Improving the Credibility of, and Compliance with, Speed Limits: a Real-World Approach

1
Improving the Credibility of, and Compliance
with, Speed Limits a Real-World Approach

• N. Hautière1, P. Charbonnier2, E. Dumont1, S.
  Glaser1, E. Violette3

1 LCPC, Paris 2 LRPC de Strasbourg, ERA 27,
Strasbourg 3 CETE Normandie-Centre, ERA 34, Rouen
2
Presentation Outline

• Introduction
• Rough points
• Our position
• Speed limits computation state of the art
• How to improve drivers compliance with speed
  limits?
• Process
• Examples
• On-board solution ARCOS
• Roadside solution SARI
• Cooperative solution DIVAS
• Perspectives

3
Introduction

• Fact
• There is a strong link between accidentology and
  speed.
• Problem
• Permanent speed limits (signs) do not help road
  users to adapt their speed in case of transient
  difficulties - Meteorology rain, fog, wet road,
  ice on the road
• - Traffic, road works, lack of maintenance
• Solution
• Adaptive and customized speed limits.

110 light rain 90 strong rain
4
Introduction Rough Spots(at least in France)

• The duality of speed limits
• Speed limits have different functions, e.g.
• to ensure homogeneous behaviours,
• to ensure coherence with road related risks.
• To comply with speed limits, road users must be
  aware that speed limits are related to the risk
  and not only to speed enforcement.
• Problem do we communicate on the speed or on
  the risk?
• Compliance
• Prior to the introduction of automatic speed
  enforcement, speed limits were designed knowing
  that they would not be respected.
• Today, posted speed limits are no longer suitable
  because they are complied with.
• Liability
• Legal issues are problematic for adaptive speed
  limits.

5
Introduction Our Position

• We focus on the scientific aspect of the problem.
• We seek to compute credible safe speed
  limits.() i.e. risk-related
• We consider isolated vehicles, only interacting
  with the infrastructure.

6
Speed Limits Computation State of the Art

• Empirical approach
• Actual speeds are measured in nominal conditions
  ? all the parameters are integrated
• Problem only appliesin nominal conditions ?
  not adaptive ? not credible

• Computational approach
• Based on physical models for specific
  situations? many parametersare omitted driver,
  car, visibility...
• Problem not customized ? needs to be
  pessimistic? not credible

7
How to Improve Drivers Compliance with Speed
Limits?

• Hypothesis
• Credible speed limits are better complied with.
• Question
• How to make speed limits credible?
• Answer
• By making them adaptive and customized.

8
Process

• Combine empirical and computational approaches
• Usual approach
• Our approach
• with SL Speed Limit
• MSL Mandatory Speed Limit
• NSL Nominal Speed Limit ? empirical model
• f(pi) speed decrement needed to maintain
  nominal risk level
• pi transient risk factor ? computational
  approach

9
ExamplesOn-board solution

• We make it customized(on-board more credible),
  
• ARCOS Project First use of risk functions
• Example the SAVV (speed warning in curves)
• Website http//www.arcos2004.com/

Source ARCOS Project
10
ExamplesRoadside solution

• We make it adaptive
• SARI / IRCAD (roadside addresses all drivers)
• Problem drivers are not aware of the risk in bad
  conditions (particularly with skid resistance)
• We must set a warning threshold in the speed
  distribution.
• Website http//www.sari.prd.fr/

Risk function
Source Lacroix Traffic
11
ExamplesCooperative Solution

• We make SLs both adaptive and customized.
• We generalize road-related risks by adding
  meteorological risks, and by combining risk
  factors (with ranking, rather than simply
  decrementing).
• This is one objective of the DIVAS Project.

Source ARCOS Project
Source PReVENT MapsAdas
12
DIVAS Dialog between Infrastructure and Vehicles
to Improve the Road Safety (1)

• Type of project French ANR 2006
• Promoted by PREDIT GO9
• Timeframe May 2007-May 2010
• Cost budget 4 M (ANR funding 1.3M)
• Coordinator LCPC
• Philippe Lepert
• Nicolas Hautière
• Consortium 15 partners

Competitiveness clusters
Source LARA (ENSMP/INRIA)
13
DIVAS Dialog between Infrastructure and Vehicles
to Improve the Road Safety (2)

• The DIVAS project is building a global a vehicles
  infrastructure information exchange system
• It aims at preparing its implementation, in terms
  of
• technology,
• acceptability,
• credibility.
• The project is focussed on the role of
• the infrastructure characteristics
• the role of the road operators in the deployment
  of such systems.
• It aims at providing each vehicle with an
  individualized safety indicator along a route,
• It mainly takes into account the road geometry,
  the road surface conditions and the visibility
  conditions.
• Web site http//or.lcpc.fr/divas-fr/
• Reference
• N. Hautière, P. Lepert. Infrastructure -
  Vehicles Dialogue to Improve Road Safety The
  DIVAS Approach. To appear in Transport Research
  Arena (TRA), Ljubljana, Slovenia, April 21 25,
  2008

14
DIVASMeasuring Actual Speeds (empirical
approach)

• Reference drivers with instructions, in
  nominal conditions.
• Record the speed profile along the road (and
  other information also) with an instrumented
  vehicle.
• Calibrate speed profiles using roadside speed
  measurements at different spots.
• Build a nominal speed profile.
• Infer nominal risk for a specific situation
  (e.g. brick wall)

Source LAVIA Project
15
1st Level ApplicationConsolidation of Vertical
Signalling

• Signs provide the permanent speed limits, posted
  by the road operator (or police).
• Posted speed should be coherent with nominal
  speed in order to be credible ? Discrepancies
  should be studied, baring in mind the duality of
  posted speed limits.

16
2nd Level Application Adaptation of Speed
Limits to Keep Constant Risks (computational
approach)

• Risk models are chosen with respect to the
  studied risk factor
• The DV is computed to have a constant risk (R)
  compared to the nominal risk (RN)
• Example brick wall risk model and wet road
• Nominal risk we compute the gravity of an
  accident at impact speed SN into a wall at t2s
• SN ? DAN ? EESN ? RN
• Wet road leads to a reduction of skid resistance
• S ? DA ? EES ? R gt RN
• Knowing the actual road surface conditions, we
  can compute
• DS / SS-DS ? RRN
• Assumption computing SS-DS is more credible
  than computing Sf(skid resistance), which was
  used for example in ALZIRA project.
• EES Equivalent Energy Speed (cf. LAB
  PSA/RENAULT)

17
Perspectives
Mid-term seminar
November 2008
Today
May 2010
May 2007
Q1
Q2
Q3
Q4
Q5
Q6
Q7
Q8
Q9
Q10
Q11
Q12
Time (Quarter)
DIVAS agenda

• We argue that credible risk-related speed limits
  would be better complied with.
• We proposed an approach to compute credible speed
  limits by making them adaptive and customized.
• We are testing the approach in the framework of
  DIVAS project dealing with cooperative systems.
• In the coming next months, we will see if our
  approach is relevant or not.