The impact of automation and connectivity on traffic flow and CO2 emissions. A detailed microsimulation study
Introduction
The impact of automation-related technologies on road transport networks is a topic studied for many years now (Kesting et al., 2010; Louwerse and Hoogendoorn, 2004). Connected and Automated vehicles (CAVs) are expected to bring significant advancements in the existing road transport systems in terms of reducing traffic congestion, energy consumption, CO2 and pollutants emissions (Alonso Raposo et al., 2017; Litman, 2015).
The picture looks promising, but as research on the field progresses, many researchers express their doubts on the anticipated benefits, either energy or traffic-wise. In Fiori et al. (2019) preliminary results were presented showing that moving from internal combustion engine vehicles (ICEVs) to plug-in electric vehicles (PEVs), the relationship between congestion and energy consumption can change, with higher energy consumption connected to the free-flow test-cases. Mattas et al. (2018) presented a simulation case study on the impact of connected autonomous vehicles, focusing on the possible benefits of connectivity. The results show that depending on the traffic demand, the autonomous vehicles can have adverse effects on traffic flow, while connected autonomous vehicles can be beneficial for the network, depending on their penetration rate.
At the same time, the tools used in various impact assessment studies have known limitations that may affect the final predictions. Results in the work of Ciuffo et al. (2018) raise concerns about the capability of the existing car-following models to reproduce observed vehicles’ acceleration dynamics and thus estimate vehicles emissions and energy/fuel consumption. Most studies focus on congested conditions where the free-flow regime is expected to play a minor role. However, recent studies (Laval et al., 2014; Marczak et al., 2015) highlight that also in congested conditions, the acceleration regime, which incorporates vehicle dynamics and driving behaviour, affects the capacity drop, the hysteresis and possibly other traffic-related phenomena.
Panis et al.(Int Panis et al., 2006) highlighted the need for a detailed analysis of not only average speeds but also other aspects of vehicle operation such as acceleration and deceleration. On the same page, Lejri et al. (2018) proposed a model that accounts for traffic speed dynamics in order to provide more accurate emissions estimations. Finally, in the recent simulation study of Stogios et al. (2019), the importance of considering the different driving behaviors is highlighted.
The prospect of CAVs in reducing the environmental impact of vehicles is of great importance. In this context, it is important to investigate how and to what extent CAVs technologies will affect vehicle energy use and reduce traffic emissions. On the other hand, if the technology does not deliver the expected results, it is essential to identify the correct traffic management strategies to help reach the desired emissions-reduction goals.
The present paper studies the impact of automation and connectivity on future road transport networks, based on certain assumptions regarding the vehicle technology, the traffic supply, and the demand. A microsimulation study on a highway network conducted using AIMSUN traffic simulation software. The results are based on various scenarios with state of the art models for the simulation of different vehicle technologies involved, that is, conventional human-driven vehicles (CVs), AVs and CAVs. The same scenarios were replicated using the same models for the congested part but accounting explicitly the vehicle dynamics on free flow. Different technologies generate different driving behaviours that impact the levels of congestion in the network. The motivation of this work is to answer whether the variation in the driving behaviours produces significant differences in emissions during rush hours. Furthermore, it is important to understand the role of realistic vehicle dynamics simulation and how this impacts traffic flow and emissions.
The network used is the ring road of Antwerp presented in (Mattas et al., 2018). Initially, we build four scenarios using state of the art models for the simulation of the different vehicle types. Each of the four scenarios refers to 3-h simulation with a) CVs, b) AVs, c) CAVs and d) CAVs with 20% increased traffic demand. Furthermore, in order to study whether the explicit simulation of vehicle dynamics can change these results, the above-mentioned car-following models were modified in order to explicitly consider realistic vehicle dynamics. In order to achieve the latter, we introduce the simulation of vehicle dynamics on free-flow, using the Microsimulation Free-flow aCceleration model (MFC) (Makridis et al., 2019a). A publicly available implementation in Python of the last version of the MFC model can be found online (https://pypi.org/project/co2mpas-driver/). The above-mentioned four scenarios are replicated using now the modified car-following models, leading to 8 scenarios in total for this work. Results demonstrate the impact of connectivity and automation on traffic flow and emissions.
The reference models used for the assessment of CO2 emissions estimates are the fuel consumption and CO2 emissions factors proposed by the EMEP/EEA guidebook, and a generic version of CO2MPAS (Fontaras et al., 2018) similar to the one described in (Tsiakmakis et al., 2017). The EMEP/EEA guidebook methodology (European Environment Agency, 2016), is more widely known by its software implementation, COPERT. It foresees an average-speed model and is frequently used in most European countries in order to estimate emissions of all major air pollutants produced by different vehicle categories. CO2MPAS is a vehicle-specific simulation model recently introduced by the EU in its vehicle CO2 certification system (Fontaras et al., 2018). Finally, it should be highlighted that the results of this study are bounded by a set of assumptions, which are summarised a) in the accuracy of the car-following models used for the simulation of different vehicle types, b) the accuracy of the emissions models, c) the accuracy of the vehicle dynamics simulation model and d) the absence of electric vehicle dynamics simulation.
Section snippets
Simulations
This section describes the driver models used in order to simulate the different technologies, the models used to provide emissions estimations and finally, the physical network and the different scenarios.
Results
Simulated data regarding the state of the network were retrieved for 10-min intervals. The results focus on three major dimensions, the impact of the vehicle/driver technology, i.e. CVs, AVs or CAVs, the impact of the simulation of vehicle dynamics, i.e. MFC and finally, the differences in the emissions estimations from EMEP/EEA and the instantaneous generic CO2MPAS.
Fig. 2 illustrates the evolution of the average harmonic speed per 10-min period over the 3-h simulation for different vehicle
Discussion and conclusions
This work proposes a microsimulation framework that investigates the impact of vehicle automation and connectivity in terms of traffic flow and emissions on a realistic highway transport network, the ring road of Antwerp, Belgium. The proposed framework takes into consideration by modelling the different vehicle technologies that generate different vehicle behaviours and traffic patterns. Moreover, it investigates the potential impact of simulating realistic vehicles dynamics through modelling.
CRediT authorship contribution statement
Michail Makridis: Conceptualization, Methodology, Writing - review & editing, Data curation, Formal analysis. Konstantinos Mattas: Conceptualization, Methodology, Writing - review & editing, Data curation, Formal analysis. Caterina Mogno: Conceptualization, Methodology, Writing - review & editing, Data curation, Formal analysis. Biagio Ciuffo: Conceptualization, Methodology, Writing - review & editing. Georgios Fontaras: Conceptualization, Methodology, Writing - review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (26)
- et al.
The effect of electrified mobility on the relationship between traffic conditions and energy consumption
Transport. Res. Part Transp. Environ.
(2019) - et al.
The development and validation of a vehicle simulator for the introduction of Worldwide Harmonized test protocol in the European light duty vehicle CO2 certification process
Appl. Energy
(2018) A behavioural car-following model for computer simulation
Transp. Res. Part B Methodol.
(1981)- et al.
Modelling instantaneous traffic emission and the influence of traffic speed limits
Sci. Total Environ.
(2006) - et al.
A parsimonious model for the formation of oscillations in car-following models
Transp. Res. Part B Methodol.
(2014) - et al.
Accounting for traffic speed dynamics when calculating COPERT and PHEM pollutant emissions at the urban scale
Transport. Res. Part Transp. Environ.
(2018) - et al.
Simulating impacts of automated driving behavior and traffic conditions on vehicle emissions
Transport. Res. Part Transp. Environ.
(2019) - et al.
Influence of connected and autonomous vehicles on traffic flow stability and throughput
Transport. Res. C Emerg. Technol.
(2016) - et al.
A simulation-based methodology for quantifying European passenger car fleet CO2 emissions
Appl. Energy
(2017) - et al.
The R-Evolution of Driving: from Connected Vehicles to Coordinated Automated Road Transport (C-ART)
(2017)