• Reduce the greater than 100,000 injuries and deaths associated with drowsiness involving both commercial and passenger vehicles.
  
  
• Develop, test, and evaluate a prototype drowsy driver detection and warning system for commercial and then passenger motor vehicle drivers.
  
  
• Implement an experimental research protocol to independently validate promising drowsiness detection systems, including algorithms previously developed under NHTSA sponsorship.
  
  
• Identify reliable psychophysiological and driver performance indicators of drowsy driving from advanced signal processing and data analysis techniques.

• Drivers are often unaware of their deteriorating condition or, even when they are aware, are often motivated to keep driving.
  
  
• Incipient drowsiness can be observed and measured well before the occurrence of episodes of involuntary sleep.
  
  
• Drowsiness has recently been detected with impressive accuracy by measuring the percentage of eyelid closure over time, i.e. PERCLOS.
  
  
• A drowsiness detection and warning system can help reduce alertness-related crashes by helping to maintain alertness until it is safe to stop and rest.

NHTSA's principal industry partner is Carnegie Mellon Research Institute (CMRI), the principal DoD partner is the Naval Health Research Center, and the principal academic partner is the University of Pennsylvania. University partners at the University of Pennsylvania (funded by NHTSA and OMCS) provide validation of detection systems using sleep deprived subjects in a controlled laboratory setting. Devices shown in the laboratory to reliably and validly detect drowsiness are further examined for reliable and valid operation using commercial drivers under actual driving conditions. CMRI utilizes a comprehensive sensor system on several tractor trailer trucks, and collect over-the-road behavioral and performance data from drivers during their normal course of business. Collectively, the team will identify and then implement the most effective detection algorithm(s) and warning system(s), which improve driver alertness and motivation to stop and rest.

Based on NHTSA's revised problem size estimates for the 5-year period 1989-1993 (Knipling & Wang, 1995):

• An average annual total of 6.3 million police reported crashes occurred during this 5 year period.
  
  
• Approx. 100,000 crashes per year (1.6% of 6.3 million) were identified on Police Accident Reports (PAR) where drowsiness was indicated in a report check-box, and from a review of "Drift-Out-Of-Lane" crashes not specifically indicated but which had drowsiness characteristics.
  
  
• Approx. 1,357 drowsy-related fatal crashes resulted in 1,544 fatalities (3.6% of all fatal crashes), as reported by the Fatal Accident Reporting System.
  
  
• Approx. 71,000 of drowsy-related crashes involved non-fatal injuries.
  
  
• The role of drowsiness in the leading causes and types of crashes may be largely underestimated due to unreported off-roadway crashes, police unable to verify drowsiness, and driver reporting error.
  
  
• In previous trucking summit meetings, drowsiness was named as the number one driving problem

95.9% of annual drowsy driver crashes (a total of 96,000 including 1,429 fatalities) involved drivers of passenger vehicles, whereas 3.3% (a total of 3,300 total including 84 fatalities) involved drivers of combination-unit trucks. However:

1. Drowsiness was cited in .82% of truck crash involvements, versus .52% of passenger vehicle crashes;
2. Expected involvements for combination-unit trucks is 4.5 times greater than for passenger vehicles due to exposure (60K vs. 11K miles/year), operational life (15 vs. 13 years), and night driving;
3. 37% of the truck related drowsy driver fatalities involved individuals outside the truck, as compared to 12% of the fatalities from drowsy passenger drivers.

Detection Models & Warning Systems

• Model Validation Performance & Physiology
• Numbers & Types of Sensors
• Normative Weighting
• Event Driven Models
• Detection & Prediction
• Levels of Alertness
• Individualized vs Generalized Monitoring
• Sleep Requirements & Sleep History
• Driving Exposure & Experience
• Detection vs Maintenance
• Passive Detection vs Active Maintenance
• Costs & Benefits
• Level of Accuracy

Operability & Acceptance

• User Maintenance & Calibration
• Obtrusive Sensors
• Driver/Vehicle Compatibility
• Ease of Use & Intuitive Operation
• Risk Compensation & Migration
• Confounding Useability
• Alertness Restoration & Maintenance
• Singular vs Sequential Warnings
• User Mediation
• Operational Reliability
• Sensors & Algorithms
• False Positives vs Nuisance Alarms
• Costs & Benefits
• Consumer Investment
• CA Technology Synergism & Reuse