Road Safety at Intersections Controlled by Traffic Lights

IVC and Risk Indexes

Bruno Dalla Chiara, Francesco Paolo Deflorio and Serena Cuzzola

Politecnico di Torino, Department DIATI – Transport Engineering, C.so Duca degli Abruzzi, 24-10129 Torino, Italy

Keywords: ADAS, Safety Index, Intersection Safety, Driver Reaction Time.

Abstract: The paper reports the results of safety analyses conceived to assess the effects and benefits which might be

generated by the forthcoming use of the infrastructure-to-vehicle (I2V) or vehicle-to-infrastructure (V2I)

communication systems at road intersections regulated by traffic lights. Road crossings are often considered

as critical areas for the occurrence of accidents, because they increase the likelihood of the event given the

confluence of traffic streams from and to different directions. The analyses are aimed at calculating a real-

time estimate of some risk indexes of accident, which might be provided on-board when approaching road

intersection regulated by traffic lights. This information can then be used by an ADAS for traffic signal

approaching. Two typologies of use of the information on the risk indexes can be identified: if data can be

detected in real time, the driver could be informed on-board of a potentially hazardous situation using

algorithms to predict the trend of the vehicle on the basis of the data detected from the monitoring; another

use would be detecting – in case the vehicle were already within the dilemma zone – the lowest risk

manoeuvre and sending a message on board to inform the driver.

1 INTRODUCTION

Quantifying the road safety risk and the effects that

the Advanced Driver Assistance Systems (ADAS)

can generate on it – i.e. the combined value of

reducing the likelihood of an accident and its

severity – is a very sensitive theme, which can today

resort to the infrastructure-to-vehicle and/or vehicle-

to-vehicle communication systems. Together with

the interest towards the quality and energy efficiency

of transport, safety is contributing to the fast

spreading of Intelligent Transport Systems (ITS),

which include – amongst the different technologies –

the ADAS. Such perspective is part of the migration

– which has been in progress for years – from the

mere passive safety to the study of systems, tools

and applications which can ensure active, preventive

and post-crash safety.

Within this context, attention is progressively

concentrating on the cooperative systems, which can

interact to one another, thus setting up

communication between the different vehicles

(Vehicle–to–Vehicle, V2V) or between the vehicles

and the infrastructures (Vehicle-to-Infrastructure,

V2I, or Infrastructure-to-Vehicle, I2V) to create ad

hoc communication networks. This paper focuses on

the effects of integrating such communication

systems with the ADAS with the aim of improving

road safety; special attention is paid to the safety of

the road users, in order to reduce both the number

and severity of the road accidents. The

communication technologies between vehicles and

between infrastructure and vehicle are suitable to

intervene at the pre-crash stage, i.e. in emergency -

assistance, where the action of the driver could still

prevent the accident or reduce its risk.

It worth reminding a basic definition of crash,

slightly reviewing the one which was proposed in

Dalla Chiara, Deflorio and Diwan (2009): the crash

phase of an accident occurs when the perception-

reaction time of a driver plus the time necessary to

actuate the procedure (e.g., braking) of the vehicle

he/she is driving is greater than or equal to the time

involved by the exogenous variation that occurs

outside the vehicle; such a perception-reaction time

of the driver plus that of the vehicle is therefore the

maximum time available for the driver to respond to

an emergency condition on the road and prevent an

accident.

The road intersections are often considered as

critical areas for the occurrence of crashes, because

they increase the likelihood of the confluence of

569

Dalla Chiara B., Deﬂorio F. and Cuzzola S..

Road Safety at Intersections Controlled by Trafﬁc Lights - IVC and Risk Indexes.

DOI: 10.5220/0004165105690577

In Proceedings of the 9th International Conference on Informatics in Control, Automation and Robotics (IVC&ITS-2012), pages 569-577

ISBN: 978-989-8565-22-8

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

traffic streams from and to different directions.

2 STATE OF THE ART

During the last years, car manufacturers and

researchers experimented many ADAS (MacNeill,

and Miller, 2003); (Maile and Delgrossi, 2009).

These systems are in-vehicle technologies that

provide support to various aspects of the driving task

and they are supposed to improve traffic safety and

traffic efficiency. In this field, the most famous and

deployed ADAS systems are the adaptive cruise

control (ACC) and the intelligent speed adaptation

(ISA), collision avoidance systems, adaptive light

control, lane departure warning, driver vigilance

monitoring, pre-crash vehicle preparation and

parking aid (Tapani, 2009); (Monteil et al., 2011).

In order to estimate the future impact of the

ADAS development process from its very early

stages, some studies were based on the use of

microscopic traffic simulation. Torday et al. (2003)

proposed to integrate the output of this tool with a

safety indicator, evaluated during the micro

simulation process. The microscopic level of traffic

description grants the opportunity of knowing the

relative position of the vehicles, their speed and

deceleration. All of these parameters thus enable the

computation of a safety indicator useful to compare

scenarios where ADAS are activated for vehicles.

Other authors (Morsink et al., 2008) provide an

overview of micro-simulation modelling for road

safety impact assessment of ADAS. Recent literature

and expert opinions identify driver behaviour sub-

models and road safety indicators as key

components. In Benz et al. (2006), several existing

models – on both the micro and macro scales –

would be adapted and used to assess safety related

effects of ITS measures. Examples of such measures

include but are not limited to ADAS and IVIS.

While the micro-models would determine the

individual vehicles' safety related behaviour, the

macro-models would investigate the network-wide

aspects.

In order to enhance the performance of micro

simulator for safety analysis, a Surrogate Safety

Assessment Model (SSAM) has been developed (US

DOT-FHWA, 2009). This technique combines micro

simulation and automated conflict analysis, which

analyses the frequency and type of narrowly averted

vehicle-to-vehicle collisions in traffic, to assess the

safety of traffic facilities, without waiting for a

statistically valid number of crashes and injuries to

actually occur. Applications of this method to road

intersection scenarios are reported in Gettman and

Pu (2006), Klunderet al. (2006) and Ki-Joon and

Jaehoon (2009). An assessment of the driver

behaviour at dilemma zone (Liu, Herman and Gazis,

1996) and of the effectiveness of safety indicators

based on the traffic conflict technique at intersection

is reported in Archer (2005) and Hurwitz (2009).

Recent international research projects have been

investigating both vehicle-based and road-based

monitoring. The European projects SAFESPOT,

COOPERS, CVIS and COVEL aimed at improving

road safety by using intelligent vehicles

interconnected to each other through a vehicular ad-

hoc network (VANET).

As regards the V2V and V2I communication

systems and their relationships with safety and

ADAS, they are a primary means for supplying

information to drivers. In recent years, V2V and V2I

communication systems have been submitted to

intensive studies, also applied to safety at

intersections (INTERSAFE-2).

In this field, FOTSIS was a large-scale field

testing of the road infrastructure management

systems needed for the operation of seven close-to-

market cooperative I2V, V2I & I2I technologies (the

FOTsis Services), which allowed assessing in detail

both their effectiveness and their potential for a full-

scale deployment in European roads.

We need to recall that the response time of a

driver can be split into a mental processing and

reaction time and a muscular time . The former

includes the time from the perception of the external

stimulus to the brain’s message to the foot to brake.

This implies the awareness of the hazard, the

emotive response and the reaction itself. The

muscular reaction time is needed for the right foot to

move onto the brake pedal. The driver’s reaction

time is influenced by quick or slow reflexes, by

his/her experience as well as by the complexity of

the dangerous scenario that has to be faced. On the

basis of tests and literature, the median perception-

reaction time of a driver results to be 0.66s,

measured under normal highway driving conditions,

with some degree of braking expectation, since the

drivers were expecting the event to happen. From

the moment the driver puts his/her foot onto the

pedal, almost 0.1s pass (inertia of the system) before

the brake starts operating; this value may increase to

0.4s in the case of slow and older braking

equipment.

The diagram reported in Johansson and Rumar

(1971), as well as on ISO technical standards and in

Dalla Chiara et al. (2009), in revised editions,

represents the distribution of a driver’s brake

ICINCO2012-9thInternationalConferenceonInformaticsinControl,AutomationandRobotics

570

perception-reaction time between 0.2 and 2.1s. The

percentile of perception-brake response times

for these same conditions was 2.0s. The findings

from this study are consistent with the relevant

literature: most drivers are capable of responding to

an unexpected incident in 2.0s or less. Thus, the

perception-reaction time of 2.5s, adopted by the

American Association of State Highway and

Transportation Officials for design reasons,

encompasses most of the driving population.

A driver who might need 0.3s of perception-

reaction time under alerted conditions might need

1.5s under normal conditions; such response time

may decrease by approximately 1s or more in an

expected situation: IVC warning systems allow one

to pass from an unexpected to an anticipated

situation, and thus influence the perception-reaction

time.

3 SAFETY ANALYSIS AT

INTERSECTIONS

This article shows the results of the analyses

developed on the effects and benefits which would

be potentially generated by the forthcoming use of

the infrastructure-vehicle (I2V) or vehicle -

infrastructure (V2I) communication systems at the

road intersections regulated by traffic lights (Fig. 1):

a theme that – as it has been highlighted – is

extensively being dealt within the literature.

Figure 1: Evocative image of I2V and V2I at a generic

intersection regulated by traffic lights (INTERSAFE-2).

The effects of the use of the I2V systems are

assessed through the proposal of indicators on the

likelihood and/or severity of the risk, which can

timely and preventively indicate potentially critical

conditions and send more or less intensive alarm

messages – depending upon the criticality– on board

the vehicles which are potentially involved by means

of the I2V communications.

For the sake of completeness of the analysis, we

also developed some proposals for the combined use

of sensors to monitor the vehicles which approach

the intersections (US Dept. of Transportation, 2008).

Our analyses assume that the use of I2V systems

would match the increased level of attention of the

driver and – consequently – the dampening of the

perception-reaction-actuation time (t

p,r,a

) of the

driver, with the subsequent increased safety margin.

The processing concerns the study of the driver’s

behaviour when the yellow light is triggered. All

such processes associate the use of the I2V

technologies to the maximum perception-reaction -

actuation time of the driver (t

p,r,a

) which – in case of

low levels of attention – has been estimated to 2.3

seconds (where 0.8s can be assumed for the

actuation phase), on the basis of Johansson and

Rumar’s distribution (1971). The studies are based

upon the assessment of the variation of new road

risk indexes as a result of the reduced t

p,r,a

time;

specifically, it is assumed that the I2V

communications are such to send on board

indications which can supply two levels of alert,

namely: the former can take the attention of the

driver back to normal levels (t

p,r,a

equal to 1.46s:

green arrow in Fig. 3.) and the latter can generate an

actual alert (t

p,r,a

equal to 1.1s: red arrow in Fig. 2).

Figure 2: Johansson and Rumar’s distribution of the t

p,r

highlighting the two values of t

p,r

which give t

p,r,a

=1.46s

(green arrow) and 1.1s (red arrow).

4 THE DILEMMA ZONE AND

ROLE OF INTEGRATED

I2V-ADAS

It is worth reminding – first of all – the concept of

dilemma zone, which has been most likely

introduced for the first time in (Liu, Herman and

Gazis, 1996). The so-called dilemma zone is the

portion of approach to the intersection the driver

might cover, starting from the time when the traffic

0.0%

0.6%

4.4%

19.6%

48.3%

64.8%

72.9%

81.6%

88.5%

90.7%

94.7%

96.0%

98.1%

98.8%

99.1%

99.4%

100.0%

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

100

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1

Percentile

Number of people, f

Time [s]

Median time value: 0.66''

RoadSafetyatIntersectionsControlledbyTrafficLights-IVCandRiskIndexes

571

light turns into yellow, without being able to either

stop in safety conditions before the stop line (or

close to it) or to fully clear the intersection at the end

of the yellow light or when the red one is triggered;

such conditions are critical and generate an actual

dilemma to the driver, who does not know what

his/her behaviour should be in order to act safely,

not to commit infractions or cause accidents. Such

area can be eliminated with a proper yellow time

calculation and if vehicle speed is lower than the

established limit, but sometimes it exists and its

position and length vary depending upon the cases

and some parameters need to be taken into

consideration.

In order to clarify the concept of dilemma zone,

the behaviour of a driver is considered

independently, i.e. irrespectively on the one adopted

by the drivers of any vehicles which precede his/her

own one. When the yellow light is triggered, the

driver is faced with a choice: should he/she stop the

vehicle or cross the intersection – even by

accelerating – so that he/she can clear the area

before the red light? We should keep in mind that –

usually – the driver does not know how long the

yellow light will last or the so-called clearance time,

i.e. the all red time. The solution depends on factors

which characterize the distance and time required to

stop the vehicle and/or clear the intersection: the

initial speed of the vehicle, the actual or possible

deceleration, the driver’s perception and reaction

time, the distance between the stop line and the

access, the position of the vehicle when the yellow

light is triggered and the extension of the

intersection. It is obvious that – as a tendency – the

drivers who are far from the intersection choose to

stop; those who are very close to it – instead –

normally try to clear the intersection and therefore –

if required – they accelerate.

In either case, the characteristics of the

manoeuvres are influenced by the perception –

reaction – actuation time (t

p,r,a

) of the driver.

The stopping distance (Xs or d

) is the minimum

level of the distance, calculated from the stop line, a

vehicle should be within in order to have a

comfortable stop and in full safety conditions

(beyond such position, the vehicle cannot be

stopped: Cannot stop in Fig. 3). By steady

deceleration, the stopping space can be calculated

through a known ratio (1).

tvdX

arpas

⋅

+⋅==

(1)

where:

 Xs is the stopping distance or stopping space of

the vehicle [m];

 v is the initial speed of the vehicle [m/s];

 tp,r,a is the perception – reaction – actuation time

[s];

 a is the deceleration [m/s

The clearance distance (Xc or d

) is the maximum

distance from the stop line below which a vehicle

can clear the intersection in full safety conditions

(Cannot go in Fig. 3) within a given yellow light

time, which – though - he/she does not know. This

was computed through ratio (2).

Xc=d

=v*(Y+R)+a*(Y+R-t

)2/2-(W+lv) (2)

where:

 Xc is the clearance distance in meters;

 W is the length of the intersection measured from

the stop line of the access which is considered in the

opposite angle, depending upon the manoeuvre to be

performed; it is expressed in meters;

 lv is the length of the vehicle, in meters;

 v is the speed the vehicles approaches the

intersection at, expressed in [m/s];

 Y is the duration of the yellow light phase

(yellow light time) relevant to the access which is

being taken into consideration, expressed in seconds;

 R is the duration of the all red stage (all red

time), in seconds;

 tp,r,a is the perception – reaction – actuation

time;

 a is the time of the acceleration (assumed as

constant) adopted to clear the intersection. In default

of more accurate data, such as the ones generated by

monitoring, the value of this parameter is assumed

through Gazi’s equation (FHWA, 2006), i.e.:

a[m/s

] = 4,9-(0,213*v[m/s])

Three different conditions can be generated on the

basis of the relationship between the two distances

which have been defined above, namely:

1. Xs>Xc

2. Xs = Xc

3. Xs<Xc

In the first case (Xs>Xc), the dilemma zone results

from the overlapping of the Cannot Stop and Cannot

Go portions. The position and length of such areas –

when existing – vary from case to case.

The second case (Xs = Xc) represents an ideal

situation: the dilemma and optional zones disappear;

a driver which would find him/herself in those

conditions could stop the vehicle or clear the

intersection comfortably and in full safety

conditions, with no doubts at all on the behaviour to

ICINCO2012-9thInternationalConferenceonInformaticsinControl,AutomationandRobotics

572

be adopted.

Figure 3: Graphic representation of the zone where the

vehicle cannot stop in safety conditions (Cannot Stop) r

cannot clear the intersection in full safety conditions

(Cannot Go).

In the last case (Xs<Xc) an Optional Zone would

generate, i.e. a portion of the access lane where the

driver of the vehicle in it may select whether to stop

comfortably and safely at the stop line or to clear the

intersection in safety conditions.

It is worth mentioning that the dilemma zone

depends on the kinematic parameters of the vehicle

(i.e. speed, deceleration or acceleration) besides on

the yellow light time, which is generally the same

for all the accesses of the intersection. The most

appropriate strategy to minimize the issue caused by

the presence of the dilemma zone consists of

determining a yellow or all red time which allows

clearing the intersection from the limit position

available to stop. Nevertheless, the variability in the

conditions of motion, of the drivers and adherence of

the carriageway might determine different

circumstances than the ones which are defined a

priori. These variations can be observed by means of

position detection systems located either on-board

the vehicle (GPS-with WAAS, as EGNOS) or on the

infrastructure (VIP, Inductive Loops, W SN based

on magnetometers, etc.).

The analyses illustrated hereinunder are aimed at

providing a real time estimate of the risk of accident

for an approach of road intersections regulated by

traffic lights: this information can then be used by an

ADAS, which supplies the driver a risk indicator of

the instrument panel; such indicator should be able

to resort to information which is usually not

available to the driver or which – in any case –

he/she cannot calculate in real time, namely: the

road in front of him/her (navigator instrument panel)

the residual time to the triggering of the red light and

the clearance time (I2V), the comparison between

the driving dynamics and the safe crossing or

stopping conditions. This would allow assessing

whether or not a situation is hazardous and – if it is –

trying to avoid the potential collision by transmitting

alert messages to the potentially involved vehicles.

In the analysis of the safety conditions, we have

applied risk indexes formulated on the basis of the

vehicle position and speed information.

5 DRIVER’S BEHAVIOUR AND

RISK INDEXES

The study of the risk of the single vehicle

approaching the intersection is strictly linked to the

study of the dilemma zone and – subsequently – to

the distances required to clear the area and stop

depending upon the course state adopted by the

vehicle. Two specific indicators have therefore been

formulated: the former is relevant to the overall

clearance of the intersection and latter refers to the

complete stop of the vehicle in correspondence to

the stop line. Literature proposes various approaches

to risk assessment (Rausand, 2011), yet those

hereafter described have been originated by our

proposal, having in mind a simple approach, at least

at this level of analysis.

On the grounds of the analyses described, ratios

have been formulated to determine – as a result of

the identification of the dilemma zone – simple risk

indexes on the basis of specific input data.

With reference to a determined time instant (at a

given spatial position D), the risk index relevant to

the stop manoeuvre (IR_stop or IR1) is defined by:

RoadSafetyatIntersectionsControlledbyTrafficLights-IVCandRiskIndexes

573

IR1= IR_stop= D_stop / D (3)

where:

 D_stop is the distance – computed from the stop

line – the vehicle needs to stop – in full safety

conditions – before or in correspondence to the stop

line (stop distance as previously defined);

 D is the distance – measured from the stop line –

where the vehicle is at the time taken into

consideration.

According to the report we have presented above, a

null or almost null risk index represents the fully

safe condition (D >>D_stop), since the vehicle can

stop without the risk of occupying the intersection,

even if partially. Values of IR_stop ≥1, on the other

hand, detect potentially hazardous conditions (D

<<D_stop) for safe stopping. Values of IR_stop

included between 0 and 1 indicate almost totally safe

or almost risky conditions, depending on whether

they are closer to zero or to one.

Likewise, a risk index has been defined as

related to complete intersection clearance

manoeuvre (IR_clearance or IR2) :

IR2 = IR_clearance = D / D_clearance (4)

where:

 D is the distance – measured from the stop line,

where the vehicle is at the instant taken into account;

 D_clearance is the clearance distance; such

distance, which is computed starting from the stop

line, ensures the vehicle the complete clearance of

the whole intersection, in full safety conditions,

during the yellow light stage (relevant to its

manoeuvre) or – in case –during the all red stage.

For the sake of greater security, the all red stage has

been considered as equal to zero (an all red stage is

present in reality, even though it is rather limited).

By this choice, we have intended to allocate the all

red as safeguard fraction for those whose behaviour

– perhaps because of slower reflexes – is not within

the average one which was computed in these

analyses; an advanced ADAS system may include

the transmission on board of the all red time,

consequently modifying the risk conditions; setting

such value to zero would allow providing a risk

indicator even to vehicles which are not equipped

with ADAS-IVC.

Values of IR_clearance close to zero identify full

safety conditions (D <<D_clearance) – i.e. where the

vehicle can fully clear the intersection by the end of

the yellow light stage relevant to its manoeuvre -

also for the risk index connected to the clearance

manoeuvre. On the other hand, values of risk

relevant to clearance which are greater than or equal

to one would identify potentially risky situations (D

>>D_clearance) for the overall clearance of the area

in full safety conditions. Values of IR_clearance

included between 0 and 1 indicate, almost fully safe

or almost risky conditions, depending upon their

being closer to zero or to one.

This section of analysis focused on the behaviour

– and relevant criticalities – of the different drivers

who approach intersections governed by traffic

lights at the moment the yellow light is triggered.

The analysis of a single vehicle is not aimed at

assessing the consequences of the potential accident;

it merely intends to evaluate how much a vehicle -

depending upon its dynamics and on the driver’s

behaviour – is exposed to the risk of accidents: it is a

kind of assessment of the exposure to the risk, rather

than an estimate of the risk itself.

A numerical calculation tool has been created for

such study so that – after the introduction of specific

data into the case in exam – the presence and

extensions of the dilemma zone could be assessed

Fig.4, as well as the value of the risk indexes (of not

completing the manoeuvres of either stop or

complete clearance of the intersection by the end of

the yellow light or – in case - all red stages) and if

there is the actual risk of accident. The tool

reproduces the motion of a single vehicle

approaching a traffic signal and provides also

graphic outputs for the variation of the risk indexes

as a function of the initial speeds which can be

assumed for the vehicle in exam.

With reference to the three t

pra

values which have

been taken into consideration in the analyses (i.e. 2.3

– 1.46 – 1.1 seconds), the presence and variation of

the dilemma area have been investigated to reach the

definition of risk indexes relevant to both the

clearance and stop manoeuvres; such indexes

highlight what the most advantageous or least

disadvantageous manoeuvres would be for the

drivers of the analyzed vehicle (see the analysis of a

specific situation in Fig. 5).

In order to detect the risk of the vehicle when

approaching the traffic signal, we can assume to

update its risk level at different positions before the

stop bar. Since the feasible deceleration rates for

vehicles fall usually in a quite limited range (a

typical range might be between 3 and 5 m/s

progressive sections along the approaching lanes can

be defined to trace its speed and compare it with the

expected value in case of stopping from that

distance.

The first of these checking points (named section

“A”) is defined assuming a deceleration rate of 3

m/s

and is 64m before the stopping bar, for a

ICINCO2012-9thInternationalConferenceonInformaticsinControl,AutomationandRobotics

574

vehicle moving with a speed of 50km /h.

ariation of position and extension of the dilemma zone

0,00

20,00

40,00

60,00

80,00

100,00

120,00

140,00

160,00

180,00

0 102030405060708090100

Speed [km/h]

Distances from the stop line [m]

Stopping distance

Clearance distance

ariation of position and extension of the dilemma zone

0,00

20,00

40,00

60,00

80,00

100,00

120,00

140,00

0 102030405060708090100

Speed [km/h]

Distances from the stop line [m]

Stopping distance

Clearance distance

Figure 4: Numerical methodology for the analysis of the

dilemma area and relevant diagrams.

Fig.5. shows – as related to this specific section

(section “A”) - how IR_stop (IR1) grows linearly

with the increase of the speed even though the other

conditions remain the same, whilst IR_clearance

(IR2) decreases in an almost exponential trend.

It is worth noticing that the portions of curves

above the threshold of IR = 1 (which is displayed in

red in the diagrams) identify risky situations. If – at

a given speed – at least one of the two IR were

below such threshold, the manoeuvre to be

recommended would be the one which corresponds

to it (by a communication on board the vehicle).

If both the IR’s were below such thresholds, then

either manoeuvre would not be severely risky and -

in any case – it would be appropriate to provide

indications on board to apply the safest one, i.e. the

one which is farther from the threshold. In those

case where - at a given speed – both indexes were

exceeding the IR =1 threshold, then – even though

the safety conditions are lacking- it would be

appropriate to provide communications on board to

apply only the manoeuvre that – between the two

ones – would involve lower risk (i.e. the one which

is closer to the threshold) or to communicate the risk

condition to the other vehicles which are

approaching the intersection.

Although here only the risk indexes related to

section “A” have been reported, an ADAS can easily

update this simple estimation, while the vehicle is

approaching the intersection and recognize critical

cases by following the evolution of these risk

indexes over space/time.

Trends of Risk Indexes in section "A"

0,00

1,00

2,00

3,00

4,00

5,00

6,00

7,00

0 1020304050607080

Speed [km/h]

Risk Indexes [-]

I.R.1

I.R.2

threshold

Hazardous Manoeuvre

Safety Manoeuvre

Figure 5: Trend of the risk indexes relevant to the

clearance and stop manoeuvres related to a given section

and to a specific t

p,r,a

as a function of the speed.

Two typologies of use of the information on the

risk indexes above can then be identified. In

particular, if the data can be detected in real time,

the driver could be informed onboard of a

potentially hazardous situation (which might occur if

he/she kept such driving behaviour) using – if

required – purposely-allocated algorithms to predict

the trend of the vehicle on the basis of the data

detected from the monitoring; another use, which is

strictly linked to the utilization of the diagrams

obtained, would be detecting – in case the vehicle

were already within the dilemma zone – the lowest

risk manoeuvre and sending a message on board to

inform the driver.

Comparison of trends of Risk Indexes in section "A"

0,00

1,00

2,00

3,00

4,00

5,00

6,00

7,00

0 1020304050607080

Speed [km/h]

Risk Indexes [-]

IR1_tp,r,a=2,3s

IR2_tp,r,a=2,3s

IR1_tp,r,a=1,46s

IR2_tpra=1,46s

IR1_tpra=1,1s

IR2_tpra=1,1s

threshold

↓

IR2

↓

IR1

Figure 6: Comparison diagram of the risk indexes as a

function of the speed, relevant to a given section,

according to the three t

p,r,a

which have been taken into

consideration.

Furthermore, the effects of using communication

technologies between the infrastructure and the

RoadSafetyatIntersectionsControlledbyTrafficLights-IVCandRiskIndexes

575

vehicle have been assessed reducing - in the

analysed situations – the driver’s t

p,r,a

from 2.3 s to

1.46 s and 1.1 s, leaving the other conditions

unchanged.

Fig.6 reports an example of a diagram which

summarizes the curves of the risk indexes assessed

for the three different t

p,r,a

. It is worth noticing how –

as a result of the reduced t

p,r,a

- the risk indexes

relevant to both the clearance and stop manoeuvres

result to be reduced as well.

Figure 7 and 8 report the trends and lengths of

the dilemma zone, related to the distance needed

respectively to free the crossroad area or to stop.

Trend and length of the dilemma zone

0,00

20,00

40,00

60,00

80,00

100,00

120,00

140,00

160,00

180,00

15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

Speed [km/h]

Distance from the stop line [m]

Length of the dilemma zone

Clearance distance

Figure 7: Trend and length of the dilemma zone, related to

the distance needed to free the crossroad area.

Trend and length of the dilemma zone

0,00

50,00

100,00

150,00

200,00

250,00

15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

Speed [km/h]

Distance from the stop line [m]

Length of the dilemma zone

Stopping distance

Figure 8: Trend and length of the dilemma zone, related to

the distance needed to stop.

6 CONCLUSIONS

In this work an ADAS for traffic signal approaching

has been analysed and two main roles have been

considered:

 provide a risk estimation for alternative

manoeuvres (stopping or clearance) and then

communicate the driver the less hazardous

manoeuvre on the basis of known, measured or

estimated parameters;

 reduce the risk level, by reducing the driver

perception and reaction time, since IVC increase the

level of attention of the driver.

The experiments run in simulation by means of a

spreadsheet have led to acknowledge – as a result of

the reduction in the t

p,r,a

, - a corresponding reduction

in the estimated risk of accidents. The positive

effects of the infrastructure–vehicle communication

have been ascertained in terms of reduced exposure

to the risk by a single vehicle (analyses of the trend

of the single vehicle approaching the intersection

regulated by traffic lights). More specifically, as

related to the behaviour of a driver at the moment

the yellow light is triggered for his/her traffic

stream, the application of I2V systems

(corresponding to a reduction in the t

p,r,a

), the

following has been observed:

 reduced extension of the dilemma zone;

 disappearance of the dilemma zone and growth

of the zone of choice : in some cases, as a result of

the increased level of attention in order to attain

standard values, i.e. tp,r,a equal to 1,46 s and –in a

large number of cases relevant to the forwarding of

alert messages – tp,r,a equal to 1.1 s ;

 the decreasing of the risk indexes relevant to the

stop (IR1 o IR_stop) and clearance (IR2 o

IR_clearance) manoeuvres, mainly in

correspondence to the speed values corresponding to

IR values which were far greater than the safety

threshold (IR=1): in correspondence to very low

speeds for IR2 and high speeds for IR1;

 the advanced knowledge of IR1 and IR2, with

the subsequent opportunity to warn the drivers on

board (possibly before they enter the dilemma zone)

about the lowest risk manoeuvre to be undertaken:

such potential is useful mainly in those cases where

both IR’s result to be above the safety threshold

IR=1;

 the opportunity to reduce instantaneously, and

therefore in real time, the risk or – better – the

exposure to the risk - of not completing in full safety

conditions the manoeuvre which is intended to be

undertaken by the end of the yellow light stage.

In short, the results of the analyses show that the use

of the I2V e V2I communication systems in the

intersections regulated by traffic lights – assumed in

the processing as directly related to a reduction of

the tp,r,a - has beneficial effects on road safety as

related to the reduction of risks of accidents.

Furthermore, the analyses performed allow

supporting also the combined use of sensors, to

enable the most viable continuous monitoring and

assess the dilemma zone and the potential risk of

ICINCO2012-9thInternationalConferenceonInformaticsinControl,AutomationandRobotics

576

accident instantaneously and in real time.

It is also worth specifying that the analyses did

not consider any actual data on the use of the I2V

technologies – since they are not available to date –

or any active intervention on the vehicle in case of

need.

The subject is in evolution and many questions

remain open. For example an investigation of

vehicle behaviour, when it is not isolated in

approaching the traffic signal, need more tests,

possibly also with a traffic micro-simulation tool. It

can be assumed that the actual potential of the

systems which have been taken into consideration

could be assessed once said technologies are widely

spread on the market.

REFERENCES

Archer J., 2005. Indicators for traffic safety assessment

and prediction and their application in micro –

simulation modelling: A study of urban and suburban

intersections. Doctoral Thesis; Royal Institute of

Technology, Stockholm, Sweden.

Benz T., Gaitanidou E., Spyropoulou I., Yannis G., and

Tapani A.,2006. Modelling road traffic safety - The

In-Safety approach. In Proceedings of the 13th World

Congress and Exhibition on Intelligent Transport

Systems and Services, ERTICO, London.

COOPERS project CO-OPerativeSystEms for Intelligent

Road Safety - (accessed March 07, 2012)

http://www.coopers-ip.eu/index.php?id=project.

CoVeL - Cooperative Vehicle Localization for Efficient

Urban Mobility (accessed March 07, 2012)

http://www.covel-project.eu/.

CVIS Cooperative Vehicle-Infrastructure Systems

(accessed March 07, 2012)

http://www.cvisproject.org/.

Dalla Chiara B., Deflorio F. P., Diwan S., Assessing the

effects of inter-vehicle communication systems on

road safety, IET Intelligent Transport Systems, June

2009, IET Intell. Transp. Syst., 2009, Vol. 3, Iss. 2, pp.

225–235, doi: 10.1049/iet-its:20080059.

Day C. M., Premachandra H., Brennan T. M., Sturdevant

J. R., Bullock D. M., Operational Evaluation of

Wireless Magnetometer Vehicle Detectors at

Signalized Intersection, Transportation Research

Record 2192.

FHWA (2006). Traffic Detector Handbook: Third Edition

Chapter 4. http://www.fhwa.dot.gov (accessed March

07, 2012).

FOTsis - European Field Operational Test on Safe,

Intelligent and Sustainable Road Operation (accessed

March 07, 2012) http://www.fotsis.com/.

Gettman D. & Pu L., 2006. Theoretical validation of

Surrogate Safety Assessment Methodology for

roundabouts and cross intersections. In Proceedings of

the 13th World Congress and Exhibition on Intelligent

Transport Systems and Services, London.

Hurwitz, D. S., 2009. Application of Driver Behavior and

Comprehension to Dilemma Zone Definition and

Evaluation. Open Access Dissertations. Paper 112,

University of Massachusetts.

INTERSAFE-2 – Cooperative Intersection Safety

(accessed March 07, 2012) http://www.intersafe-

2.eu/public/

Johansson G., Rumar K., 1971. Driver's brake reaction

times. Human Factors, 13(1), pp. 23-27

Ki-Joon K., Jaehoon S., 2009. Development of

Intersection Traffic Accident Risk Assessment Model

- Application of Micro-simulation Model with SSAM

to Sungnam City. In IRTAD CONFERENCE, Seoul,

Korea.

Klunder G., Abdoelbasier A., Immers B., 2006.

Development of a micro-simulation model to predict

road traffic safety on intersections with surrogate

safety measures. In Proceedings of the 13th World

Congress and Exhibition on Intelligent Transport

Systems and Services, London.

Liu C., Herman R. and Gazis D.C., 1996. A review of the

yellow interval dilemma, Transportation Research Part

A: Policy and Practice, 1996, vol. 30, issue 5, pages

333-348.

MacNeill, P. Miller, R., 2003. A new technology for a

cruise control system A new technology for a cruise

control system. IEEE Vehicle Technology Conference

Maile M., Delgrossi L., 2009. Cooperative intersection

collision avoidance system for violations (CICAS-V)

for avoidance of violation-based intersection crashes.

Enhanced Safety of Vehicles.

Monteil J., Billot R., El Faouzi N. E., 2011. Towards

cooperative traffic management: methodological

issues and perspectives. In Australasian Transport

Research Forum 2011 Proceedings, 28 - 30 September

2011, Adelaide, Australia.

Morsink P. L. J., Wismans L. J. J., Dijkstra A., 2008.

Micro-simulation for road safety impact assessment of

advanced driver assistance systems. InEuropean ITS

Congress, Geneva.

Rausand M., Risk Assessment, 2011. Theory, Methods,

and Applications, Norwegian Univ. of Science and

Technology, ISBN: 978-0-470-63764-7, September

2011.

SAFESPOT Integrated Project (accessed March 07, 2012)

http://www.safespot-eu.org/.

Tapani A., 2009. Traffic simulation modelling of driver

assistance systems. In 16th World congress on ITS.

Torday A., Baumann D., Dumont A. G., 2003. Road

safety assessment using micro simulation based

indicators. Proc. of the ITS World Congress, Madrid.

U.S. Department of Transportation Federal Highway

Administration,2009.Surrogate Safety Assessment

Model (SSAM).

U.S. Department of Transportation, Cooperative

Intersection Collision Avoidance System Limited to

Stop Sign and Traffic Signal Violations (CICAS-V),

Task 8 Final Report, Prototype Build and Testing,

(Appendix F), June 24, 2008, August 29, 2008.

RoadSafetyatIntersectionsControlledbyTrafficLights-IVCandRiskIndexes

577