Unity-based Simulation Scenarios to Study Driving Performance

David González-Ortega, Francisco J. Díaz-Pernas,

Mario Martínez-Zarzuela and Míriam Antón-Rodríguez

Department of Signal Theory and Communications and Telematics Engineering,

Telecommunications Engineering School, University of Valladolid, Valladolid, Spain

Keywords: Driving Simulation, Unity, Driving Learning, Driving Efficiency.

Abstract: In this paper, we present two 3D driving simulation scenarios and a vehicle model developed to analyze

driving efficiency and safety. The scenarios include different roads, traffic, and events so that drivers can

acquire driving competences and their performance can be tested. They were developed with Unity game

engine, which allows to create 2D and 3D games and applications and to export them to many different

platforms. The first scenario is urban with varied roads, crossroads controlled by traffic lights, and

roundabouts. The second scenario is interurban with different road sections connected by roundabouts. The

vehicle engine has been modeled setting the dependency between engine torque and rpm and the

dependency of each gear and speed on fuel consumption. Automatic or manual gear shifting can be selected.

During a simulation, the speed, rpm, gear, consumption, and traffic offences are showed in real time and

stored in files for further processing. Seven people drove in the scenarios and their stored data enabled us to

analyze differences in driving performance between them. The simulator was ranked positively by all the

participants regarding ease of interaction, similarity to real driving, and usefulness for driving learning.

1 INTRODUCTION

Event simulation is being increasingly used in many

fields. With the advance of technology, it is more

feasible to create very realistic environments that let

users an easy immersion in the simulators, achieving

best results in their use. These simulators can speed

up the process of acquisition of basic skills and are

configured as tools of great learning potential in

many fields.

Particularly, driving simulators achieve driver

learning and reeducation of drivers through the

inclusion of varied routes and situations where

safety can be compromised depending on the driver

behavior. Many companies such as Renault are

investing large amounts of money in driving

simulators (Munir et al, 2017). Simulation can let

future drivers get used to many situations that take

place rarely but that can result in traffic accidents,

such as landslides, animals in the middle of the road,

or a previous accident. The development of

simulation environments is important not only for

their utility to drivers but also because it allows to

recover data that can be analyzed to study different

aspects that have influence on traffic safety. This

data can also be used to discover shortages in the

education and abilities of drivers that can lead to

dangerous situations. Simulators can help to

maintain the decreasing trend in the rate of traffic

casualties present in the last years in the European

Union. 26 out of the 28 EU countries have decreased

the number of traffic casualties with respect to 2010

(Mobility Transport. European Comission, 2018).

In this work, two driving simulation scenarios

with different road sections and events and a vehicle

model to analyze the driving efficiency and safety

are presented. The scenarios and the vehicle model

have been developed with the Unity game engine

(Unity, 2018). Unity is one of the most used engines

in low budget projects and allows to develop 2D and

3D games and graphical applications. Besides, it is

possible to export the applications to 27 different

platforms including PC, mobile, and console ones.

An important feature of the last version of Unity is

its integration with virtual reality platforms such as

Oculus Rift.

The driving simulator that integrates the two

scenarios shows in real time and stores in files for

further processing, not only information about the

driver vehicle (position, speed, rpm (revolutions per

González-Ortega, D., Díaz-Pernas, F., Martínez-Zarzuela, M. and Antón-Rodríguez, M.

Unity-based Simulation Scenarios to Study Driving Performance.

DOI: 10.5220/0006914601830189

In Proceedings of 8th Inter national Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2018), pages 183-189

ISBN: 978-989-758-323-0

183

minute), gear, and fuel consumption), but also the

traffic offences committed during simulation. This

information can be used to analyze the different

simulations fulfilled by the drivers so that conclusions

can be drawn about their driving competence.

The rest of the paper is organized as follows.

Section 2 presents the state of the art on driving

simulators. Afterwards, Section 3 explains the

developed driving simulation scenarios and user’s

vehicle. Section 4 details the obtained experimental

results and, finally, Section 5 draws the main

conclusions about the presented work.

2 DRIVING SIMULATORS

A simulator is a hardware and software configure-tion

in which, through calculation algorithms, the behavior

of a particular process or physical system is replicat-

ed. In this process, the real situations are substituted

by other artificially created that can serve to acquire

competences that will be transferred later to a real life

situation in an effective way. In the education field,

simulators are means not only for concepts formation

and knowledge building but also for the application of

them to new contexts that people, for several reasons,

cannot have access to from the methodological

context where their learning is developed. Simulators

have educational advantages such as the providing of

open learning environments based on real models and

that users adopt an active role, turning themselves into

the builders of learning from their own experience. To

name some examples, simulation has been applied to

fields so different as power engineering (Zimmerman

et al., 2011) and spine surgery (Ryu et al., 2017)

education with satisfactory results.

All the formerly mentioned is clearly applicable

to driving simulators. Although there are driving

simulators aimed to the learning of future drivers

(DriveSim, 2018) from which the user can learn to

drive from scratch, most of these simulators aim to

entertain leaving aside the development of

competences towards responsible driving. A large

number of users, including race drivers, highlight

that these types of simulators usually provide a very

realistic experience and very detailed circuits that

lets them know and adapt to new competitions, in a

way that the routes are known before doing them for

real (iRacing, 2018). Simulators such as iRacing,

Project Cars, Gran Turismo, or Forza Motosport are

recognized worldwide for their fidelity in the

representation of racing vehicles’ behavior.

Concerning these kind of simulators, (Stinchcombe

et al., 2017) found a statistically significant

association between video game experience and

risk-taking behaviors (large values of speed and

crashes) when the participants had to drive in a

scenario developed specifically to assess a particular

skill such as handling. As a consequence, simulators

for learning and assessing safe and efficient driving

should be carefully designed to reinforce proper

driving behaviors through the provision of feedback.

In the last years, driving simulators have been

used in many research studies for a wide variety of

purposes. With them, drivers’ reaction to different

situations, which cannot be replicated in a real road,

can be evaluated (Bham et al., 2010). Their

suitability to assess the behavior of drivers and as a

means to learn safe and efficient driving skills has

been shown. (Sullman et al., 2015) showed the

potential of the training in efficient driving skills

with a simulator for bus drivers. (Jamson et al.,

2015) studied with a simulator the tradeoff between

driving efficiency and safety using systems

embedded in the vehicle to advise the driver about

the use of the throttle pedal. They showed that

efficiency can be improved using these systems.

Simulators were also used to compare younger and

older driver performance (Stinchcombe et al., 2010)

and to correlate driver cognitive measures and driving

performance (Yamin et al., 2016). (Hooft van

Huysduynen et al., 2018) used the Multidimensional

Driving Style Inventory (MDSI) to study the relation

between self-reported driving style and the driving

behavior in a simulator, concluding that there is a

modest correlation between them.

In driving simulators, scenarios to analyze the

level of driving efficiency and safety, the relation

between them, and the results produced as a function

of them can be included. When a person sees itself

in a real driving environment such as a city with

heavy traffic or a complex junction, he may think

that he will be unable to face that challenge.

Particularly, there are people suffering from

amaxophobia, characterized by the fear of vehicles

or travelling in a vehicle (Gossling, 2017). Driving

simulators enable users to face many of these

problems in a virtual environment without

endangering the user or other drivers. Simulators

enable the users to face problems and to obtain

driving skills to overcome these problems, in such a

way that these skills can be applied when they face

similar situations in a real environment. The users

can also get used to the vehicle controls through

simulations, manipulating them with the steering

wheel, pedals, and gear lever similarly to in a real

vehicle. For that purpose, different types of

peripherals can be used, from controls to complete

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

184

driving cabin, including devices such as Logitech

G27. Obviously, simulators enable users to repeat

simulations in a scenario and to face certain events

many times. Besides, it is possible to simulate real

situations that would be practically impossible to do

voluntarily in a real environment. Situations such as

a vehicle stopped in the middle of a road or a

pedestrian crossing a road unexpectedly are events

that can be replicated in a simulator easily. This will

enable simulator users to be better prepared for

unusual situations that can find in the future.

Simulators can also be used to do studies about how

certain agents, such as fatigue, physical or

psychological state of the driver, medicine, drugs, or

alcohol consumption or age, can influence driving

capacity significantly. Moreover, they can be used to

study the influence of Advanced Driver Assistance

Systems (ADAS) (Gonçalves et al., 2014), or of the

presence of children in the vehicle (Olaverri-

Monreal et al., 2014), on driving performance easily

and without danger. Lastly, it is important to

mention that simulators can gather data in a simple

way. Thus, users can have access to information of

the fulfilled simulations and observe their evolution.

This enables drivers to see their mistakes and

committed traffic offences and what aspects they

should improve to avoid them in the future.

3 DEVELOPED DRIVING

SIMULATION SCENARIOS

AND USER’S VEHICLE

We have developed two 3D driving simulation

scenarios using the Unity game engine with a view

to covering a wide range of environments and

situations that drivers should know to react properly.

We have also modeled the user’s vehicle. As a

peripheral input module, we have selected the

Logitech G27 device, which includes steering wheel,

clutch, brake, and throttle pedals, and gear lever.

3.1 Urban Scenario

The first scenario is urban. We have used the Road

& Traffic System plugin from the Unity Asset Store

to create the road sections and to include the traffic.

This plugin can be used to create dynamic traffic

networks in roads with one, two, and three lanes per

direction, crossroads controlled by traffic lights, and

roundabouts.

Pedestrians were included in the scenario. They

use a 3D model of a person with its linked skeleton

and animations that once applied to the 3D model

make it behave with the intended motion. There are

three different animations: Idle (standing still

pedestrian), Walk Forward (pedestrian walking

forward) and Hit (pedestrian being hit by a vehicle).

The driver is guided in the crossroads and

roundabouts to follow an established route. Fig. 1

shows a schematic view of the scenario where

sections with different lanes, 4-legged and T

crossroads and roundabouts (green circles) can be

observed. Fig. 2 shows images from this scenario. In

these images, the vision of the rear view mirror at

the top center, the current gear at the bottom center,

and the rpm meter and speedometer at the bottom

right of them can be seen.

Figure 1: Schematic view of the urban scenario.

(a)

(b)

Figure 2: Urban scenario images.

Unity-based Simulation Scenarios to Study Driving Performance

185

3.2 Interurban Scenario

The second scenario is interurban with road sections

between 0.5 and 2.5 kilometers long interconnected

with 2-lane and 3-lane roundabouts. Roundabouts

are very common in roads throughout the world as

they have replaced traditional 4-legged crossroads

(Nikitin et al., 2017) to improve traffic flow and

safety. The road sections are also of 2 types: some

have one lane per direction and others have two

lanes per direction. Fig. 3 shows a schematic view of

the scenario. Similarly to the first scenario, drivers

are guided with signs informing, while approaching

a roundabout, about the exit they have to take. The

pedestrians were also included similarly to the first

scenario. The traffic vehicles were included by

applying two scripts to them, one for the car as a

whole and one for each wheel without using the

Road & Traffic System to control their behavior so

that adaptation to interurban traffic can be better

achieved. Each vehicle has to know if there are other

vehicles in the surroundings to change its speed or

lane accordingly. Fig. 4 shows some images from

the scenario. Fig. 4(c) shows an image where the

scenario projected on a screen and the Logitech G27

steering wheel can be observed.

3.3 User’s Vehicle

Regarding the engine modeling of the user’s vehicle

(with internal combustion engine), the dependency

between the engine torque and horsepower and rpm

were set as shown in Fig. 5. Other factors to

compute the overall force of the engine have to be

set, such as differential ratio, gear ratio (per each

gear) and wheel rpm. To compute the fuel

consumption, its dependency on the current gear and

speed was set as shown in Fig. 6. It is possible to

select manual or automatic gear shifting before the

beginning of a simulation. There are several cameras

that can be selected by the user to change the field of

view. One camera is outside the vehicle as shown in

Fig. 4(c). Other camera is inside the vehicle, as

shown in Figs. 4(a) and 4(b), to have a view similar

to the driver in a real vehicle. With this view, the

vision of the rear view mirror of a vehicle is

included. Besides, the user can turn the view left or

right pressing particular buttons on the steering

wheel. While Fig. 4(a) shows the frontal view inside

a car while approaching a roundabout, Figs. 7(a) and

7(b) show the view turning left and the view turning

right that a user had in nearby places during a

simulation, respectively.

Figure 3: Schematic view of the interurban scenario.

(a)

(b)

(c)

Figure 4: Interurban scenario images.

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

186

In each route in a scenario, the simulator stores

for further processing the position (x, y, and z

coordinates), the current and average speed, covered

distance, rpm, current gear, and current, average,

and total fuel consumption every second.

Figure 5: Graphs of the engine torque and horsepower as a

function of rpm of the modeled vehicle.

Figure 6: Graphs of the fuel consumption as a function of

the current gear and speed of the modeled vehicle.

4 EXPERIMENTAL RESULTS

Seven people drove in the driving simulation

scenarios ranging from 21 to 42 years old. The

participants drove in the first scenario twice, firstly

with automatic gear shifting and secondly with

manual gear shifting to analyze the influence of the

different gear shifting. They drove in the second

scenario with manual gear shifting. Figs. 8 and 9

show the average speed and the average fuel

consumption of the users in the three simulations,

respectively. As expected, the average speed is much

lower in the urban scenario, where there were small

differences in the routes using automatic and manual

gear shifting. Five out of the seven participants

drove a bit faster with automatic gear shifting than

manual gear shifting in the first scenario as they did

not have to pay attention to the gear shifting.

Regarding the fuel consumption, it was higher in the

urban scenario as expected and five out of seven

participants has a higher fuel consumption with

manual gear shifting as they did not change gear as

efficiently as using automatic change.

(a)

(b)

Figure 7: Interurban scenario images.

Figs. 10 and 11 show the rpm, standard deviation

of rpm, and average consumption using manual gear

shifting for the urban and interurban scenario,

respectively. The standard deviation of rpm is much

larger in the urban scenario as expected due to

frequent stops in it. In both scenarios, there is a

correlation between large values of rpm, the

standard deviation of rpm and average consumption

as economic driving implies soft driving trying to

keep the value of rpm low and constant.

While User #1 had both the lowest value of

standard deviation of rpm and average consumption

in the urban scenario, he clearly decreased his

driving performance in the interurban scenario

comparatively as he had both the second largest

value of standard deviation of rpm and average

consumption in it. The rest of the users had

comparative values of the standard deviation of rpm

quite similar in the two scenarios. User #7 was the

participant with the worst driving performance as he

had the largest value of standard deviation of rpm

and the third largest value of average consumption

in the urban scenario and had both the largest value

of standard deviation of rpm and average

consumption in the interurban scenario.

Unity-based Simulation Scenarios to Study Driving Performance

187

Figure 8: Average speed of the users in the simulations.

Figure 9: Average fuel consumption of the users in the

simulations.

Figure 10: rpm, Standard deviation of rpm, and average

consumption using manual gear shifting for the urban

scenario.

Figure 11: rpm, Standard deviation of rpm, and average

consumption using manual gear shifting for the interurban

scenario.

After using the simulator, the seven participants

were asked about their experience driving in the

scenarios. They were asked about the ease of

interaction, the similarity to real driving, and the

usefulness for driving learning, obtaining an average

score (out of 10) of 8.5, 7.5, and 7, respectively.

5 CONCLUSIONS

In this paper, two driving simulation scenarios,

which were developed with the Unity game engine,

have been presented. The simulator makes it

possible the acquisition of driving skills and the

assessment of driver performance through the

inclusion of varied road configurations, user’s

vehicle, traffic, and events so that the safety and

efficiency are affected by the driver behavior. The

continuous technological advances have enabled the

development of low-cost simulators as the presented

with realistic scenarios and events impossible to

have in a real driving environment without

endangering people or vehicles. People that took

part in the experiments driving in the simulation

scenarios highlighted the ease of interaction, realistic

experience, and usefulness for driving learning of

them. The data stored in the simulation will allow to

carry out deep comparative analysis to associate the

different drivers and their driving styles with their

100

User

Scenario 1

(automatic)

Scenario 1

(manual)

Scenario 2

(manual)

User

Scenario 1

(automatic)

Scenario 1

(manual)

Scenario 2

(manual)

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

188

safety and efficiency level, and assess the evolution

in the development of driving competence as more

drivers will use the simulator.

The simulator can be extended to new scenarios

and events and virtual reality devices. Oculus Rift

virtual reality device is planned to be integrated in

the simulator to achieve a more realistic and

immersive driving experience together with new

scenarios and different user’s vehicles regarding size

(bus, truck) and motor (electric).

ACKNOWLEDGEMENTS

This work was supported by the National

Department of Traffic (DGT) of the Ministry of the

Interior (Spain) under research project SPIP2017-

02257.

REFERENCES

Bham, G. H., Mathur, D., Leu, M., Vallati, M., 2010.

Younger drivers evaluation of vehicle mounted

attenuator markings in work zones using a driving

simulator. Transportation Letters: the International

Journal of Transportation Research, 3: 187-198.

DriveSim, 2018. Available at http://drivesimsimulator.

com/. [Accessed 17 Apr. 2018].

Gonçalves, J. S. V., Rosetti, R. J. F., Jacob, J., Gonçalves,

J., Olaverri-Monreal, C., Coelho, A., Rodrigues, R.,

2014. Testing Advanced Driver Assistance Systems

with a Serious-Game-Based Human Factors Analysis

Suite. In: Proceedings of the IEEE Intelligent Vehicles

Symposium, 13-18.

Gossling, S., 2017. The psychology of the car, Elsevier.

Amsterdam. 1st edition.

Hooft van Huysduynen, H., Terken, J., Eggen, B., 2018.

The relation between self-reported driving style and

driving behaviour. A simulator study. Transportation

Research Part F, 56: 245-255.

iRacing, 2018. Available at https://www.iracing.com/.

[Accessed 17 Apr. 2018].

Jamson, S. L., Hibberd, D. L., Jamson, A. H., 2015.

Drivers’ ability to learn eco-driving skills; effects on

fuel efficient and safe driving behaviour.

Transportation Research Part C: Emerging

Technologies, 58: 657-668.

Mobility Transport. European Comission, 2018. Available

at https://ec.europa.eu/transport/road_safety/specialist/

statistics_en# [Accessed 19 Apr. 2018].

Munir, S., Hovd, M, Fang, Z., Olaru S., Kemeny, A.,

2017. Complexity reduction in motion cueing

algorithm for the ULTIMATE driving simulator. IFAC

PapersOnLine, 50-1: 10729–10734.

Nikitin, N., Patskan, V., Savina, I., 2017. Efficiency

Analysis of Roundabout with Traffic Signals.

Transportation Research Procedia, 20: 443-449.

Olaverri-Monreal, C., Gonçalves, J., Bengler, K.-J., 2014.

Studying the Driving Performance of Drivers with

Children Aboard by Means of a Framework for

Flexible Experiment Configuration. In: Proceedings of

the IEEE Intelligent Vehicles Symposium, 7-12.

Ryu, W. H. A., Mostafa, A. E., Dharampal, N., Sharlin, E.,

Kopp, G., Jacobs, W. B., Hurlbert, R. J., Chan, S.,

Sutherland, G. R., 2017. Design-Based Comparison of

Spine Surgery Simulators: Optimizing Educational

Features of Surgical Simulators. World Neurosurgery,

106: 870-877.

Stinchcombe, A., Gagnon, S., 2010. Driving in dangerous

territory: Complexity and road-characteristics influen-

ce attentional demand. Transportation Research Part

F: Traffic Psychology and Behaviour, 13(6): 388-396.

Stinchcombe, A., Kadulina, Y., Lemieux, C., Aljied, R.,

Gagnon, S., 2017. Driving is not a game: Video game

experience is associated with risk taking behaviours in

the driving simulator. Computers in Human Behavior,

69: 415-420.

Sullman, M. J. M., Dorn, L., Niemi, P., 2015. Eco-driving

training of professional bus drivers – Does it work?

Transportation Research Part C: Emerging

Technologies, 58: 749-759.

Unity, 2018. Available at https://unity3d.com/es/

[Accessed 16 Apr. 2018].

Yamin, S., Stinchcombe, A., Gagnon, S., 2016. Deficits in

attention and visual processing but not global

cognition predict simulated driving errors in drivers

diagnosed with mild Alzheimer’s disease. American

Journal of Alzheimer’s Disease and Other Dementias,

31(4): 351-360.

Zimmerman, R. D., Murillo-Sánchez, C. E., Thomas, R. J.,

2011. MATPOWER: Steady-State Operations,

Planning, and Analysis Tools for Power Systems

Research and Education. IEEE Transactions on Power

Systems, 26(1): 12-19.

Unity-based Simulation Scenarios to Study Driving Performance

189