Evaluation of different vehicle types by using modelling algorithms in road traffic virtual networks platforms

With the growing demand for simulating and analyzing road traffic, the scope of project planning, modelling, and execution will no longer be restricted to domain experts. In the future, a significantly larger portion of engineering and traffic management professionals will be required to carry out these activities. In this context, it is necessary to have several modelling and analytic algorithms that are based on the current state of affairs. Currently, the research in this subject is facing issues related to driving style variables and the integration of vehicles with the Intelligent Transportation System (ITS). Recently, there has been a focus on studying the significance of driving style for vehicle classes in connection to traffic flow, enhanced comfort, road safety, and the decrease of chemical and noise pollution. The configuration of traffic patterns is constantly changing, leading to the increasing prevalence of the idea of micro-mobility in recent years. The authors provide a comprehensive overview of the primary mathematical models employed in specific applications, as discussed before.


Introduction
The algorithms that are used to simulate driving styles make use of physical models and algorithms that have been calibrated using data from the actual world.However, due to the high level of complexity, there are currently far too few of these physical models that are accessible for widespread usage.The usage of simulated road networks is one approach that may be taken to address this issue; by doing so, one can also address the more basic issue of knowing how various road traffic situations and modifications could behave in the actual world [1].Simulation models may also assist us in conceptualizing systems that are far more complicated than what we are able to represent by only employing straightforward mathematical calculations and minute alterations to the flow of real-time traffic.
The use of driving style for different vehicle classes in is an essential component of many of the driving style algorithms that are incorporated into virtual road traffic modelling [2,3].Using this parameter, cars are able to enter gaps that are too small to respect the stability of the car following algorithm (how vehicles respond to the vehicles in front of and surrounding them).
Either the vehicle that is changing lanes or the car that is behind it will have to use the brakes at a pace that is up to twice as fast as their maximum rate of deceleration.

Theoretical considerations with reference to model-building methods
There are multiple considerations with regards to driving style for vehicle classes algorithms used for virtual networks simulations [4].If the vehicle in the network is considering the decision to make a alter the path, then the decision it is used in the virtual network.If the vehicle in the network is making the actual execution to go to the new path, then the network and the path of the vehicle is adapted to satisfy the virtual network as whole.
For a virtual simulation of road traffic, the driver comportment is assimilated with virtual vehicle within the network.In a virtual simulation of road traffic, the behavior of the driver is represented by a virtual car that operates inside the network.The virtual car is programmed to follow traffic rules and respond to various scenarios encountered on the road.It mimics the actions of a real driver, such as accelerating, braking, and changing lanes.
The algorithms that decide the behavior of a vehicle can be divided into one of two primary categories: tactical or operational [3,4].When using tactical category, the virtual cars (virtual drivers) make judgments depending on the existing and predicted characteristics of the traffic.At the same time, the operational choice is solely dependent on the knowledge, which may be restricted depending on the circumstances, on the characteristics of the neighboring traffic.The traits and actions of the driver have a significant impact on whether or not they have a desire to adjust the driving style, and whether or not they are required to do so.The changing in driving style may be necessary for a variety of reasons, including slower traffic, standing in a line, or being forced to slow down by the car in front of you, among other things.The driver may choose then to change the behavior if the car that they are driving has to slow down because of the vehicle in front of them [5].The acceleration of the vehicle in the lane that it is now in can be obtained by using the equation 1, bellow: where:   is the sensitivity coefficient   is the velocity of the vehicle in question   is the speed exponent (ranging from -2 to +2)  ∆ is the velocity of the leading vehicle subtracted from the velocity of the vehicle in question   is the distance gap between the leading vehicle and the vehicle in question   is the distance headway exponent (ranging from +4 to -1).If  +1 is less than zero, it indicates that the driver will need to slow down his vehicle, which is something that no driver wants to do [6].In such scenario, the driver wants to switch lanes (behavior change), but if they don't provide their approval, the vehicle will stay in the same lane.
The model incorporates a series of choices that are representative of the algorithm shown in the picture below.or not a change is absolutely necessary [7].The second stage is to determine whether or not it would be desirable to make a change based on the existing situation in the context of the anticipated improvement in traffic circumstances.The last stage of the model involves determining whether or not the behavior change is possible, after which the switch is actually made.
The decision model developed by Gipps takes into account both the targeted speed of the network vehicle and the maximum speed at which it is safe to go [8,9].These speeds are examined simultaneously in order to avoid the effect that may be caused by slow-moving network cars or congestions that are located a significant distance from the vehicle [10].

Network design
The behavior algorithms that are included into the Aimsun platform are being used for the road traffic evaluation that is being done in connection to driving styles using modelling algorithms.For the purpose of this evaluation, we used a digital model of the road traffic network in Slatina city.
A modeling of the whole region analyzed was carried out in the city of Slatina, and it was based on research that was done on the present traffic.The data on the current traffic was obtained by employing virtual inductive loops that were based on Quercus SmartLoop cameras and also human counters done in different key points.Several calibration components have been included into the modeling as a consequence of the early simulations as well as the need to calibrate the virtual network.These calibration elements are as follows:  Nicolae Titulescu Boulevard was introduced into the virtual network, beginning at the intersection with Oituz Street and continuing in a clockwise direction toward Vederii Street  Creating different centroids for secondary and tertiary roads that junctions with the main arteries  Created were two scenarios lasting two hours each, one for morning and one for afternoon.The created virtual system is made up of the following components:  9 km length sections  12 km lane traffic lanes  106 sections  30 intersections (intersections are all the connections between 2 adjacent sections)  17 centroids

Calibration of current traffic conditions
After the virtual system of the street frame in the examined region was built, the next step was to fill the virtual network with types of vehicles based on traffic measures as well as the data collected from inductive loops and counting.This was done so that the network could be as close to the real traffic conditions.
For this part of the process, there were a total of 17 centroids employed, which are sites from which vehicles enter and depart the network.A destination origin matrix representing the complete model was constructed using these centroids as the foundation.

Modelling of a virtual network
It is required to employ a dynamic simulation of road traffic in order to get a high degree of detail of the system that is being simulated in order for us to be able to provide a wide variety of reports.The amount to which the model's quality is determined by the availability and precision of the input data is a significant factor.Therefore, the following information must be provided.
 Road traffic network  Current traffic demand in the region being investigated  Current plans for road traffic signaling; In order to analyze the road traffic assessment in connection to driving style and vehicle class by using modeling algorithms, we have used three distinct scenarios based on three different types of vehicles on which the variations are connected to the lane shifting behavior.The following circumstances have been altered from the default scenario: remaining in the lane designated for overtaking, undertaking, and reckless lane shifting.Following the implementation of the aforementioned changes, the virtual network received the following three categories of vehicles:  cars (light vehicles that are calibrated to be similar to personal vehicle in traffic)  van (vehicles that used for transport goods in the city and are usual around 3.5 tones)  truck (heavy vehicles that are used to transport a higher number of goods) To arrive at an approximation of the road traffic assessment in connection to driving style and vehicle class, we analyze the scenario using the virtual vehicle settings described above, taking into account the fact that all cars entering the network are distributed evenly across the vehicle types used.Under these circumstances, we do an analysis of the situation based on the following criteria: time spent waiting, driving speed, fuel consumption, carbon dioxide emissions, particulate matter emissions during intercity travel.

Simulation results
The simulation was carried out using the genuine data set, which was derived from the measurements that were taken in the field.The research is carried out in the span of two hours, beginning at 7:30 and ending at 9:30.Following the completion of the simulation, the findings are shown in a graphical format, which may be seen in the figures that are provided below.

Conclusion
As a general rule, vehicle class and driving stile have a major impact on the overall features of traffic flows in an area because of the influence that the different maneuvers have on the traffic that is around them.This is because the flow induces the surrounding traffic to behave in a certain way.Because they may be simply added as an additional resource to already existing lane-changing control and prediction systems, the lane-changing decision models that are based on the characteristics of the traffic are, in many circumstances, the best option.For instance, they may be effective in a project aimed at easing traffic in an area that has a high concentration of automobiles packed into a relatively small space.This quality may be detected in the virtual network that was used to assess the driving style in respect to the road traffic.
The findings of the simulation demonstrate very clearly that the vehicles in the network behave differently based on the driving style of each vehicle class.This suggests that the driving style of a vehicle class has a significant impact on the overall behavior and performance of the network.Furthermore, understanding these differences can help in developing more efficient and tailored strategies for managing traffic flow in mixed vehicle environments.
By analysing the data collected from the simulation, it becomes evident that certain vehicle classes tend to exhibit more aggressive driving behaviours, while others adopt a more cautious approach.This knowledge can be leveraged to create targeted interventions and regulations that encourage safer and smoother traffic flow.Additionally, the findings emphasize the importance of considering the diversity of vehicle classes when designing transportation systems, as the driving style of each class directly influences the overall efficiency and effectiveness of the network.

Figure 1 .
Figure 1.Lane change, a classification of the possible models

Figure 2 .
Figure 2. Gap acceptance schematics for lane change Following the schematics shown above, the initial stage of the model involves determining whetheror not a change is absolutely necessary[7].The second stage is to determine whether or not it would be desirable to make a change based on the existing situation in the context of the anticipated improvement in traffic circumstances.The last stage of the model involves determining whether or not the behavior change is possible, after which the switch is actually made.The decision model developed by Gipps takes into account both the targeted speed of the network vehicle and the maximum speed at which it is safe to go[8,9].These speeds are examined simultaneously in order to avoid the effect that may be caused by slow-moving network cars or congestions that are located a significant distance from the vehicle[10].

Figure 4 .
Figure 4. O/D matrix from 07:30 to 09:30 for the small vehicles classThe Aimsun platform creates various characteristics for each of these three kinds vehicles by basing those characteristics on general aspects like the vehicle's length, speed, consumption, acceleration, and response time, among other things.