Yes, quoting the risks paper, "policy makers must define performance criteria for these devices". What amount of error rate is low enough? Even air bags "occa

Yes, quoting the risks paper, "policy makers must define performance criteria for these devices". What amount of error rate is low enough? Even air bags "occasionally fail to deploy in higher speed crashes or deploy at lower speeds than anticipated, resulting in injuries and deaths that would otherwise not be expected". A quick look at the literature says DADSS tests your scenario. In "DRIVER ALCOHOL DETECTION SYSTEM FOR SAFETY (DADSS) – PILOT FIELD OPERATIONAL TESTS (PFOT) VEHICLE INSTRUMENTATION AND INTEGRATION OF DADSS TECHNOLOGY" at https://www-esv.nhtsa.dot.gov/Proceedings/26/26ESV-000262.pdf https://www-esv.nhtsa.dot.gov/Proceedings/26/26ESV-000262.pd... > After comprehensive research that investigated optimal sensor placement in numerous locations within the vehicle, the sensor was adapted for installation in the DADSS research vehicles in four different positions: above the steering column, above the glove box, in the passenger’s door panel and in the driver’s door panel. These positions improved analysis of the impact of cabin air flow and the driver’s position on alcohol measurements as well as optimized performance. It also allows a sober driver to operate the vehicle while a dosed passenger can provide samples for analysis. and in proposed field testing: > Drivers will not be permitted to consume alcohol and only alcohol-free measurements will be obtained from the driver side sensors. Passengers will be asked to consume two different amounts of alcohol, approaching a breath alcohol concentration (BrAC) of 0.02-0.03% and 0.04-0.05%, respectively.