Leadership: Vitaly Stepanyants

Contact e-mail address: [email protected]


Научно-учебная группа технологий подключенного и беспилотного транспорта

Проектная команда МИЭМ объединяет беспилотный и подключенный транспорт

Development - Project Repository

CAVISE

Research - Structured knowledge base

(Private, Coming Soon)

Collaboration Pages

(Private, Coming Soon)


Latest News

Title | Text | Illustration (optional)

Announcements (our & adjacent events)

Date | Event Name | Description | Time & Place | Entry Requirements


CAVISE: A Connected Automated Vehicle Integrated Simulation Environment

CAVISE development plan presentation (in Russian) TBU

CAVISE development plan presentation (in Russian) TBU

Introduction

TBU

There are more and more vehicles using automated and connected vehicle (CAV) technologies, and business models of sharing and Mobility-as-a-Service, combined in the concept of Intelligent Transportation Systems (ITS), which aims to reduce traffic congestion, improve safety for road users, as well as reduce energy consumption and harmful emissions. It is necessary to ensure that these technologies can actually help achieve these goals, and therefore they require extensive testing (verification and validation) before they can be allowed on public roads. Computer simulation is used to speed up the testing and development process.

There is a need for the creation of large-scale models (in terms of area and dimensions), which allow modeling, for example, the impact of using transport sharing models or automated vehicle technologies on transportation system characteristics. At the same time, in order to obtain accurate results, high detailisation of the models at the lower levels must be maintained. Such a tool would allow for more efficient urban resource planning and sustainable development.

Currently, there are many microscopic models of connected and automated vehicles, but such models do not take into account the influence of environmental objects and many other parameters on the behavior of vehicles. For example, based on analysis, most models assume the presence of fully functional automated vehicles (AVs). However, from a practical point of view, their possible functionality is divided by the Society of Automotive Engineers (SAE) into five levels. As a result, there may be a mixed proportion of cars with different levels of autonomy on the roads. This creates a gap in current research on the behavior of AVs using microscopic simulators. Recent research also suggests that the presence of AVs may influence the behavior of human-driven vehicles. However, current research based on microscopic simulations suggests that human-driven vehicles will interact equally with human-driven cars and automated vehicles, ignoring the adaptation of human behavior.

The key hypothesis is that a more accurate description of environmental objects and their influence on machine perception and signal transmission will lead to quantitative and qualitative changes in the parameters obtained as a result of modeling.

Also, there are many approaches and tools for modeling objects at different levels of detail and from different domains (areas of knowledge) that affect the operation of ITS. However, these approaches and tools are not integrated with each other. Given the importance of planning the trajectories of connected and automated vehicles in microscopic simulators, joint simulation systems have been used in recent years. Joint simulation strategies show good prospects for modeling the behavior of connected and automated vehicles with a high level of detail. Most of the research on joint simulation is at the initial levels and largely focuses on the development of joint simulation frameworks. Given the computational limitations in joint simulation tools, research focuses on modeling one or more vehicles with limited traffic volume and transport network sizes. Joint simulation tools are heavily focused on planning the trajectories of connected and automated vehicles, and pay less attention to the characteristics of traffic flows and networks than with microscopic modeling. However, due to increasing avaliable computing power and the availability of APIs for joint simulation, scenarios can be expanded to more connected and automated vehicles controlled based on sensor perception of the environment, and human-driven vehicles controlled using behavioral models. However, when launching nanoscopic models that take into account external factors, the computing capabilities of most computers impose restrictions on the size of scenarios, because of this, some parameters of transport behavior are simplified.

At the moment, most models are created only for testing in a certain city or country. They do not take into account the cultures of different countries, traffic patterns and other factors. To use such a tool effectively, it is also necessary to create a format for a quick and structured description of the simulated scenarios, allowing you to take into account all the parameters essential for this task and simplify their configuration for software users.

A significant problem is also the redundancy of scientific papers and developed individual tools for development and modeling, only a few of which are actively used. The CAVISE project was tasked with exploring all currently available open source tools for modeling connected and automated vehicles, selecting the best and most compatible of them, and integrating them. In this way, it is planned to save time on development within the project, as well as to efficiently use the time already spent by other developers.