Preliminary Studies of Haptic Displays for RearEnd Collision Avoidance

1
Preliminary Studies of Haptic Displays for
Rear-End Collision Avoidance
Louis Tijerina Ford Motor Co.
2
Acknowledgements

• Work completed while on staff at Transportation
  Research Center Inc. under NHTSA funding
  
• Scott Johnston, Hung Pham, Marc Winterbottom of
  TRC and Frank Barickman, NHTSA prepared
  instrumentation
  
• Edwin Parmer served as experimenter
• Full Report No. DOT HS 809 151

3
Haptic Display Conceptsfor Manual and ACC Driving

• Haptic Braking for collision warning
• Monopulse braking for rear-end crash imminent
  warning (e.g., Kiefer et al., 1999)
  
• Triple pulse braking used in Intersection CAS
  (e.g., Bittner, et al., 1999)
  
• Pulse braking might indicate ACC deceleration
  authority reached

4
Study Objectives

• Study 1 Monopulse braking display parameter
  setting
  
• Study 2 Assessment of monopulse braking
  displays in car following for True Positive and
  False Positive conditions

5
Study 1 Monopulse Brake Parameter Setting
Approach

• Subjects N6 TRC drivers 27 to 56 years
• Test Vehicle 1995 Lumina, MicroDAS
• Pulse Brake Torque motor attached via wire cable
  pulley to arm of brake pedal, under computer
  control
  
• Distracter Magellan RGS destination entry
• Masking noise Ocean surf sound played over
  headphones

6
Second-order, Central Composite Study Design
in 2Variables, Alpha 1 (Orthogonal Design)
Treatment Coded Uncoded
Comb. X1 X2 J (g/s) D (s)
1 -1 -1 0.08 0.25
2 -1 1 0.08 1.00
3 1 -1 0.32 0.25
4 1 1 0.32 1.00
5 0 0 0.20 0.65
6 1 0 0.32 0.65
7 -1 0 0.08 0.65
8 0 1 0.20 1.00
9 0 -1 0.20 0.25

7
Study 1 Procedure

• Subject oriented to destination entry task,
  vehicle, and test site
  
• All trials on TRC skid pad 1 mile straightaway
• Subject drove a nominal 45 mph (72 kph) , no lead
  vehicle present
  
• Half distracted, half non-distracted trials
• Subject instructed that when haptic display was
  noticed bring vehicle to a controlled stop.

8
(No Transcript)
9
(No Transcript)
10
Study 1 Conclusions

• Display parameter effects linear, additive
• Individual differences accounted for most
  response variability across measures
  
• 0.32 g/s jerk rate and 0.65 s duration profile
  was best
  
• Unexpected finding magnitude of braking response
  was determined by haptic display

11
Study 2 Pulse Braking Displays Car Following
Events

• Subjects N7 TRC drivers, 21 to 65 years, new to
  study
  
• Test Vehicle Same as Study 1
• Pulse Brake Same as Study 1
• Distracter Same as Study 1 but only Distraction
  Trials were run
  
• Masking noise Same as Study 1

12
Study 2 Car Following Procedures

• Surrogate Lead Vehicle (LV) towed rear-end
  shell of 1997 Ford Taurus rear end,brake lights
  disabled
  
• Adaptive cruise control in Test Vehicle developed
  for consistent initial conditions
  
• 45 mph (72 kph) 2.0 s time headway, nominal
• LV braked occasionally at 0.35 g to a stop
• Following Vehicle behind Subject Vehicle

13
Study 2 Procedures (cont.)

• Same As Study 1, except
• Distraction trials only
• All combos of pulse braking displays presented
  when surrogate LV braked at 0.35 g (TP trials)
  _at_Time to Collision
  
• All pulse braking displays presented when
  surrogate LV was not braking (FP trials)
  
• Catch trials (True Negatives) interspersed
• Instruction ...As soon as you notice that the
  lead vehicle is braking to a stop, bring your
  vehicle to a controlled stop

14
(No Transcript)
15
True Positive (TP) results

• No significant effects of Jerk Rate or Duration
  on
  
• Brake Reaction Time
• Maximum Brake Pedal Force
• MinimumTime-to-Collision (TTC)
• Max Deceleration
• Interpretation Drivers braked according to
  constraints of lead vehicle behavior, not display
  parameters

16
False Positive Trials Inappropriate Brakings
Duration, X2
Jerk Rate, X1
17
False Positive TrialsInappropriate Braking
Actions
18
False Positive (FP) Results

• Inappropriate brake applications on 20 of 63
  trials
  
• Somewhat idiosyncratic 40 of inappropriate
  brakings from 1 test participant
  
• Inappropriate braking actions were mild, brief
• Interpretation Ideomotor compatibility theory
  prediction holds, in part.

19
Concluding Remarks

• Haptic braking displays are, de facto, present in
  ACC
  
• Pulse braking may provide useful display when
  visual channel is occupied
  
• Present results suggest guarded optimism
• Pulse braking has many implementation issues as
  collision avoidance, e.g., multimodal
  presentation, acceptability