A portable alcohol detection system based on exhaled breath analysis has been developed. The system consists of a breath sensor, a smartphone to control the sensor and communicate various data, and a data cloud system. The system can be used to monitor a driver’s status from a remote location.

The breath sensor consists of four separate sensors. The first is a water vapor sensor that is used to verify if the applied gas is human breath. The others are semiconductor gas sensors to detect ethanol, acetaldehyde, and hydrogen. The detector is connected to a smartphone, and the driver’s alcohol check results are automatically sent to a data cloud system. To prevent abuse of this detector by blowing substitute gas, it can recognize human exhaled breath by detecting small amounts of metabolites as well as saturated water vapor. The ethanol concentration is obtained from the voltages of the three semiconductor gas sensors. Each sensor has a specific sensitivity for ethanol, acetaldehyde, and hydrogen. We apply the differential evolution algorithm to the relationship between each sensor’s output voltage and its calibration curve. Then we calculate the ethanol concentration of the human breath. This multi sensor method is more accurate than a single sensor method that only uses one gas sensor to measure ethanol concentration. We also employ an original humidity sensor that was suitably designed to detect highly humid (saturated) water vapor using a comb-shaped electrode pattern. From our field tests, in which we used more than 30 sets of detectors, we investigated the performance of exhaled breath recognition. When the field test users of our detector did not consume alcohol, their ethanol, acetaldehyde, and hydrogen concentration levels, as a result of natural human metabolism, were 1.8, 1.9, and 0.1 ppm on average, respectively. Based on these data, we set the threshold level on each gas sensor to recognize human breath. As a result, the detector could measure the breath alcohol level within 3 s. It could also successfully distinguish human exhaled breath from ambient air or spray gas.

Since the detector is small (Size:​ 75 mm(W), 55 mm(D), 20 mm(H), Weight:​ 20 g, approx.) and connected to a smartphone, it can be used at any time and can contribute to the safety of professional drivers by enabling remote and central management of alcohol test data.

It is possible to connect this detector to a vehicle engine interlock system by wireless communication so that the system can recognize the driver’s status, even before they enter a vehicle. This will help prevent drunk driving. We believe that our detector is helpful to decrease alcohol-related accidents and enable remote management of alcohol inspection.