RelativeGAP is proud to be involved in a Connected and Autonomous Vehicles research programme with the Budapest University of Technology and Economics.
We are responsible for macroscopic modelling in the research using PTV Visum and we are looking to analyse various operational scenarios including thought-provoking topics such as the cyber security of CAV systems.
Modelling Urban Autonomous Transport Systems
Modelling Urban Autonomous Transport Systems. The advent of fully autonomous transport is among the most promising opportunities in transportation in the next decades. Nevertheless, we need to bear in mind the challenges this brings to future transport systems. It is the mission of modelling professionals to explore several engaging topics of emerging mobility.
Accordingly, the objective of this study is to define a generally usable methodological framework, which is appropriate to predict the network level benefits related to autonomous systems and which is adequate to identify possible expenses associated with the introduction of such a mobility system. During the research the Unified Traffic Model of Budapest (EFM) has been applied. Using autonomous vehicle penetration as an external parameter, the investigation has considered a permanent increase in the number of autonomous cars. This study has also considered the most important parameters of transport systems involving connected and autonomous vehicles including following distance, reaction time, parking issues, access and egress cost, etc. There are many other macroscopic modelling aspects of emerging mobility which can be taken into account in the future phases of this research. According to the result of the investigation, it can be concluded that in case of a 100% penetration, up to 2 billion hours journey time saving can be achieved assuming a 15-year-long operation time and a city with about 1.7 million car trips per day.
During the investigation traditional macroscopic transport modelling methodology has been applied, built up from a demand and a supply side. The demand side is represented by origin-destination matrices of the considered demand segments, whilst the supply side of the model is represented by a complex directed graph describing the road network of the modelled area. The comparison of the different alternatives is based on the main indicators of the traditional transport modelling approach (e.g. travel time savings, traffic volumes, speeds, etc.). Besides, a comprehensive vehicle technology framework has been set up to evaluate the impact of initiating an autonomous transport system. Thus, the output of the study is a general benchmarking of a traditional transport system and autonomous transportation in accordance with the framework of this study.
Cyber Security Challange
Cyber Security Challenge. By now information technology revolution has reached our vehicles. More advanced electronic innovations, connected vehicles, self-driving functionalities or even the use of navigation systems are making our lives easier but at the same time they pose security challenges. This study aims to estimate the effects of such risks on transportation at a macroscopic level.
This cyber security research has been using our model built for macroscopic transport modelling of autonomous vehicles. The basis of the model was the Unified Transport Model of Budapest (EFM) maintained by the Centre for Budapest Transport (BKK). This analysis examined the effects of cyber security attacks. The presumed attacks are divided into the following two categories:
Intervention in the route choice: These types of attacks need to be considered not only for vehicles with advanced self-driving functionality but when drivers are using any navigation software. When looking at attacks we sought to find out how traffic flow can be diverted from the equilibrium state by over-preferring or avoiding certain routes or links.
Decrease in capacity by immobilising vehicles: Our assumption was that we induced a bottle neck effect by decreasing the capacity at certain strategic points. To assess the effects, the daily model had to be converted into a peak hour one and we used the blocking back functionality of PTV Visum to be able to replicate queues.
In our study we did not specifically aim to examine the changes in journey time or vehicle kilometres in the model, but the main goal was to create a widely applicable methodology. Therefore, the achievement of our research is the modelling process itself that allows for applying this methodology in the future to answer specific system vulnerability questions.