Alcohol-impaired driving contributes to more than 10,000 traffic fatalities every year in the United States. A direct way to reduce alcohol-impaired driving is to identify and immediately check those vehicles in which driving behavior is impaired. Various alcohol-impaired driving detection techniques have been developed including ignition interlock and in-vehicle systems to measure vehicle lateral deviations;however, all these existing alcohol-impaired driving detection techniques rely on special devices installed in a vehicle and most vehicles currently on the road do not include these devices. Moreover, few of these techniques systematically integrate vehicle motion signals with the knowledge and experience of police officers in detecting alcohol-impaired driving. Objective: To address this problem in the detection of alcohol-impaired driving, the current work designs and tests a novel intelligent 3D camera-based alcohol-impaired driving detection and interruption system. This new system will integrate top-down knowledge of experienced police officers with bottom-up data of vehicle motion and other visible information as well as historical accident data and information of time of day and day of week. It will be designed to automatically identify alcohol-impaired driving without relying on any special device in a vehicle and send wireless alerts to police cars nearby to check the problem vehicle immediately. Methods and Aims:
Aim 1 is to collect data on vehicle motion in both alcohol-impaired and normal driving situations as well as the experience and knowledge of experienced police officers in identifying alcohol-impaired driving. It includes three sub-studies: an experimental driving simulator study (60 participants), field study (30 participants), and an interview study (30 experienced police officers).
Aim 2 adopts system design approaches to design and develop this system, and the algorithms of the system will be trained based on the data obtained from the three sub-studies of Aim 1. After the system is developed, it will be installed, improved and field tested on real roads with the assistance of local police (Aim 3). Impacts: It is anticipated that this new system will change the traditional external method for detection of alcohol-impaired driving (i.e., subjective observation of police officers) to automatic detection, thereby increasing the number of problem vehicles being checked, and eventually decreasing fatalities caused by alcohol-impaired driving.

Public Health Relevance

The significant number of fatalities (over 10,000 in 2009) caused by alcohol-impaired driving remains a major concern for public safety and health in the United States. This work proposes the development of a novel intelligent 3D camera-based alcohol-impaired driving detection and interruption system using vehicle motion, other visible information of vehicles, alcohol-impaired driving detection experience of police officers, historical accident data, and information on time of day and day of week. The system will also be tested and improved using real road data with an overall goal of significantly reducing the number of fatalities related to alcohol-impaired driving.

National Institute of Health (NIH)
National Institute on Alcohol Abuse and Alcoholism (NIAAA)
Exploratory/Developmental Grants (R21)
Project #
Application #
Study Section
Health Services Research Review Subcommittee (AA)
Program Officer
Bloss, Gregory
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
State University of New York at Buffalo
Engineering (All Types)
Schools of Engineering
United States
Zip Code
Xiang, Sen; Yu, Li; Chen, Chang Wen (2016) No-Reference Depth Assessment Based on Edge Misalignment Errors for T+D Images. IEEE Trans Image Process 25:1479-94