× close
The system aims to detect drunk driving, drowsy driving, and distracted driving by detecting five indicators that can be observed on a driver’s face and body.Credit: Michigan Engineering
Engineers at the University of Michigan have shown that a camera similar to those already installed in new cars, combined with facial recognition tools, can read information on a driver’s face and upper body impairments.
This low-cost system has the potential to effectively detect drunk, drowsy, or distracted drivers before they hit the road or while they are driving. As part of the Infrastructure Investment and Jobs Act of 2021, a new federal requirement for all new passenger vehicles to have this protection could be passed as early as 2026.
The standards these new safety systems must meet are currently being discussed, and the National Highway Traffic Safety Administration’s comment period ends on March 5th. Although details are unclear, the UM team is confident that their system can meet the new standards. Meet your requirements in a cost-effective manner.
“These 3D camera technologies are already being used in products such as smartphones, tablets, and mixed reality devices,” said Mohamed Islam, a University of Massachusetts professor of electrical engineering and computer science who is leading the project. Masu. “And these are small, inexpensive cameras that can be easily mounted on the rearview mirror, steering column, or anywhere else in the driver’s seat.
“Many new vehicles already have advanced driver assistance system (ADAS) cameras to track driver attention. These cameras are already mature and cost-effective.”
Islam’s team suggests augmenting existing ADAS cameras with infrared detection and ranging (LiDAR) or structured light 3D cameras that cost around $5 to $10. A proof-of-concept experiment in which data captured by a 3D camera is interpreted using artificial intelligence tools allows him to identify five signs that a driver may be at fault.
- Increases blood flow to the face. When you drink alcohol, your superficial blood vessels dilate, allowing more blood to be pumped to your face, resulting in the redness and swelling associated with being drunk. Blood absorbs the infrared radiation emitted by the 3D (LiDAR) camera, so the increased blood flow creates a difference from the driver’s baseline readings.
- Heart rate. Infrared cameras can also monitor the driver’s pulse by observing blood vessels.
- eye movements. AI can identify eyelid droop and decreased blinking.
- head position and body posture. The driver’s baseline head position and overall posture changes are measured and monitored.
- Breathing rate. The 3D camera can observe the rise and fall of the chest, and the AI can compare the driver’s breathing rate to a baseline. Ingesting alcohol usually slows down your breathing rate.
The research team demonstrated that the system can measure vital signs, detect drowsiness, and provide data that correlates with breathalyzer readings. The researchers are currently working with tier 1 automotive suppliers, including Denso, and tier 2 suppliers, which manufacture cameras, to further develop and potentially commercialize the technology.
The system is cheaper and harder to cheat than in-vehicle breathalyzers, which can cost as much as $200 per vehicle. Breathalyzers can be defeated by having another person perform the test on the driver’s behalf, or by opening a window or diluting the air near the driver.
Although using light to measure blood may raise concerns about fairness after pulse oximeters were shown to give inaccurate readings in patients with darker skin tones. , the infrared light used in 3D cameras is not affected by melanin, which is the cause of skin color.
3D cameras also overcome two other challenges that prevent traditional cameras from performing this role effectively. By measuring only the infrared radiation emitted by the camera, the system is able to ignore changes in ambient light. Additionally, the 3D camera can track the driver’s movements, which prevents the driver’s face from looking different from different angles.
Drunk driving remains one of the leading causes of death on American roads. In 2021, the most recent year for which data is available, there were 13,384 drunk driving-related deaths, an increase of 14% from the previous year, according to NHTSA. The agency estimates that an average of 37 people die each day in drunk driving accidents.
The technology was developed with support from DENSO, Omni Science, and the University of Michigan. Omni Sciences is a UM startup founded by Islam and in which Islam has a financial stake. UM is looking for partners to bring this technology to market.