Technology Solutions to Driver DistractionOverload

1
Technology Solutions toDriver Distraction/Overloa
d

• Paul Green
• University of Michigan
• Transportation Research Institute (UMTRI)
• Human Factors Division
• pagreen_at_umich.edu,
• www.umich.edu/driving

Presentation at Nation Safety Council
International Symposium on Distracted
Driving October 14-15, 2008, Arlington, VA
2
(No Transcript)
3
What is the problem? (tasks)

• Drivers are under pressure to do work while
  driving (conference calls, email, browse
  Internet, etc.).
• Drivers may want to communicate with others not
  in the vehicle while driving (text message (DWT),
  email, phone).
• Drivers may want to be entertained while driving
  (MP3 player, satellite radio, TV)
• The data indicates that doing other tasks while
  driving increases crash risk.
• Setting and monitoring warning systems can burden
  drivers.
• So, drivers are being asked to/try to do more
  while driving.
• --gt Distraction and Overload

4
What is distraction?

• Distraction
• 1 the act of distracting or the state of being
  distracted especially mental confusion, 2
  something that distracts especially AMUSEMENT.
  
• Distract
• 1a to turn aside DIVERT
• b to draw or direct (as one's attention) to a
  different object or in different directions at
  the same time
• (Merriam-Webster Online dictionary)
• Task switching is disrupted

5
So, what is overload?
Capturing attention vs. too much to do Why
must you pick up the phone when it rings?
What happens when you are engaged in a call
driving?
6
Workload theory and overload

• If the aggregate demand of some collection of
  tasks (total workload) exceeds the drivers
  capability, then the outcomes described early
  occur. Degradation is gradual.
• If the demand exceeds a resource, then
  performance suffers.
• visual can only look at 1 place
• auditory can only listen to 1 auditory stream
• cognitive can only process limited info
• psychomotor only have 2 hands 2 feet

7
What can be done to prevent the crashes and
injuries induced by distraction and overload?
New drivers selection, training, licensing no
generalized time sharing skill Educate, PR -gt
social norm change (time) lessons from alcohol,
seat belts Corporate policy Ban Workload
manager Design out hazards- reasonable and
expected use and misuse
8
What are some design solutions?

• Reduce the demands of the task
• Decrease the time to perform the task(e.g.,
  fewer keypresses to reduce exposure)
• Decrease the visual demand of the task(e.g.,
  speech to reduce cost of exposure, also reduce
  cognitive demand)
• Terminate the task when the driver (feedback if
  driver looks too long)
• Reduce the workload when the demand of driving is
  high (workload manager)

9
Design standards and guidelines

• ANSI Canvass process
• Consensus output
• Input from all affected parties
• Public review of drafts
• Consider and respond to comments
• Right of appeal

10
Design standards and guidelines

• UMTRI, Batelle, HARDIE guidelines
• JAMA
• ESoP
• AAM
• ISO 16673 (Occlusion), 26022 (Lane Change)
• SAE J2364 (15 second rule)
• static, occlusion method

11
ISO Reports and Standards

• ISO 15005 Dialog Management
• ISO 15006 Auditory Information
• ISO 15008 Legibility
• ISO 16352 Warnings Literature Review
• ISO 16951 Message Priority
• ISO 17287 Design Process
• ISO 15008 Visual Behavior

12
How does a workload manager work?

• Measures the factors that affect the primary task
  of driving (traffic, road geometry, vehicle
  capability, road surface, visibility (both
  affected by weather) -gt determines total workload
  (by equation or table look up)
• Determines individual capability
  (age/experience-related), parent setting for
  teen?
• Determines the visual, auditory, cognitive, and
  psychomotor demands of other tasks-gtdetermines
  conflict/overload

13
What does the system measure?

• Traffic distance to lead vehicle, lead vehicle
  acceleration, distance and acceleration of other
  nearby vehicles (requires Lidar/radar from ACC)
• Road Geometry road class (expressway, major
  arterial, etc.), straight or curve (and curve
  radius), grade, posted speed (requires navigation
  system)
• Vehicle Capability speed, safety systems
  present
• Road Surface coefficient of friction (from
  traction control)
• Visibility sight distance (from cameras,
  weather radio, wiper defroster state, or other
  sources)

14
What does a workload manager then do with that
information?

• Lock out some tasks
• When the workload exceeds some level (raining
  in heavy traffic etc.), the system locks out
  some tasks,
• receive incoming phone calls
• delay some system reminders
• enter street address, intersection, etc.
• search certain web site
• watch TV
• Alter presentation of warnings

15
Implications for other distraction/workload
mitigation approaches

• For now, banning use of all cell phones and other
  devices (except for emergency use) make sense,
  but in the future we may need to revisit the ban
  for certain workload managers
• Workload manager data may be insightful in
  training people to drive safely

16
Implementation Challenge Carry Ins

• Often the secondary task
• Need data for workload of carry-in tasks
• Need way to communicate the tasks and its
  workload to the vehicle (no protocol now)
• Need agreement that vehicle can control the
  carry-in device
• Regulation?

17
Initial Implementation

• Maneuver based (in process vs. expected)
• Turn at intersection
• Merge or exit freeway
• Change lanes
• Park (parallel, into stall)
• Back up
• Reliable/easier to detect
• More likely to be accepted by consumers

18

• This will help some, but many crashes occur under
  benign conditions. In fact, most distraction
  crashes occur in good weather and the driver was
  not drunk.

Eventually, we will need to go beyond
maneuver-based implementations.
19
Other reasons workload managers matter
Comparison of studies Design of roads
20
What needs to be done to support the development
of workload managers?

• 1. Overall workload
• Publish and discuss the equations from the
  SAVE-IT project
• Continue the research from the SAVE-IT project to
  consider a broader range of test conditions
  (night, bad weather, other road types)
• Validate the simulator estimates on the road
• Encourage others to report aggregate workload
  values based on SAVE-IT and other studies in
  reports and articles

21
What needs to be done to support the development
of workload managers?

• 2. Task demands
• Expand the collection of tasks for which VACP and
  duration values are report
• For published studies, compute the values
• Look at the relationship between VACP values,
  duration, and measures of workload

22
What needs to be done to support the development
of workload managers?

• 3. Develop models of driving
• Develop human performance models of workload.
  The lack of models make automotive HF research
  appear second class.

23
Workload managers should be part of the solution
of the driver distraction/overload problem.
Paul Green UMTRI pagreen_at_umich.edu
www.umich.edu/driving