top of page
Robotec_logo_2022-RGB (1) (2).png

 

As technology advances, driver monitoring systems (DMSs) are becoming powerful tools in preventing human-error-related incidents and making roads safer for everyone. They are progressing in detecting driver drowsiness and distraction, and some attempts are being made to teach them to recognize stress or cognitive load (for more insights, see our previous posts on mental workload, regulatory aspects of driver distraction, or visual-based DMS approaches,  and download our handbook on driver stress detection). Driving under the influence (DUI) of alcohol is one of the next steps to address – it is forbidden and penalized. However, different blood alcohol concentrations (BACs) are set in different countries/states as legal limits, from total prohibition up to 0.08% BAC – if you’re from North America or Australia, you probably know such a notation, which means grams of ethanol per one deciliter of blood (g/dL, g%). Europeans more commonly use promilles (permilles), which are equal to grams of ethanol per liter of blood (g/L, g‰), and some countries like the United Kingdom or New Zealand measure BAC in milligrams of ethanol per 100 milliliters of blood (mg%, mg/dL). For reference and conversion, see Table 1 below. 

 

 

Table 1. BAC formats and units. Source: Wikipedia.org. 

Format 

Units 

Used in (among others) 

1 mg% 

1 mg/dL = 0.01 g/L =  = 1 mg/100 mL 

United Kingdom, Ireland, New Zealand 

1 percent (%), 1 g% 

1 g/dL = 1 cg/mL =  = 10 g/L = 1 g/100 mL 

USA, Australia, Canada 

1 permille (‰) 

1 g/L = 1 mg/mL =  = 100 mg/1 dL 

Austria, Belgium, Bulgaria, Czech Republic, Denmark, France, Germany, Latvia, Lithuania, Netherlands, Poland, Portugal, Romania, Russia, Slovenia, Spain, Sweden, Switzerland, Turkey 

 

Is driving while intoxicated a real-world problem? Here's a handful of statistics:  

  • About 32% of all traffic crash fatalities in the United States involve drunk drivers (with BACs of 0.08% or higher; NHTSA, n.d.). In 2022, there were over 13,500 people killed in these preventable crashes – that’s one person every 39 minutes. 

  • According to the European Commission estimates, 25% of all road fatalities across the EU are alcohol-related, and around 6,500 deaths could be prevented each year if all drivers obeyed the law on drunk driving (European Transport Safety Council, n.d.). Yet, according to the European Road Safety Observatory (2006, p. 10), alcohol-related crashes are underreported in the official statistics of most European countries. 

  • World Health Organization (WHO, 2023) estimates that in high-income countries about 20% of fatally injured drivers have BAC levels above the legal limit, while in low- and middle-income countries between 33 and 69% of fatally injured drivers had consumed alcohol before the crash. 


The numbers are horrifying—no wonder the regulatory authorities focus on eliminating driver impairment caused by alcohol intoxication. The issue of drunk driver detection was already directed by the US Congress to the NHTSA (National Highway Traffic Safety Administration, 2024) and mentioned by Euro NCAP’s vision for 2030 (Euro NCAP, 2022), where the latter body declares DUI detection a key real-world priority. 

 


 

What happens when you drink? 

Alcohol is a substance that reduces the proper functioning of the brain, impairing thinking, reasoning, and muscle coordination – abilities that are essential for operating a vehicle safely (NHTSA, n.d.). Intoxication in drivers may negatively influence their cognitive and psychomotor functions, which may result in visible (measurable) changes in their reaction times, perception, non-driving-related tasks (NDRTs) performance, or automation complacency (e.g., Tivesten et al., 2023; Zemblys et al., 2024). According to the reviewed research papers, there are some driving performance indicators affected by the use of alcohol, that include, among others (e.g., Van Dyke & Fillmore, 2014): 

  • the increase of lane position standard deviation (LPSD) following alcohol compared with placebo, indicating less driving precision under the drug, 

  • the increase of steering rate under alcohol compared to placebo (faster and more abrupt movements), 

  • the increase in the total number of centerline and road edge crossings under alcohol compared to placebo, 

  • the increase in average drive speed of the sample under alcohol compared to the placebo, 

  • the increase of inhibition failures under alcohol compared with placebo. 


A general review of the effects of alcohol intoxication on cognitive driving tasks (conducted by Garrisson et al., 2021, as cited by European Commission, 2023) concludes that: 

  • deficits in aspects of visual perception begin at a BAC of 0.3 g/dL, 

  • impairments in vigilance start at a BAC of 0.3 g/dL, 

  • deficits in divided attention and sustained attention commence at BACs between 0.5 g/dL and 0.8 g/dL, 

  • problems with dividing attention over several tasks begin at a BAC of 0.8 g/dL. 

Moreover, individuals with poorer baseline skill levels show greater impairments in response to alcohol (Harrison & Fillmore, 2005). This may be a reason why, in some countries, the legal BAC limit for novice drivers (holding their license for less than two years) is lower than for other vehicle operators. 


NHTSA (2024) initiated rulemaking that would gather the information necessary to develop performance requirements for advanced drunk and impaired driving prevention technology. In this document, the typical and predictable effects of driving over a range of BAC levels are presented (see Table 2). However, BAC is a measure of the amount of alcohol in the bloodstream rather than a reliable indicator of the degree of impairment. Also, ethanol absorption after oral intake depends on various factors, such as the amount ingested, the speed of drinking, the nature and concentration of the alcoholic beverage, previous meal consumption, or physical, biological, and psychological factors (e.g., gender, body weight, and body water composition). What follows, individuals are not always accurate in estimating their level of intoxication (MacDonald, Zanna, & Fong, 1995; Turrisi & Jaccard, 1991, as cited by Gustin & Simmons, 2008) and thus, may drive when they are unaware that they are intoxicated. 

 



Table 2. Typical and predictable effects of driving over a range of BAC levels.  Source: National Highway Traffic Safety Administration, 2024. 

 

Many countries, including the United States, Australia, and Sweden, introduced alcohol interlocks as a means of preventing drunk driving in previous DUI offenders, being part of rehabilitation programs alternative to driving bans. In Sweden, they are also used in government and company fleet cars. An alcohol interlock is a device that requires the driver to blow into an in-car breathalyzer before starting the ignition. If breath alcohol concentration (BrAC) exceeds the set limit, it will not be possible to start the vehicle.  Research has shown that alcohol-interlock programs (AIPs) reduce the risk of recidivism by 75% during the period when the interlock is operational, they reduce alcohol-related fatalities, and they have favorable cost-benefits (sources as cited by the European Commission, 2023). This year, alcohol interlock installation facilitation has already become mandatory for new cars, vans, trucks, and buses in Europe (Official Journal of the European Union, 2019). 



 

What are the next steps? 

More and more research on DUI has been done in recent years. However, it should be mentioned that most of the tests found in the literature were conducted on “scientific grade”, but stationary driving simulators. The outcomes of such studies gave a little more insightful view of the performance of intoxicated drivers and behavioral changes associated with alcohol consumption. However, while it is very important to have access to any data on the physiology and behavior of drunk drivers, it is of great interest to the automotive industry that this data is collected under more naturalistic conditions. It is, therefore, encouraging to see that in the last couple of years, there have been several papers written about research conducted on test tracks (e.g., Tivesten et al., 2023; Zemblys et al., 2024). 


Driver Monitoring Systems may be of great importance in detecting signs of impaired driving. Camera- and sensor-based solutions can warn and intervene whenever a driver’s eye movements and gaze patterns, steering behavior, or presence of alcohol gas in the air indicate potential impairment. There are, of course, some difficult issues in creating robust DUI-detecting DMS. First of all, alcohol consumption effects are nonspecific – it is impossible to accurately determine a person's BAC based solely on their observable physical appearance. Secondly, alcohol impairment is most apparent in complex tasks or tasks that demand coordination on multiple dimensions – this highlights the importance of preparing or selecting the relevant test protocol. Last but not least, is that the effects of drowsiness are similar to those of alcohol, making it difficult to create an algorithm specific to alcohol (Arnedt et al., 2001). On the other hand, both conditions contribute to the driver's inability to drive, and both might merit similar countermeasures (e.g., safety maneuvers bringing a vehicle to stop), so distinguishing between the two is not necessarily a requirement for an effective system functioning. 


We believe that comprehensive efforts shall be taken to ensure sobriety on roads – starting from education, enforcement, and license revocation to preventive instruments. Although drunk driving prevention systems are (as of today) mostly based on sensor modules measuring the alcohol content in the exhaled breath of a driver (such as alcohol interlocks), some camera-based measures seem to be promising means to recognize and reduce the level of alcohol-impaired driving. 


Our vision is to introduce more widely available DUI detection – the one supported by Driver Monitoring Systems, as such systems can be functionally adapted to help reliably identify drivers impaired by alcohol. If you have similar goals and would like to explore potential research- or project-related activities, contact us at humanfactors@robotec.ai!  


References 


Arnedt, J. T., Wilde, G. J., Munt, P. W., & MacLean, A. W. (2001). How do prolonged wakefulness and alcohol compare in the decrements they produce on a simulated driving task?. Accident; analysis and prevention, 33(3), 337–344. https://doi.org/10.1016/s0001-4575(00)00047-6  


European Commission (2023). Road safety thematic report – Alcohol and drugs. European Road Safety Observatory. Brussels, European Commission, Directorate General for Transport. 


European New Car Assessment Programme (2022). Euro NCAP Vision 2030: a Safer Future for Mobility. https://www.euroncap.com/en/press-media/press-releases/euro-ncap-vision-2030-a-safer-future-for-mobility/  


European Road Safety Observatory (2006) Alcohol, retrieved January 25, 2007 from www.erso.eu  

European Transport Safety Council. (n.d.). Drink Driving, https://etsc.eu/issues/drink-driving/, access: 18.08.2024 


Gustin, J. L., & Simons, J. S. (2008). Perceptions of level of intoxication and risk related to drinking and driving. Addictive behaviors, 33(4), 605–615. https://doi.org/10.1016/j.addbeh.2007.11.010  


Harrison, E. L., & Fillmore, M. T. (2005). Are bad drivers more impaired by alcohol? Sober driving precision predicts impairment from alcohol in a simulated driving task. Accident; analysis and prevention, 37(5), 882–889. https://doi.org/10.1016/j.aap.2005.04.005  


National Highway Traffic Safety Administration. (2024). Advanced impaired driving prevention technology. Federal Register, 89, 830-857. https://www.federalregister.gov/documents/2024/01/05/2023-27665/advanced-impaired-driving-prevention-technology  


National Highway Traffic Safety Administration. (n.d.). Drunk Driving. https://www.nhtsa.gov/risky-driving/drunk-driving#resources, access: 18.08.2024 


Official Journal of the European Union. (2019). Regulation (EU) 2019/2144 of the European Parliament and the Council of 27 November 2019. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32019R2144&from=EN.  


Tivesten, E., Broo, V., Ljung Aust, M. (2023). The influence of alcohol and automation on drivers’ visual behavior during test track driving. Transportation Research Part F: Traffic Psychology and Behaviour, 95: 215-227. http://dx.doi.org/10.1016/j.trf.2023.04.008 


Van Dyke, N., & Fillmore, M. T. (2014). Acute effects of alcohol on inhibitory control and simulated driving in DUI offenders. Journal of safety research, 49, 5–11. https://doi.org/10.1016/j.jsr.2014.02.004 


World Health Organization. (‎2023)‎. Global status report on road safety 2023. World Health Organization. https://iris.who.int/handle/10665/375016. License: CC BY-NC-SA 3.0 IGO 


Zemblys, R., Ahlström, C., Kircher, K., & Finér, S. (2024). Practical aspects of measuring camera‐based indicators of alcohol intoxication in manual and automated driving. IET Intelligent Transport Systems. https://doi.org/10.1049/itr2.12520  

 

 

 

 

bottom of page