in the field of optimizing the transport processes of autonomous transport systems.
As the first step of the research, we developed an optimization model with which the occurred traffic demands can be realized by optimizing the load on the network through linearly representing the traffic management problem. In order to influence the transport needs at the strategic level, we implemented the integration of the toll structure assigned to the routes of the network into the model. We demonstrated the operation of the basic models on example networks, we also performed a sensitivity analysis, thus substantiating our further investigations.
The increasingly detailed resolution of space and time causes a significant increase in computational complexity during optimization. Increasing the resolution of the partition reduces the problem but, at the same time, reduces the efficiency of the optimization. Therefore, we have developed procedures to reduce the number of constraint expressions to reduce the model's runtime. The established procedures can exclude unnecessary and redundant constraints and examined locations that are certainly not relevant to the given trip (e.g., not available due to speed restrictions). We evaluated the individual and combined effects of the developed methods on example networks, also checking that the methods do not reduce the efficiency and reliability of the solution. Based on our results, the procedures reduce the computational complexity by up to 84–96%.
Using the presented model, we created a basis for traffic safety analyses in autonomous transport systems.
For this purpose, we defined indicators evaluating the safety of traffic flow. The developed indicators examine the number and distance of crossing vehicle movements and the values and homogeneity of individual and network speeds of vehicles. On the one hand, the indicators can be used to assess and rank the safety of the traffic distribution determined during the optimization. On the other hand, they provide an opportunity to examine the safety effects of different transport network structures or their interventions.
Beyond our ongoing efforts to improve the safety of highly automated vehicle systems, we also consider examining automotive security challenges a critical task.
In the field of automotive cybersecurity, our group has reached a significant milestone: we have completed the 1st phase of the IMP MASPOV project aimed at laying the methodological foundations of the National Automotive Cybersecurity Certification Centre.
Phase I of MAPSOV was also used to bring the planned collaborations to life. As a first step, we prepared for closer collaboration with the Karlsruhe Institute of Technology (KIT). We presented the results of Phase I and the research processes planned in Phase II to the Karlsruhe Institute of Technology (KIT), UTIMACO and the Ostbayerische Technische Hochschule Regensburg. Following the presentation, the partners expressed their interest in the project, and accordingly we launched the process of signing the joint MOU (Memorandum of Understanding). This process has already been completed for KIT and UTIMACO. In addition to the above, we continue the joint scientific work with the Fachhochschule Campus Wien (FHCW).
Furthermore, we developed special techniques based on both supervised and unsupervised learning methods to investigate the future research challenges in automotive security. In accordance with this, we indicated those most relevant fields which are expected to change vehicular security's future. Accordingly, our research study also supports automotive Original Equipment Manufacturers (OEMs) to identify the relevant development directions that can reduce critical vulnerabilities even in the earliest development phase. According to analysis, machine learning techniques and V2X related cybersecurity applications are expected to lead the automotive development efforts (Pethő, Török, & Szalay, 2021) in the future.
Pauer, G., & Török, Á. (2021). Binary integer modeling of the traffic flow optimization problem, in the case of an autonomous transportation system. Operations Research Letters, 49(1), 136-143.
Pauer, G., & Török, Á. (2022). Introducing a novel safety assessment method through the example of a reduced complexity binary integer autonomous transport model. Reliability Engineering & System Safety, 217, 108062.
Pethő, Z., Török, Á., & Szalay, Z. (2021). A survey of new orientations in the field of vehicular cybersecurity, applying artificial intelligence based methods. Transactions on Emerging Telecommunications Technologies, 32(10), e4325