An important need exists in the transportation industry for a reliable, unobtrusive device to warn vehicle operators of reduced alertness. Researchers suggest that 35-50% of all highway fatalities are caused by measurable levels of reduced driver alertness. Unfortunately techniques presently used to measure driver alertness are impractical for widespread use. Studies do agree that the autonomic, spontaneous eye blink response is the single most reliable indicator of alertness. Proposed is a device that precisely measures multiple blink characteristics in real time and outputs audio alarms based on the degree of estimated driver dysfunction. Alarms are reset when blink characteristics return to normal. The low cost, wireless system envisioned is compact and can readily interface with data acquisition devices and on-vehicle computer systems. Phase I will demonstrate blink sensor signal quality by comparing time coded video records with computer acquired and enhanced blink data. Phase II will concentrate on automation of signal interpretation, alarm algorithms, miniaturization and validation.
