Roadway Lighting and Visibility Anchorage.ppt

12/22/2008

1

Roadway Lighting and Visibility

D R ld B GibbDr. Ronald B. GibbonsGroup Leader, Lighting and Infrastructure

Technology

Motor Vehicle Crashes – Magnitude of the Problem

Cause and Number of Deaths

Other Adults

R

A

N

K

Infants Under 1

Toddlers

1-3

Young Children

4-7

Children

8-15

Youth 16-20

Young Adults 21-24 25-34 35-44 45-64

Elderly

65+

All Ages

Years of

Life Lost2

1

Perinatal Period 13,734

Congenital Anomalies

496

MV Traffic Crashes

533

MV Traffic Crashes

1,546

MV Traffic Crashes

5,979

MV Traffic Crashes

4,136

MV Traffic Crashes

6,759

Malignant Neoplasms

16,569

Malignant Neoplasms

139,785

Heart Disease 582,730


Malignant Neoplasms

23%(8,614,131)

2


5,513

MV Traffic Crashes

421

Malignant Neoplasms

400

Malignant Neoplasms

829

Homicide

2,414

Homicide

2,738

Homicide

5,204



Malignant Neoplasms

390,214

Malignant Neoplasms

553,768

Heart Disease

22%(8,110,571)

Heart Accidental Exposure to Suicide Suicide Suicide Suicide MV Traffic Stroke Stroke Stroke MV Traffic

Top 10 Leading Causes of Death in the United States for 2001, by Age Group1

National Center for Statistics and Analysis

3

Heart Disease

479

Drowning 393

pSmoke/Fire

178

447

1,879

1,924

5,070 Crashes

6,891

15,518

144,486

163,538 Crashes

5%(1,700,952)

4 Homicide

332

Homicide

362


168

Homicide

391

Malignant Neoplasms

814

Accidental Poisoning

771

Malignant Neoplasms

3,994

Suicide

6,635

Diabetes

14,913

Chronic Lwr. Resp. Dis.

106,904


123,013

Stroke

5%(1,687,683)

5

Septicemia

312

Malignant Neoplasms

321

Accidental Drowning

164


324


566

Malignant Neoplasms

768

Heart Disease

3,160

HIV

5,867


14,490

Influenza/ Pneumonia

55,518

Diabetes

71,372


4%(1,444,745)

6


299

Heart Disease

200

Homicide

133

Accidental Drowning

293

Heart Disease

398

Heart Disease

543


2,507


5,036

Chronic Liver Disease 13,009

Diabetes

53,707


62,034

Suicide

3%(1,079,822)

7

MV Traffic Crashes

139

Exposure to Smoke/Fire

170

Heart Disease

82

Heart Disease

273

Accidental Drowning

326

Accidental Drowning

211

HIV

2,101

Homicide

4,268

Suicide

9,259

Alzheimer’s

53,245

Alzheimer’s

53,852

Perinatal Period

3%(1,070,154)

8

Nephritis/ Nephrosis

133

Septicemia

96

MV NonTraffic Crashes

51

Exposure to Smoke/Fire

140


244


206

Stroke

601

Chronic Liver Disease

3,336

MV Traffic Crashes

8,750


33,121

MV Traffic Crashes 42,443

Diabetes

3%(1,014,201)

9

Stroke

108


92

Benign Neoplasms

46

MV NonTraffic Crashes

125

Accidental Falls 114

HIV

167

Diabetes

595

Stroke

2,491

HIV

5,437

Septicemia

25,418


39,480

Homicide

3%(924,263)

10

Meningitis

78

Perinatal Period

63

Septicemia

33

Chr. Lwr. Resp. Dis.

102

Acc. Dischg. Of Firearms

114

Accidental Falls 134


458

Diabetes

1,958


5,106

Hypertension Renal Dis.

16,397

Septicemia

32,238

Chronic Liver Disease

2%(623,998)

ALL3 27,568 4,288 2,703 6,672 15,851 14,940 41,683 91,674 412,204 1,798,420 2,416,425

All Causes 100%(36,866,317)

1When ranked by specific ages, motor vehicle crashes are the leading cause of death for age 2 and every age 4 through 33.. 2Number of years calculated based on remaining life expectancy at time of death; percents calculated as a proportion of total years of life lost due to all causes of death. 3Not a total of top 10 causes of death. Source: National Center for Health Statistics (NCHS) CDC, Mortality Data 2001 Note: The cause of death classification is based on the National Center for Statistics and Analysis (NCSA) Revised 68 Cause of Death Listing. This listing differs from the one used by the NCHS for its reports on leading causes of death by separating out unintentional injuries into separate causes of death, i.e., motor vehicle traffic crashes, accidental falls, motor vehicle nontraffic crashes, etc. Accordingly, the rank of some causes of death will differ from those reported by the NCHS. This difference will mostly be observed for minor causes of death in smaller age groupings.

Motor Vehicle Crashes – Magnitude of the Problem

Distribution of global injury mortality by cause:World Report on Road Traffic Injury Preventionp j y

World Health Organization, Geneva, 2004

“In low-income and middle-income countries, the phenomenon of pedestrians and vehicles not being properly visible is frequently a serious problem. In these places, there are fewer roads with adequate ill i ti d t b lit t ll ”

Suicide, 16.9%

Violence, 10.8%

Drowning, 7.3%

Road traffic injuries, 22.8%

Source: WHO Global Burden of Disease project, 2002, Version 1.

illumination and some may not be lit at all.”From Chapter 3: Risk Factors (pg 86)

War, 3.4%

Other intentional injuries, 0.2%

Other unintentional injuries, 18.1%Poisoning, 6.7%

Falls, 7.5%

Fires, 6.2%

Driver IssuesAccident Statistics show that more than 50% of fatal accidents occur during the night hours while g gonly 25% of the vehicle miles are driven.• Night driving has been described as a situation for

which humans have not evolved, leaving our visual system inadequate and inefficient for certain tasks (Rumar, 1990).

Determining the nat re of the dri ing task isDetermining the nature of the driving task is critical

12/22/2008

2

100 Car “Naturalistic” Approach• Data collection in a “naturalistic” setting to obtain

crash/pre-crash/near-crash/conflict data as well as distributions of driver performance100 d i i th i ( l d) hi l ith• 100 drivers in their own (or leased) vehicles with specialized instrumentation, on public roads, as close to unobserved as possible.

• Subjects use instrumented vehicles for an extended period (up to 13 months) without an experimenter present.

• Subjects are not coached or instructed to perform any j p yspecific actions other than drive as they normally do.

• Instrumentation is unobtrusive and inconspicuous to other drivers, but not invisible.

Naturalistic Data Collection ApproachHighly capable instrumentation (well beyond EDRs)

• Five channels of digital, compressed video• Four radar sensors front, rear (for all 100

vehicles), and side (for 20 vehicles)• Machine vision-based lane tracker• Many other sensors: GPS, glare, RF,

acceleration, yaw rate, controls, etc.• Cell phone, wireless internet, or hardwire

downloadTi i hi l k b i h• Tie into vehicle network to obtain other sensor information

• Rugged, crash tested, all solid state• Crash detection, Fault detection• Remote Access

The Naturalistic “100 Car” Driving Study:

Database Statistics42,300 hours of driving data collected82 Crashes and collisions• Defined as any contact between the subject vehicle and another

vehicle, fixed object, pedestrian pedacyclist, animal.761 Near crashes • Defined as a conflict situation requiring a rapid, severe evasive

t id hmaneuver to avoid a crash.8295 Critical incidents• Conflict requiring an evasive maneuver, but of less

magnitude than a near crash.

Driver Behavior - Spinny

12/22/2008

3

Driver Behavior – Silly Driver Behavior – Dopey

Driver Behavior - Sleepy Data Analysis

12/22/2008

4

ResultsThe primary cause of all of incidents is secondary task distraction• Doing other things in the vehicle

Any visual task must overcome the distraction of the driver in the vehicle

Lighting and Driver SafetyWhat we know…

Box [1972] showed that the night/day accident ratio was 66% higher on unlighted freeways than on lit ones.• 0.5 lux appeared to be the illuminance level which provided

the lowest accident rateOsner [1973] and Nishimori[1973] both showed a 56% reduction in accidents when lighting was added to a roadway.CIE Pub. No 93 “Road Lighting as an AccidentCIE Pub. N 93 Road Lighting as an Accident Countermeasure” rigorously analyzed 62 lighting and accident studies from 15 countries.• “(S)tatistically significant results show reductions (in

nighttime accidents) of between 13 and 75 percent.”

Motor Vehicle Crashes – Implications of Darkness

Vehicle occupant deaths, FARS, 1987-2003

Motor Vehicle Crashes – Implications of Darkness

Pedestrian deaths, FARS, 1987-2003

12/22/2008

5

Evaluating Impact of Light Pedestrian Fatalities – Fall PM Return to Standard Time

250

300Light Dark

100

150

200

PedestrianRun-off-road

Cra

shes

90-90

50

Weeks Before and After Return to Standard Time

Visual ActivityAs drivers, the visual task is a very complex activity• Detection of road hazards• Lane keeping• Wayfinding• Monitoring of the instrument panel• Observing other drivers• Pedestrian Detection• Sightseeing?

Mi di th th t i th hi l• Minding the other occupants in the vehicleWe distribute our visual resources between all of these activities• We allot attention to the task which seems most demanding• Not necessarily the most important

The Eye

12/22/2008

6

The RetinaCaptures photons and send nerve impulses to the brainTwo important regions of the retina are:• Fovea

• The central and most sensitive part of the visual field

• Highest Acuity• Almost entirely ConesAlmost entirely Cones

• Periphery• Low Acuity• High range of sensitivity• Almost entirely Rods

Visual ProcessThe visual process is partially automatic and partially consciousp y• Lane keeping is automatic

• Visual vection process• Object Detection and Wayfinding is conscious

Object Detection Process• Visual Search

• We have a standard search pattern as we drive• Looking for objects• Looking at signage• Following the road path

• Detection• Through the visual search, we find an object of interest

• This detection can be peripheral• Spotted to the side as a result of motion or through high

conspicuity• This detection can be foveal

• Found through the visual search patterng p• Recognition

• We attend to the objects of interest• This is a foveal task

• Reaction• We decide what the appropriate course of action

• Braking, steering etc.

Object Visibility

12/22/2008

7

Visibility with Glare Visibility with Extreme Glare

Visibility of ObjectsWe see objects based on their contrast to the background• This can either be color contrast or luminance contrast• In roadway lighting design, color is not considered

The IES lighting design requirements are consensus standards based on experience• Lighting design can be performed based on illuminance,

luminance or STVTh STV i i ht d f i f l l t d• The STV is a weighted average of a series of calculated Visibility Levels for a defined target

• Current target is a flat 7" square with 50% reflectance which represents the smallest object which will collide with a vehicle

• VL is calculated as: VL CC

L LL

act

th

Target Background

th

= =−Δ

IES RP-8Illuminance Design Criteria

Road and Pedestrian Conflict Pavement Classification

Area (Minimum Maintained Average Values) Uniformity Veiling Ratio Luminance Ratio Luminance

Road Pedestrian R1 R2 & R3 R4 Ratio Conflict Area Lux/fc Lux/fc Lux/fc Eave/Emin Lvrnax/Lavg

Freeway Class A 6.0/0.6 9.0/0.9 8.0/0.8 3.0 0.3 Freeway Class B 4.0/0.4 6.0/0.6 5.0/0.5 3.0 0.3

High 10.0/1.0 14.0/1.4 13.0/1.3 3.0 0.3 Expressway Medium 8.0/0.8 12.0/1.2 10.0/1.0 3.0 0.3

Low 6.0/0.6 9.0/0.9 8.0/0.8 3.0 0.3 High 12.0/1.2 17.0/1.7 15.0/1.5 3.0 0.3

Major Medium 9 0/0 9 13 0/1 3 11 0/1 1 3 0 0 3Major Medium 9.0/0.9 13.0/1.3 11.0/1.1 3.0 0.3 Low 6.0/0.6 9.0/0.9 8.0/0.8 3.0 0.3 8.0/0.8 12.0/1.2 10.0/1.0 4.0 0.4

Collector High Medium 6.0/0.6 9.0/0.9 8.0/0.8 4.0 0.4 Low 4.0/0.4 6.0/0.6 5.0/0.5 4.0 0.4 6.0/0.6 9.0/0.9 8.0/0.8 6.0 0.4

Local High Medium 5.0/0.5 7.0/0.7 6.0/0.6 6.0 0.4 Low 3.0/0.3 4.0/0.4 4.0/0.4 6.0 0.4

12/22/2008

8

IES RP-8Luminance Design Criteria

Road and Pedestrian Conflict Average Uniformity Uniformity Veiling Area Luminance Ratio Ratio Luminance

R ti RatioRoad Pedestrian L

Conflict Lave/Lmin Lmax/Lmin Lvmax/Lavg

Area ( cd/m2) (Maximum Allowed)

(Maximum Allowed)

(Maximum Allowed)

Freeway Class A 0.6 3.5 0.5 6.0

Freeway Class B 0.4 3.5 6.0 0.3

High 1.0 3.0 5.0 0.3

Expressway Medium 0.8 3.0 5.0 0.3

Low 0.6 3.5 6.0 0.3

Hi h 1 2 3 0 5 0 0 3 High 1.2 3.0 5.0 0.3

Major Medium 0.9 3.0 5.0 0.3 Low 0.6 3.5 6.0 0.3 High 0.8 3.0 5.0 0.4 Collector Medium 0.6 3.5 6.0 0.4 Low 0.4 4.0 8.0 0.4 High 0.6 6.0 10.0 0.4 Local Medium 0.5 6.0 10.0 0.4 Low 0.3 6.0 10.0 0.4

IES RP-8STV Design Criteria

Road and Pedestrian Conflict STV Luminance Criteria Area Criteria

Road Pedestrian Weighting UniformityLavg Lavg Road Pedestrian Weighting Uniformity Conflict Area Average

Lavg Cd/m2

Lavg Cd/m2 Ratio

VL Median Median Lmax/Lmin <7.3m >7.3m (Maximum

Allowed)

Freeway "A" 3.2 0.5 0.4 6.0

Freeway"B" 2.6 0.4 0.3 6.0

Expressway 3.8 0.5 0.4 6.0

High 4.9 1.0 0.8 6.0 Major Medium 4 0 0 8 0 7 6 0Major Medium 4.0 0.8 0.7 6.0

Low 3.2 0.6 0.6 6.0 High 3.8 0.6 0.5 6.0 Collector Medium 3.2 0.5 0.4 6.0

Low 2.7 0.4 0.4 6.0 High 2.7 0.5 0.4 10.0 Local Medium 2.2 0.4 0.3 10.0

Low 1.6 0.3 0.3 10.0

IES RP-8 – The Next Revision

Separation of street and roadway lighting• Different Visual TasksDifferent Visual Tasks

Adaptive Possibilities• Allowing for changing of road class based on

pedestrian and vehicle traffic changes

Draft is currently under review and should be completed within 1 yearp y

Roadway Lighting and Driver Safety – What we don’t know…The impact of lower lighting levels on driver safetyy• Are we over-lighting?

The impact of mesopic lighting levels on driver safety• Is White truly better?

12/22/2008

9

Threshold VL LevelsIESNA Specifies a Level of 1.6 to 4.9 for STVActual threshold values are much higherHi h li h i l l lik l i dHigher lighting levels are likely not required• Contrast control is required• Perfect uniformity may not be the best lighting condition

Visibility Level At Threshold: White Clothed Objects

50

60

70

80

0

10

20

30

40

50

Non

e

Hyb

rid

3UV

A

5UV

A

Non

e

Hyb

rid

3UV

A

5UV

A

HLB HID HHB HOH HLB-LP

IR-T

VES

VL

CyclistPerpendicularParallelStatic

The White IssueWhite Light might provide equivalent visual task performance at a lower illuminance level than pnon-White Sources

Equivalent Performance?• The task is performed with the same speed and

accuracyy• In outdoor lighting this equates to driver safety,

pedestrian safety, way finding performance, comfort

Why?So why is there a possible white light benefit?• The physiology of the eye lends itself to light sourcesThe physiology of the eye lends itself to light sources

which have radiation in the entire visual spectrum• Remember

• Our eye was developed to interact with the sun –The ultimate full spectrum light source.

• The effect is a result of the difference in the spectral sensitivity of the various photoreceptors in the retinasensitivity of the various photoreceptors in the retina

PhotoreceptorsRods• Sensitive to low levels of radiation• The Spectral Sensitivity of the rod is maximum at 507nm• The rods define the pupil size in all viewing types

(Photopic, Mesopic and Scotopic)

Cones• Cone are sensitive at high levels of radiation• There are 3 cone types defined by their spectral sensitivities• There are 3 cone types defined by their spectral sensitivities

• Long• Medium• Short

12/22/2008

10

Photoreceptor Sensitivity Cone Sensitivity

Relative Eye Sensitivities Overall Spectral Sensitivity of the Eye

Photopic Vision• Cone Vision - Using V(λ)Cone Vision Using V(λ)• > 3.5 cd/m² adaptation Luminance

Scotopic Vision• Rod Vision - Using V´(λ)• < 0.035 cd/m²

Mesopic VisionMesopic Vision• Mixture of Rod and Cone Vision• Transition state between photopic and scotopic

12/22/2008

11

Scotopic LumenA lumen is defined as:

• P(λ) is the spectral power distribution (SPD) of the light source

• V(λ) is the spectral sensitivity of the eye• k = 683 lumens/Watt• Φ = Luminous Flux in Lumens

A scoptic lumen represents the luminous flux evident to scopt c u e ep ese ts t e u ous u ev de t tothe rods.• It is calculated using V’(λ) instead of V(λ).

The S/P Ratio is the ratio of Scotopic Lumens to Photopic Lumens

Mesopic LumensThe Mesopic lumen is calculated based on using a Mesopic sensitivity curve rather than a Photopicp y p• Difficulty:

• The Mesopic Sensitivity changes with the adaptation luminance

• There are 2 models• LRC

l b i d b h• Move – Currently being accepted by the CIE• Mesopic Lumens are not uniform across the retina

What is the Actual Benefit?Better Visual AcuityHigher Equivalent LuminanceHigher Equivalent Luminance

Better Visual AcuitySome researchers suggest that light sources with a higher S/P Ratio are more beneficial as they force the pupil to be smaller, allowing the observer to have better visual acuity and visual performanceIssue for outdoor lighting:• This research was only performed for indoor lighting levels and

may not apply to outdoor applications

12/22/2008

12

Higher Perceived LuminanceLight sources with a blue content create a higher response in the cones and therefore a higher p gperceived luminance• Higher Mesopic Luminance• Higher S/P ratio

Issue:• It is important to remember that this change is only p g y

evident in the periphery of the eye and not in the fovea• Rods must be present to create the effect

For outdoor lighting, this is the most critical aspect of white light.

How big is the benefit?There is conflicting research• LRC has shown that an object which appears at 20º

peripherally is perceived at a lower level in white light thanperipherally is perceived at a lower level in white light than in amber

• Lewis has shown similar effects for peripheral targetsIt has been proposed that Luminance Equivalence Multipliers be applied to lighting levels to account for the increased benefit of white light.• These values are based on the Mesopic lumen rating of the

light sourcelight source• The applicability value of the scaling factors is a topic of

debate Important:• The use of these scaling factors has not been approved by

the IESNA for Lighting Applications

Proposed LEMs for White Light

Source 1 fc Illuminance

0.1 fc Illuminance

Metal Halide 1 1Incandescent 1.5 2.9Mercury Vapor 2 4 4 4Mercury Vapor 2.4 4.4High Pressure Sodium

3.9 7.8

Low Pressure Sodium

4.8 14.6

Conflicting ResearchVTTI Crosswalk Lighting Results• Metal Halide under performs HPS in a object detectionMetal Halide under performs HPS in a object detection

task in a roadway environment• For the same vertical illuminance on the object,

detection distances were longer under HPS than MH for Black Clothed Objects and Equivalent for Denim Clothed Objects

12/22/2008

13

Crosswalk Lighting – Pedestrian Detection

Impact of Pedestrian Clothing On Pedestrian Detection

800

1000

1200

1400

1600

1800

2000

Dis

tanc

e (f

t)

Black No GlareDenim No GlareSurrogate No GlareWhite No Glare

0

200

400

600

6 10 20 30

Lamp Level

Crosswalk Lighting – Black Clothed

Impact Of Lamp Type Pedestrian Detection For Black Clothed Pedestrians

300

400

500

600

700

800

900

Dis

tanc

e (f

t)

HPS No GlareMH

0

100

200

300

6 10 20 30

Lamp Level

D

Crosswalk Lighting – Denim Clothes

Lamp Type and Denim Clothed Pedestrians

400

600

800

1000

1200

Dis

tanc

e (f

t)

HPS No GlareMH No Glare

0

200

400

6 10 20 30

Lamp Level

Why the difference in the Research Results?

Experimental Methods differences• Many experiments uses fixed geometry to determine the effects

Th lt th i t t d t l th ti• The results are then interpreted to apply across the entire visual field

• This is invalid as the visual field is not uniform across the retina

• Crosswalk Lighting Investigation• Use non fixed geometry – Free Driving• Most objects appeared foveally

• No Mesopic White Light effect is evident• No Mesopic White Light effect is evident• Older Drivers – 65+ years old

• The lens yellows with age and may impact performance for a non-yellow source

Neither of these methods truly represent what is happening in a vehicle

12/22/2008

14

Other White Light IssuesHigh Color Temperatures with a high blue content can cause greater sky scatter and sky glowg y y g• Rayleigh scattering is spectrally sensitive

• Collision of light with atmospheric molecules• Which is why the sky is blue• Low angle blue content light scatters more

• We have evolved under the moonlight• We believe the color temperature of the moonlight

is about 4200K

Research GoalIn order to fully assess the impact of lighting on the driver, a visual model must be developed , pwhich is accounts for all of these issues• Peripheral vs. foveal issues• Color impacts• Spectral effects• Object movement• Visual Search parameters

VTTI ResearchVTTI is currently working on a project to establish this visual model• Using eye tracking to determine visual search• Using eye tracking to determine visual search

characteristics• A wide variety of visual tasks, light sources and driving

conditions• Free driving will be the basis of the experimentation

• No fixed geometry• Naturalistic data collection

Our experiment here provides validation to some of the work performed in the experiment• Color

• 4 different target colors – Gray, Green, Blue, Red• Light Sources – 4200k LED and induction

The VehicleWe collected:• IlluminanceIlluminance• Luminance• Color• Observer Input• GPS

12/22/2008

15

Color CameraCollecting RBG of road viewAssists in determining conflicts

Luminance CameraCollecting Photometrically yAccurate Road Characteristics

Illuminance5 illuminance meters• 4 measuring horizontal illuminance on the roof of the4 measuring horizontal illuminance on the roof of the

vehicle• Wheel path and centerline of the vehicle• 2 along centerline to measure gradient

• 1 measuring vertical illuminance at the windshield• Estimates of glare impact

Where the Results GoWe will be analyzing the detection distances and the luminance in each area.• A short report will be prepared

The data will add to our entire database of data.

Documents

Roadway Lighting and Visibility Anchorage.ppt