ROADSIDE INFRASTRUCTURE FOR SAFER EUROPEAN ROADS

1
SafetyNet WP 5
Task 5.2 In-depth accident causation
data Prague, Czech Republic, May 2006 Helen
Fagerlind, Chalmers University of Technology
Project co-financed by the European Commission,
Directorate-General Transport Energy
2
Involved Partners
Chalmers SE (Task Leader)DITS ITMUH DETNO NLVA
LT FIVSRC UK
VALT
Chalmers
TNO
VSRC
MUH
DITS
3
Background

• Use an existing accident investigation network to
  capture data on approx. 1000 road accidents to be
  recorded in an in-depth accident causation
  database
• Contributing to reducing number of accidents
  within the EU
• Provide a road accident databank on the causation
  of accidents for data users

4
Potential Use of the Information

• Informing and monitoring road and vehicle safety
  policy
• Development of future accident avoidance
  technologies
• Education to help prevent similar future
  accidents
• Assessing road infrastructure weaknesses

5
Pilot phase

• From 1st Nov 2005 to 31st January 2006
• Each partner collected minimum 5 cases
• two cases collected and database feedback
  provided by 15th of December 2005
• three cases (or more) collected and more database
  feedback provided by 31st January 2006

6
Review phase

• Aimed to assess proposed data gathering
  practises, make amendments to procedures
• Examine whether partners had successfully managed
  to retrieve high quality data
• Assessed the usability and effectiveness of the
  database

7
Phases and progress to date

• Needs of data users ?
• Development of database ?
• Infrastructure and team training ?
• Pilot phase ?
• Review of procedures ?
• Full data collection (started 1st May 2006)
• Data analysis

8
Methodology

• General variables
• SNACS (SafetyNet Accident Causation System)
• is used for causation case analyses
• has a Man-Technology-Organisation perspective
  (Road user, Vehicle, Road environment)
• is based on DREAM (Driving Reliability and Error
  Analysis Method)
• implies that a variety of interacting factors
  creates the accident
• surveys the causes of accidents
• does NOT focus on blame

9
SNACS Methodology

• Three tools
• A system to describe the Common Performance
  Conditions (CPC) that affects all drivers
  regardless (weather, light, etc)
• A categorisation system (Critical Events and
  Contributing Factors) which lists possible causes
  and consequences in an accident/incident event
• A step-by-step procedure description for how to
  perform the case analysis

10
General Variables

• to some extent the same as in WP 5.1
• Accident details (approx. 15 variables)
• Vehicle details (approx. 15 variables)
• Roadway details (approx. 25 variables)
• Road user details (approx. 25 variables)

11
SNACS Variables

• Critical Events
• Timing
• Duration
• Force/(power)
• Distance
• Speed
• Direction
• Object
• Sequence

• Contributing Factors
• Road User
• Vehicle
• Infrastructure
• Organisation

12
SNACS Variables - Contributing Factors
13
Example Case

• Wednesday, 1130, December
• 50 km/h, dry road, sunny weather
• Undivided roadway (single carriageway)
• Involves a Mazda 626 and a Suzuki Jimny
• Impacted front to front, off-side (left)
• One person (driver) in each car
• Only slight injuries on both drivers

14
Accident Scene
15
Accident Scene
Mazdas travelling direction
Suzukis travelling direction
16
Accident Description

• Suzuki Driver
• female 40-50 years old
• on her way to go shopping with her mother
• stopped for petrol and turned out onto the main
  road
• on this part of the road there is quite a long
  and straight stretch where the visibility is good
  in both directions

• Mazda Driver
• 69 year old female
• drives in the opposite direction towards the
  Suzuki
• been shopping and is on her way home
• lives close to the petrol station
• looking at a newly built fence when she pass it
• the fence has made some of the local people a
  little bit irritated

17
Accident description

• the Suzuki driver sees the Mazda crossing the
  median, where it continues driving towards the
  Suzuki
• the Suzuki driver believes it is someone who
  knows her, who is just having a bit of fun (since
  her car is quite easily recognized)
• the Suzuki driver thinks the Mazda will return to
  its own lane so the Suzuki driver does not take
  action to avoid the situation until the very last
  second when she realises they are about to
  collide
• at that point the Suzuki driver tries to steer to
  the right to avoid the collision but the two cars
  collide and impact mainly on the left front side
  of each car
• the Mazda comes to rest on the road and the
  Suzuki is going down into the ditch where it
  impacts a utility post with its rear

18
S
M
19
S
M
20
S
M
21
S
M
22
S
M
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
SNACS analysis of the Mazda
Temporary inability The driver may have fallen
asleep. (She had been up since 0600).
Inattention The driver was not paying attention
to where the vehicle was heading.
Direction- incorrect The driver drove the car
onto the roadway of the oncoming traffic.
External competing activity The driver is busy
looking at a newly built fence which is annoying
some of the local people.
Missed observation The driver was not
observing/paying attention to where the vehicle
was heading.
Distraction The drivers primary focus was not
on the road.
29
SNACS analysis of the Suzuki
Cognitive bias The driver of the Suzuki thinks
the Mazda driver is crossing the median on
purpose, along the line of thinking that that no
"one wants to impact".
Faulty diagnosis The driver thinks the vehicle
coming at her is driven by someone who recognizes
her car and wants to joke a little with her.
Timing- late action The driver realized quite
late that the oncoming car was actually going to
collide with her vehicle and she started taking
action rather late.
Inadequate plan The driver believed it was the
responsibility of the other driver, who entered
the wrong lane to take action and return to the
right lane.
Error in mental model The driver believed it was
the responsibility of the other driver, who
entered the wrong lane, to take action and return
to the right lane.
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
SNACS Results

• The results of the SNACS analysis can be
  aggregated to show causation patterns for
  multiple accidents

Aggregated results from 7 Intersection accidents
- Left turning vehicle