,

State-of-the-Art DrivingSimulators, a Literature Survey

Slob, J.J. (Jelmer)

DCT 2008.107

DCT report

Eindhoven University of TechnologyDepartment Mechanical EngineeringControl Systems Technology Group

Eindhoven, August 2008

Contents

1 Motion Simulation 2

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 What is Motion Simulation? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3.1 The Need for Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.4 Areas of use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.4.1 Entertainment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.4.2 Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.4.3 Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Driving Simulators 8

3 Discussion 12

Bibliography 13

i

List of Figures

1 Degrees of Freedom (DOF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Antoinette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Stewart Platform, or Hexapod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.3 Classical Washout Filter [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.7 SHERPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.8 UoLDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.9 VTI-III . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1 IFAS (1984) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.2 MARS (2004) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2.3 Ultimate (2004) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.4 Vitrtex (2001) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.5 BMW (2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.6 WIVW 1999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.7 NADS-1 (2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.8 Toyota (2007) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

ii

Nomenclature

φ Roll

ψ Pith

θ Yaw

x Surge

y Sway

z Yaw

ADAS Advanced Driver Assistance System

CWF Classical Washout Filter

DOF Degree of Freedom

DSC Driving Simulator Conference

FB Fixed Base

FS Fixed Screen

Heave Translational motion defined positively in the opposite direction of the gravitational field, z-direction in Figure 1.

hi-fi High fidelity

HMD Head Mounted Display

HMI Human Machine Interface

IVIS In-Vehicle Information System

MCA Motion Cueing Algorithm

MSC Motion Simulator Conference

Pitch Rotational motion around y-axis in Figure 1.

Roll Rotational motion around x-axis in Figure 1.

Surge Motion in longitudinal direction of the vehicle, x-direction in Figure 1.

Sway Motion in lateral direction of the vehicle, y-direction in Figure 1.

Yaw Rotational motion around the z-axis in Figure 1.

iii

z, heave

x, surgey, sway

, pitchψ

, yawθ

, rollϕ

Figure 1: Degrees of Freedom (DOF)

1

Chapter 1

Motion Simulation

1.1 Introduction

Driving simulation is a type of motion simulation that is currently undergoing an enormous changeas the demand for more advanced and sophisticated driving simulators increases, the past decades.This survey focusses to get insight to the current state-of-the-art of driving simulation and simulators.Motorcycle, bicycle and train simulators are not considered.

The driving simulator has a broad range of applications: in the purpose of the simulation as well asthe used type of simulator. One can find driving simulators at driving schools, psychological researchcentres, amusement parks, car manufacturers etc. This survey will make a subdivision in fidelity.

First attention is paid to motion simulation in general and everything that contributes to it. What ismotion simulation? How does motion cueing work?

Only motion based driving simulators are considered in this survey. Motion simulators without actu-ated motion are referred to as mid-level simulators [2] and consist of a car in front of a fixed screen.

Terminology and acronyms used throughout this report can be found in the Nomenclature at page iii.

1.2 History

Motion simulation started with flight simulation in the beginning of the 20th century. Flight trainingschool “Antoinette” developed the first flight simulator [3]. In 1948 Gough developed a parallel manip-ulator for the purpose of testing tires. It was not until 1962 when D. Stewart reintroduced the parallel6DOF (degree of freedom) system consisting of two platforms and 6 actuators [4]. The so called Stew-art platform was very well received in the world of motion simulation. It is a relatively compact design,comprising 6DOF-actuation and allowing relatively large payloads [5].

The first driving simulator designs consisted of few actuated DOF’s. 3DOF Designs were built bijVolkswagen (early 1970s) and the Swedish Road and Traffic Research Institute VTI, who were inspiredby the Volkswagen design [6].

Mazda [7] has build a 4DOF actuated simulator in 1985, inspired by Volkswagen. That same year,the first 6DOF actuated simulator comes from Daimler-Benz [8]. Throughout the 90s, several 6DOFactuated were built (FORD [9], JARI, BMW, Renault, WIVW, Nissan [10]).

At the North American Driving Simulation Conference(DSC) in 2003, the University of Iowa amazedthe world of motion simulation with their National Advanced Driving Simulator (NADS-1) [11]. It

2

Figure 1.1: Antoinette

123456

A A

B B

C C

D D

EIGE

NDOM

VAN

REX

ROTH

HYD

RAUD

YNE

B.V.

BOX

TEL.

VERM

ENIGVU

LDIGING

OF M

EDED

ELING

AAN

DERD

ENIN W

ELKE

VOR

M DA

N OO

K, IS

ZON

DER

SCHR

IFTE

LIJK

ETO

ESTE

MMING

VAN

EIGE

NARE

S NIET

GEO

ORLO

OFD.

GEBR

UIK

SLEC

HTS

TOEG

ESTA

AN IN

KAD

ER V

ANOP

DRAC

HT A

AN O

F VA

N RE

XROT

H HY

DRAU

DYNE

B.V

.

PROP

ERTY

OF

REXR

OTH

HYDR

AUDY

NE B

.V. B

OXTE

L,

THE

NETH

ERLA

NDS. R

EPRO

DUCT

ION, O

R DISC

LOSU

RE

TO T

HIRD

PAR

TIES

IN A

NY F

ORM

WHA

TSOE

VER, N

OTAL

LOWED

WITHO

UT W

RITT

EN C

ONSE

NT O

F TH

E PR

OPRIET

ORS. U

SE O

NLY

ALLO

WED

WITHIN

THE

SCOP

EOF

AN

ORDE

R TO

OR

FROM

REX

ROTH

HYD

RAUD

YNE

B.V.

0 100

P.O. BOX 32 · NL-5280 AA BOXTEL · THE NETHERLANDS

SIZE REV.

DIMENSIONS :

DRAWN :

CHECKED :

PROJ. :SCALE : MASS (kg) :

00A2PAGE 1 OF 1

Oktal

EMotion-4000-6DOF-700-MK2

Motion Base AssemblyDATE :

CRolink

WdE

mm

TITLE :

DOCUMENT :

1:20

02-13-1000-ADM-0000NL000648

PROJECT :

31-03-06

31-03-06

Bosch GroupRexrothRexroth Hydraudyne B.V.Systems & Engineering

MRP 1937

,1

PayloadC.O.G.

X

Z

X

Y

L.J.C.

PayloadC.O.G.

PayloadC.O.G.

Motion System Complete Retracted (Hardstop) Motion System Complete Extended (Hardstop)

1023

,513

0210

1748

Max Payload: 4000 kg Ixx: 18470 kgm^2 Iyy: 13857 kgm^2 Izz: 17001 kgm^2

Y

Z

X

Figure 1.2: Stewart Platform, or Hexapod

was the largest so far and consisted of a turntable, mounted on a hexapod, which was actuated inx- and y-direction using an xy-table. A complete car was fitted inside the dome. Several car andtruck simulators were built in the beginning of the 21st century (SimuSys, Mark III, TUTOR, Katech,SimCar, UoLDS, see Chapter 2) and some were upgraded (FORD, VTI-III, BMW, MARS RenaultULTIMATE, see Chapter 2). It was not until 2007, when the Toyota Motor Corporation came with adesign, which exceeded the NADS-1 in size. Whether the Toyota Simulator also exceeded the NADS-1in fidelity is an interesting issue (Chapter 3).

There is still no consensus on which motion system design suits the demands of a realistic drivingsimulator best. Therefore first motion cueing will be introduced and the purpose of driving simulatorswill be explained, before getting to driving simulators.

Over the past decades the use of hydraulic actuation has been replaced by electric servo technology inmany industrial applications [12]. This development is also noticeable in driving simulators. The im-provements in electric motor performance have been enabled by the development of new high energyproduct magnetic materials, advancements in low cogging, low torque ripple sinusoidal designs andthe availability of high resolution optical encoders [12].

1.3 What is Motion Simulation?

Simulation is defined as an imitation of some real thing, state of affairs, or process [13]. Motion simula-tion is all about perception. The human body has two inputs for motion perception: inertial stimulantson the body and environmental motion with respect to the body. The inertial stimulation stems fromthe gravitational force and the external forces and moments on the body [14]. The vestibular system,located in the inner ear (left and right), is the prominent sense that provides the perceptual systemwith information about linear and angular inertial accelerations of the body. The prominent senseof speed of the human body is obtained through the visual system. Sense of speed is also acquiredthrough the acoustic system.

Motion cueing describes the presentation of visual, acoustic, vestibular and haptic information (cues)with the aim to resemble real movements in virtual environments [10]. The most popular way of mo-tion cueing is the Classical Washout Filter (CWF). The principle of the Classical Washout Algorithmis shown in Figure 1.3. Only the high-frequency components of the translational and rotational ac-celerations are reproduced by the simulator. This is required to keep the simulator motion withinthe capabilities of the system. The visual system will provide the low-frequency information. For

3

the longitudinal and lateral acceleration, tilt coordination is used to mimic the sensation of sustainedacceleration [15]. Tilt coordination cannot be used for sustained heave motion.

High PassGain

1/g Low Pass

High Pass

Rate-Lim

High PassGain 1/s2

Translational accelerations

Rotational accelerations

Simulator Translations

Simulator rotations

Figure 1.3: Classical Washout Filter [1]

An extensive survey about motion perception is carried out by Van der Steen [14], where a modelof self-motion perception is presented. A driver should feel the same accelerations as he visuallyperceives. This is true to a certain extent. Threshold values determine an indifference zone. Thesevalues vary as a function of DOF, frequency and workload [16], but are determined to be typically atan angular velocity of ±3◦/s and a maximum acceleration of ±0.6 m/s2 [17]. Staying within this zoneand therefore below these values, allows the motion system to move without the occupant noticing it.

1.3.1 The Need for Motion

When people play computer games or watch movies on large or multiple screens they already feel asif they are part of the virtual world they are looking at. They sense the motion that is represented onthe screen. Then why is it required to have a large and expensive motion system to excite this virtualworld accordingly?

Discussion in the late 70s arose about the need for motion in simulation. The anti-motion positionconsisted of Jacobs and Roscoe, 1975; Woodruff, Smith, Fuller, and Weyer, 1976; Gray and Fuller, 1977;Martin and Waag, 1978; Pohlmann and Reed, 1978 [18]. P.W. Caro [19] notes that the anti-motionposition is not supported by evidence. McCauley concludes in his report [18] that motion contributesto the performance of the pilot, particularly for experienced pilots and that it is possible that motioncues may be beneficial for flight training in unstable aircraft and tasks involving disturbance cues,although the evidence is weak.

There are several reasons why a motion system is important in simulation, outlined by Piatkiewitz [15]and Advani [20]. The main reason is to prevent simulator sickness. Symptoms of simulator sicknessresult from a difference in the latency of the visual and motion system and false cues. Therefore themotion latency should be less or equal to the visual latency [21] and false cues should be avoided [15].

A driver uses his sensory inputs to obtain the required input to base his decisions on. The bandwidthof the required input signal should be in accordance to the driver’s task [15]. For cognitive decision,e.g. when to turn on the screen wipers or direction indicator, relatively low frequency information issufficient. When cornering on a busy road, high frequency information is required.

The presence of motion increases the workload of the driver. Accelerational cues and high frequencymotion add a feeling of realism. High frequency motion can be used to simulate special events likeroad rumble.

4

Automobile manufacturers also benefit from the use of a motion based driving simulator, especiallyin the early development phase of the car [22]. Evaluation of manoeuvres like lane change, slalom andbraking in turn are determined by subjective sensation, which is not available in computer models.Use of a motion system contributes in early decision-making in the design phase and allows integra-tion of suppliers in this phase. A variety of subjective evaluations, e.g. thresholds and aging effects,can be systematically analyzed.

1.4 Areas of use

Driving simulators are found in various areas of use; from entertainment to research and advancedtraining. This survey presents a subdivision in fidelity: low-level, mid-level and high-level fidelity.These simulator types are defined as follows.

At a low-level simulator, the driver sits in a car seat, preferably inside a car, which is fixed to the ground.This is called a fixed base (FB) simulator. The driver looks to a screen, which is fixed too (FS). Thescreen is designed such that it the view angle is as large possible. This can be done using multiplescreens, but preferably using a large convex screen. Steering wheels and pedals are often equippedwith force feedback and sound system takes care of the audio feedback.

Figure 1.4 Figure 1.5

The mid-level driving simulator consists of a car, which is accelerated in one degree of freedom (DOF).The screen can be fixed (FS) or can move along with the car. Force feedback is applied to the steeringwheel and a sound system takes care of the audio feedback. The actuated DOF at a 1DOF simulator isoften a y-sled, x-sled or a yaw-table.

A high-level driving simulator actuates the payload in at least 6DOF’s. The payload might be acceler-ated in additional DOF’s (introducing redundant DOF’s), e.g. to allow longer planar excursions. Thepayload consists often of a dome with a car(’s interior) inside of it and graphics projected such that aview angle of at least 220◦ is covered. The largest driving simulators consist of dome on a turntable,mounted on a hexapod, which is fixed to an xy-table.

As an alternative to FS, some simulators use a head-mounted display (HMD). HMD’s are found in anumber of simulation research applications [23].

5

1.4.1 Entertainment

Many simulators exist in entertainment industries and comprise mainly fixed screen simulators with0 or 1 actuated DOF. Figures 1.4–1.6 show some of these simulators. The benefit of the lower-levelsimulators should not be underestimated as technical development, driven by the military and enter-tainments industries, will ensure continual progress towards even higher fidelity systems [24]. Thisalso includes video games, which have the largest market share in entertainment. The increase ofprocessing power contributes to more realistic simulators in recent years [13].

1.4.2 Research

Driving simulators are used at research facilities for many purposes. The driving simulator has abroad range of applications, from studies concerning driver behaviour, the human-machine interface(HMI) and the effects of tiredness and drugs to projects concerning environmental issues, road andlandscape design, tunnel design, the reactions of the body, drivers with reduced functionality and newsubsystems in vehicles [25], [26].

In addition to studying driver training issues, driving simulators allow researchers to study driverbehaviour under conditions in which it would be illegal and/or unethical to place drivers. For instance,studies of driver distraction would be dangerous and unethical (because of the inability to obtaininformed consent from other drivers) to do on the road. A lot of fixed base simulators are used in

Figure 1.6 Figure 1.7: SHERPA

research and offer an alternative for research groups that do not require a hi-fi motion system. Thistype of simulator is cheaper in purchase, as well as in usage and maintenance.

Vehicle manufacturers operate driving simulators mainly to design the interior or to test AdvancedDriver Assistance Systems (ADAS). With the increasing use of various in-vehicle information sys-tems (IVIS) such as satellite navigation systems, cell phones, DVD players and e-mail systems [21],simulators are playing an important rule in assessing the safety and utility of such devices.

The effects of noise and vibrations on driver performance are examples of other areas that may bestudied using the simulator [22]. One exciting area is how the new technology influences driving, forexample the use of mobile telephones [26].

Simulators have also been used for road planning [25], e.g., in order to determine the positioning ofroad signs and for reasons of aesthetic design.

In the driving simulator it is possible to recreate any traffic situation as many times as you want,in an efficient, risk-free and cost-effective manner. One scenario may involve driving through the

6

countryside and suddenly being confronted by a large animal leaping out in front of the car. Anotherscenario could be in a city environment, with dense traffic, where cars unexpectedly cross the roadway.A certain scenario can also be recreated under different preconditions, to study, for example, thedifferences between driving with and without time pressure.

1.4.3 Training

Driving simulators are being increasingly used for training drivers all over the world. Research hasshown that driving simulators are proven to be excellent practical and effective educational tools toimpart safe driving training techniques for all drivers. They can be economical in cases where thealternative (real object) is much more expensive. An advantage is the reduction of risk as the driver isoften put in a complex scenario [27], which might be hazardous when applied in non-virtual environ-ment, i.e. real environment. Another advantage is the ability to monitor students and that they canbe advised as they practice [28]. There are various types of driving simulators that are being used astraining simulators, like train simulators, truck simulators, bus simulator, car simulator, etc. [21].

Figure 1.8: UoLDS Figure 1.9: VTI-III

7

Chapter 2

Driving Simulators

This section will discuss some of the most high-level fidelity, human-in-the-loop, moving base drivingsimulators. The simulators are presented in a chronological order.

A lot of low-level FB/FS simulators exist in psychological research. Most of them are described byInrets [29]. The development of 6DOF/FS simulators started in 1999 with a design of Renault, [30]),but still 6DOF/FS simulators are built.

- The very FTM Truck-simulator [31] at the Munich Technical University (2004)

- SHERPA, built by LIMAH [32]

- The Pennsylvania Department of Transportation uses a Truck Driving Simulator [33] at theirVehicle Simulation Research Center (VSRC) very recently a truck simulator;

- and many others are listed [29].

Unless the 6DOF/FS is considered a high-level simulator, it won’t be further discussed, in order tofocuss on the more sophisticated simulators.

Volkswagen The first driving simulator was built by Volkswagen [34] in the early 70s. It concerneda car on a 3DOF motion system. The motions were driven by a turntable (yaw) and a roll andpitch mechanism. A single, flat screen is mounted in front of the driver sitting on its seat at aplatform. No further car interior is created at the platform.

Figure 2.1: IFAS (1984)

Institute for Automotive and Power Train Engineering

Automotive Engineering

Prof. Dr.-Ing. Martin Meywerk

Ins t i tu t für Fahrzeugtechnikund An t r i ebssys temtechn ik

Institute for Automotive and Power Train Engineering

Automotive Engineering

Prof. Dr.-Ing. Martin Meywerk

Ins t i tu t für Fahrzeugtechnikund An t r i ebssys temtechn ik

Driving Simulator MARS

The driving simulator MARS (Modular Automotive Research Simulator) at the IFAS is based on an electrically driven hexapod motion-system, which is supported by three pneumatic actuators for carrying the static load. This motion system is mounted on a moving sledge, which provides via an also electrically driven actuator for linear motion a 7th degree of freedom with a parallel translation of up to 1500 mm. In combination with the drivers cab and the visual system on the upper frame an interactive simulation of the drive of wheeled vehicles on- and offroad is possible. For calculating the vehicle dynamics in real-time the simulation tool ORSIS is used. This program has been developed at the institute.

The driving simulator MARS

Specifications

position velocity acceleration

Surge ± 0.48 m (± 0.75 m) ±0.60 m/s ±7.0 m/s2

Sway ± 1.17 m ±0.60 m/s ±7.0 m/s2

Heave -0.31 / +0.35 m ±0.50 m/s ±7.0 m/s2

Roll ±21.5 deg ±30 deg/s >140 deg/s2

Pitch -23.6 /+27.4 deg ±30 deg/s >140 deg/s2

Yaw ±23.7 deg ±30 deg/s >140 deg/s2

Fields of Application The following listing gives an overview over fields of application of the simulator in research at IFAS: - Simulation of on– and offroad drive - Research into interaction of visual and vestibular information - Development of motion systems - Development of motion algorithms (motion cueing, wash-out) - Ergonomic testing (legibility of instruments) - Investigations in controlling behaviour of the human being (e.g. reaction tests) - Physiological studies (stress in road traffic, alcohol, drugs) - Road planning (design and layout of roads, sign posting) - Previous Research a)

The picture on the left shows the influence of the 7th actuator on the presentation of a sway acceleration. It is evident that because of the major traverse path the superimposition of high-frequent parts, represented by a horizontal movement, and low-frequent parts, represented by inclination, improves.

b) The diagram shows results of an study, where the influence of the visual information on the perception of pitch angles has been analysed. The picture presented to the test subject consisted in a straight road. It was possible to present sight and motion independently from each other.

0

1

2

3

pitc

h an

gle

4

5

6

[deg

]

3,86 4,00

c)

0 200 400 600 800

time [s]

0

50

100

150

200

250

300 skin resistance response (SRR) [kOhm/s]

skin resistance level (SRL) [kOhm]

pulse frequency [1/min]

Using the driving simulator MARS further investigations have been carried out. Here the physiological reactions of a human being on the occurrence of certain events during a predefined driving task have been analysed. The diagram shows a typical run of measured parameters while passing through a circuit using the simulation tool ORSIS.

Visitors address: Holstenhofweg 85, H8 22043 Hamburg

Phone: (040) 6541 - 2728 / Sekr.: 2159 Fax: (040) 6541 - 2742 E-mail: office-ikk@unibw-hamburg.de

E-mail: martin.meywerk@unibw-hamburg.de http://www.unibw-hamburg.de/MWEB/ikk/fkw/fkw.html

Visitors address: Holstenhofweg 85, H8 22043 Hamburg

Phone: (040) 6541 - 2728 / Sekr.: 2159 Fax: (040) 6541 - 2742 E-mail: office-ikk@unibw-hamburg.de

E-mail: martin.meywerk@unibw-hamburg.de http://www.unibw-hamburg.de/MWEB/ikk/fkw/fkw.html

0 2500 5000 7500 10000

time [ms]

0

0.5

1

1.5

2se

nso r

ed s

way

acc

eler

atio

n [m

/s2]

input

6-DOF-system

7-DOF-system

input

6-DOF-system

7-DOF-system

amplified sight

normal sight

reduced sight

different sight

unmoved sight

inverted sight

type of visual information

1,421,68

2,292,45

Figure 2.2: MARS (2004)

8

IFAS In 1984 IFAS [35], former IKK, produced a 1DOF simulator with the visuals projected in abox in front of the driver. The actuated DOF was a hydraulically driven y-sled (Figure 2.1). Thesimulator is still used at the University of the German Armed Forces in Hamburg.

IFAS extended their simulator with 6DOF’s in 2004 with the MARS Driving Simulator [36] atHelmut Schmidt University. A y-sled is present in the new design, but has a hexapod mountedon top of it (Figure 2.2).

VTI The Swedish Road and Traffic Research Institute VTI is an active player in the world of drivingsimulation and swear by a y-sled in their designs. They started in 1984 with the VTI-I [34],a 4DOF simulator. A half car with a screen fixed in front of it on a motion platform is to beaccelerated in roll, pitch and yaw, on a y-sled (sway).

A the end of the 1980s VTI rebuilds their VTI-I to the VTI-II, by request of the Swedish insur-ance company Trygg Hansa who requires a truck simulator, which has a higher payload than thepassenger car simulator [37].

In 2004 this design is upgraded to the VTI-III [6]. The VTI-III (Figure 1.9) is increased in sizeand equipped with a vibration table, allowing high frequent road rumbles to be experienced bythe driver.

Daimler-Benz The initial “Daimler-Benz” Driving Simulator was first introduced at 1985 [8] andwas the first driving simulator to be driven in 6DOF. An hydraulic hexapod, which was a specialdesign for this simulator, realized the largest motion envelope at that time. A car or truck cabinis situated inside a dome on which 6 CRT projectors display a 180◦ field of view.

In 1993 the simulator was upgraded to the “Advanced Driving Simulator” [22]. The main differ-ence to the previous design was the extension of the motion system in lateral direction. Anotherhydraulic cylinder realized a 5.6 m excursion in lateral direction (sway). This modification thequality of the simulator was improved considerably [8].

Mazda Mazda introduced their 4DOF driving simulator in 1985 to decrease the number of trafficaccidents, which grew rapidly with the spread of motorization [7]. The design is very similar tothe VTI simulators.

Ford Ford introduces the Virttex [9] in 1994, a dome on a hydraulic hexapod. The Virttex is renewedin 2001 [9] (Figure 2.4).

Figure 2.3: Ultimate (2004) Figure 2.4: Vitrtex (2001)

BMW BMW develops a 4 m high, hydraulic hexapod, with a small screen mounted onto the mo-tion platform, together with a full-size car [38]. This system is completely rebuilt in 2003 [38].The platform is now provided with a dome and the driver enters the simulator through a tun-nel/catwalk, to give the driver the idea he enters a car and not a simulator (Figure 2.5).

JARI, Nissan and WIVW Other 6DOF, hydraulic simulators are built in this period by JARI(1996) and Nissan (1999). The hexapod of IZVW/WIVW [39] built in 1999 is statically compen-sated using three additional cylinders (Figure 2.6). The hexapod doesn’t seize the payload at the

9

Figure 2.5: BMW (2003) Figure 2.6: WIVW 1999

bottom, but at the height of the driver’s head. The idea behind this concept is, that it is easier totilt the payload around the driver’s vestibular system, i.e. inside his head.

NADS-1 In 2002 the North American Driving Simulator (NADS-1) is presented at the Universityof Iowa [11]. At that time it is by far the most advanced driving simulator. The NADS-1 2.7 is a9DOF simulator, consisting of an xy-table on which a hexapod travels. On top of the hexapod, aturntable is mounted, which provides yaw-acceleration. A dome, with full-size car inside, rotateson top of the turntable.

SimuSYS SimuSYS, developed in 2003 [40], is a relatively small 4DOF simulator, yet it is also rel-atively high. The stacking of the different subsystems, which actuate a different DOF each,is constructed such that DOF can be actuated independently, but at the expense of simulatorheight. Therefore an error in the first DOF can be seen again in each following DOF.

Renault In 2004 Renault replaced its 6DOF/FS system with a hexapod on an xy-table, called theULTIMATE [30]. The design of the motion system is carried out by Bosch–Rexroth.

Tutor TUTOR is a combined bus and truck simulator [41], for professional driver training. It isdeveloped by Lander Simulation & Training Solutions and Installed at INTA (Spain) in 2004.A similar design is found at the truck simulator Mark III from Transim (2005), which is anupgrade from the FB/FS Mark II. It is used for truck driver training at MPRI (US), to improvedriving behaviour and skill.

Figure 2.7: NADS-1 (2003) Figure 2.8: Toyota (2007)

Katech, SimCar In the year 2005 two advanced 6DOF driving simulators were built. Katech in-troduced the Katech Advanced Automotive Simulator KAAS [42] and at the German AerospaceInstitute (DLR) an “inverted” hexapod was built by SimCar [43]. The inverted hexapod holdson to the payload at its top side. This was done to have the rotation point of the simulator ata higher level, as stated at MSC2007. The rotation point of a hexapod, however is virtual andcan be placed anywhere within the hexapod’s range by means of coordinate transformation inmotion software.

10

UoLDS The University of Leeds owns its own UoLDS (Figure 1.8) since 2006 [26] and claims to beone of the world’s most advanced driving simulators in research environment [44]. The researchin which the simulator is currently employed, involves intelligent speed adaptation, effects of au-tomated systems on safety and improved driver comprehension on traffic management signing.

Toyota In 2007 the NADS-1 simulator is exceeded in size by the Toyota Driving Simulator (Fig-ure 2.8), built at Toyota’s Higashifuji Technical Center in Susono City [45]. The design is verysimilar to the NADS-1, but then larger, and the main difference is found in the turntable. At theToyota Driving Simulator, the car yaws inside the dome, whereas the NADS-1 yaws the entiredome, with the car inside of it. The simulator will be used for driving tests that are too danger-ous to conduct in the real world, such as the effect of drowsiness, fatigue, inebriation, illnessand inattentiveness.

11

Chapter 3

Discussion

There is still no consensus on which motion system design suits the demands of a realistic drivingsimulator best. Generally every proud owner of a motion simulator claims to have the best simulatorsuitable for its purpose. This announced issue is a very critical topic and it should be noted thatchanges in fidelity affect training effectiveness [46]. As it comes to driving simulation one has towonder what the simulator will be used for [20]. What does the training of research purpose requirefrom a driving simulator? Unfortunately economical and political factors play a role in the simulatordesign as well and often drastically affect the design. An advanced simulator requires motion, butshould be applied very accurately, to prevent false cues [15]. Large excursions, high accelerations andhigh frequent motions contribute to a realistic perception of motion. Large excursions allows the MCAto use the tilt coordination more gradually.

12

Bibliography

[1] Su-Chiun Wang and Li-Chen Fu. Predictive washout filter design for vr-based motion simulator.Taiwan, 2004. National Taiwan University.

[2] David H. Weir and Allen J. Clark. A survey of mid-level driving simulator. Number 950172. SAEpaper, 1995.

[3] L’Aerophile Collection. Ecole de pilotage “antoinette”. In L’Aerophile Collection. M. Branger,Mourmelon-le-Grande, France, 1911.

[4] D. Stewart. A platform with six degrees of freedom. IMechE, 180(15):371–385, 1965.

[5] S.K. Advani. The Kinematic Design of Flight Simulator Motion-Bases. PhD thesis, Delft Universityof Technology, Delft, 1998.

[6] S. Nordmark, H. Jansson, and H. Sehammar G. Palmkvist. The new vti driving simulator. multipurpose moving base with high performance linear motion. Paris, 2004. Driving SimulationConference 2004.

[7] Kenichi Yoshimoto and Takamasa Suetomi. The history of research and development of drivingsimulators in japan. Tsukuba, May/June 2006. DSC/Asia Pacific.

[8] Joerg J. Breuer and Wilfried Kaeding. Contributions of driving simulators to enhance real worldsafety. Tsukuba, 2006. DSC Asia/Pacific.

[9] J. Greenberg, R. Curry, M. Blommer, K. Kozak, B. Artz, L. Cathey, and B. Kao. The validity oflast-second braking and steering judgments in advanced driving simulators. Paris, France, 2006.DSC2006.

[10] M. Fischer. Motin platform technology and motion cueing algorithms. Braunschweig, 2007.MSC2007.

[11] Chris Schwarz, Tad Gates, and Yiannis Papelis. Motion characteristics of the national advanceddriving simulator. Michigan, US, 2003. DSC-NA2003.

[12] Yvan Pelchat, Denis Pelletier, and Peter Jarvis. Application of electromechanical motion systemson level d simulators. Montreal, Canada, 2006. CAE Inc.

[13] Anonymous. Wikipedia. http://en.wikipedia.org/wiki/Simulation, 2008.

[14] F.A.M. van der Steen. Self-motion perception. PhD thesis, Delft University of Technology, June1998.

[15] Philippe Piatkiewitz. Motion Cueing Tuning Manual. Bosch – Rexroth, Boxtel, The Netherlands,2004. Internal document.

[16] Sunjoo K. Advani and Ruud J.A.W. Hosman. Revising civil simulator standards – an opportunityfor technological pull. (6248), 2006.

13

[17] Dagdelen M., Reymond G., Kemeny A., Bordier M., and Maïzi N. Mpc based motion cueingalgorithm: Development and application to the ultimate driving simulator. 2004.

[18] McCauley. Do army helicopter training simulators need motion bases? Technical report, U.S.Army Research Institute for the Behavioral and Social Sciences, 2006.

[19] P.W. Caro. The relationship between flight simulator motion and training requirements. HumanFactors, 21(4):493–501, 1979.

[20] S.K. Advani. Private conversation, 2008.

[21] Anonymous. Wikipedia. http://www.wikipedia.org, 2008.

[22] Ludger Dragon. The vehicle function culture of driving in the early development phase.MSC2007, September 2007.

[23] Geoff Blackham. Visual display systems for car and truck simulators. West Sussex, England,1999. SEOS Displays Ltd.

[24] A.M. Parkes and A. Flint. Degree of simulation, or fidelity of simulation: relative contribution tosynthetic training. Paris, 2004. Driving Simulation Conference 2004.

[25] Vti. http://www.vti.se/, 2008.

[26] Hamish Jamson. Driving me round the bend – behavioural studies using the new university ofleeds driving simulator. Braunschweig, 2007. MSC 2007.

[27] Marcos Fernández, Inmaculada Coma, Gregorio Martín, and Salvador Bayarri. A hierarchical ob-ject oriented architecture for the management of complex driving simulation scenarios. Valencia,Spain, 1999. ARTEC -Robotics Institute – University of Valencia.

[28] Winfried Tomaske. A modular automobile road simulator mars for on and off-road conditions.Hamburg, Deutschland, 1999. Universität der Bundeswehr Hamburg.

[29] Institut national de recherche sur les transports et leur sécurité Inrets. Driving simulators.http://www.inrets.fr, 2008.

[30] Mehmet Dagdelen, Jean-Charles Berlioux, Francesco Panerai, Gilles Reymond, and Andras Ke-meny. Validation process of the ultimate high-performance driving simulator. Paris, France,2006. DSC2006.

[31] Technische Universität München. Ftm driving simulator, aktive sicherheit durch fahrerassitenz.http://www.fahrzeugtechnik-muenchen.de/content/view/11/64/, 2008.

[32] Simulateur de conduite automobile de l’université de valenciennes. http://www.univ-valenciennes.fr/simulateur–sherpa/, 2008.

[33] Vehicle simulation research center (vsrc). http://www.vss.psu.edu/VSRC.htm, 2008.

[34] Staffan Nordmark. Driving simulators, trends and experiences. Paris, France, 1994. DrivingSimulation Conference 1994.

[35] T. Fortmüller. The development of driving simulation at ifas. Soesterberg, The Netherlands,2007. Motion Cueing Workshop.

[36] Martin Meywerk. Driving simulator mars. http://www.hsu-hh.de/meywerk/, 2008.

[37] S. Nordmark, H. Jansson, and H. Sehammar G. Palmkvist. The new vti driving simulator. multipurpose moving base with high performance linear motion. Paris, 2004. Driving SimulationConference 2004.

14

[38] Alexander Huesmann and Josef Nauderer. Applications to driving simulation and their require-ments to the tool. Braunschweig, 2007. Motion Simulation Conference 2007.

[39] Wuerzburg institute for traffic sciences gmbh. http://www.wivw.de, 2008.

[40] SimuSys. Innovative high-performance motion simulation system for entertainment, researchand training applications. Synthesis Report CRAFT – IST – 1999 – 56418, University ofZaragoza, Zaragoza, Spain, 2003.

[41] Lander Simulation & Training Solutions. Tutor. http://www.landersimulation.com, 2008.

[42] Si bok Yu, Soo-Young Lee, Young-Wook Son, Hee-Seok Moon, and Hyeong-Ju Noh. The firststage development of katech advanced automotive simulator. Tsukuba, 2006. DSC Asia/Pacific.

[43] M. Brünger-Koch. Motion parameter tuning and evaluation for the dlr automotive simulator.Orlando, Florida, 2005. North American Driving Simulator Conference 2005.

[44] Hamish Jamson. University of leeds driving simulator. http://www.its.leeds.ac.uk/facilities,2008.

[45] John Challen. Reality bytes. Driving Simulators, pages 50–53, March 2008.

[46] J.C.F. de Winter, P.A. Wieringa, J. Dankelman, M. Mulder, M.M. van Paassen, and S. de Groot.Driving simulator fidelity and training effectiveness. Delft, Netherlands, 2007. Delft Universityof Technology.

15