Method of detecting of driving under influence of alcohol based on MQ3 sensor and ultra wide band radar

Patent metadata

Field	Value
Publication number	US-11554665-B1
Application number	US-202217897606-A
Country	US
Kind code	B1
Filing date	Aug 29, 2022
Priority date	Aug 29, 2022
Publication date	Jan 17, 2023
Grant date	Jan 17, 2023

Abstract

Disclosed are a drinking under influence (DUI) detection system and a DUI detection method based on an MQ3 sensor and an ultra wide band (UWB) radar. The system includes an alcohol detection module, a respiratory detection module, a signal matching module and a drunk driving alarm module, and each module cooperates to complete a detection. The MQ3 sensor-based alcohol detection module is used for alcohol signal capture, alcohol sequence processing and DUI threshold detection; the UWB radar-based respiratory detection module is used for respiratory signal capture, respiratory signal processing and passenger separation and positioning; the signal matching module is used for periodic signal alignment, sequence feature matching and drinker identity confirmation; and the DUI alarm module is used for a system alarm prompt, a data visualization interface and a DUI information uploading.

First claim

What is claimed is:

1. A method of detecting driving under influence of alcohol based on an MQ3 sensor and a ultra wide band (UWB) radar, wherein an MQ3 sensor-based alcohol detection module, a ultra wide band (UWB) radar-based respiratory detection module, a signal matching module and a driving under influence of alcohol alarm module are used, and each module cooperates to complete a detection; and the method specifically comprises: Step 1 , carrying out an alcohol concentration value threshold judgment and a respiratory signal separation, specifically comprising: Step 1 - 1 , reading an output voltage value of a center console caused by a change of an alcohol concentration in a cab in real time by using the MQ3 sensor included in the alcohol detection module, and calculating a corresponding alcohol concentration value according to a sensitivity characteristic curve of the MQ3 sensor and a circuit output voltage of the MQ3 sensor, with a following expression:
alcohol concentration value=pow(11.5428*35.904* V RL /(25.5−5.1* V RL ),0.6549),

wherein VR L represents the output voltage of the MQ3 sensor, and pow is a standard library function of c; Step 1 - 2 , filtering alcohol concentration value sequences obtained in the Step 1 - 1 through a window, removing outliers and smoothing data, and then preliminarily judging whether a respiratory detection is needed based on an empirical threshold; if the threshold is exceeded, going to Step 1 - 3 , otherwise, circularly executing the Step 1 - 2 to obtain the more accurate alcohol concentration and facilitate a subsequent respiratory matching; Step 1 - 3 , starting the respiratory detection module, collecting respiratory signals of all passengers in a motor vehicle in real time by using the UWB radar, reconstructing collected I- 0 signals, and further obtaining amplitude information and phase information, wherein I and Q are mutually perpendicular components in a digital modulation process; Step 1 - 4 , drawing a distance-time respiratory intensity two-dimensional matrix in the motor vehicle by using the amplitude information and time stamp information, and distinguishing the passengers with different seats according to different distance information between the different passengers and the UWB radar by using different respiratory data presented on the matrix; Step 1 - 5 , extracting a respiratory pattern of each user from the distance-time respiratory intensity two-dimensional matrix, and calculating respiratory rate after preprocessing such as outlier removal, baseline deviation, discrete wavelet transform noise reduction and Butterworth filtering; Step 1 - 6 , establishing a respiratory curve of each user by using the respiratory pattern of each user, and matching the respiratory curve of each user with a real-time alcohol concentration value curve obtained by the MQ3 sensor; Step 1 - 7 , determining a drinker by sorting scores obtained by a feature matching algorithm when a number and identities of the users are known; further, judging that the drinker is drunk driving if the drinker has the lowest score, and then going to Step 1 - 8 ; otherwise, going to the Step 1 - 2 ; and Step 1 - 8 , prompting by the vehicle alarm after obtaining the information of the drunk driving motor vehicle and the drunk driving driver, and then uploading vehicle information to a traffic management department through an Internet of Things communication device of networked vehicles for subsequent processing; Step 2 , executing an abnormal movement detection algorithm, specifically comprising: Step 2 - 1 , using the respiratory signals captured in the Step 1 - 3 as training data if the alcohol concentration value calculated in the Step 1 - 2 exceeds the threshold, and predicting an abnormal value of the respiratory signals by inputting it into a long short-term memory (LSTM); judging that there is an abnormal movement if a difference from the real-time data exceeds the specified threshold, and going to the Step 2 - 2 ; otherwise, circularly executing this step; Step 2 - 2 , verifying whether the corresponding time data in the alcohol concentration curve is an abnormal point by using an exponential moving average algorithm and according to the time information of abnormal data; going to Step 2 - 3 if the corresponding time data is the abnormal point, otherwise, returning to the Step 2 - 1 ; and Step 2 - 3 , concluding that the passenger corresponding to the respiratory curve of the abnormal point is the drinker, and completing a rapid drunk driving detection; and Step 3 , matching respiratory curve features, specifically comprising: Step 3 - 1 , extracting feature values such as the respiratory rate, an amplitude mean square deviation, an autocorrelation coefficient and a waveform factor of each group of signal curves based on the respiratory signals captured in the Step 1 - 3 if the alcohol concentration value calculated in the Step 1 - 2 exceeds the threshold; Step 3 - 2 , using the above feature values as the training data of a support vector machine classification model, and establishing a detection model corresponding to breathers; and Step 3 - 3 , calculating the feature values such as the respiratory rate, the amplitude mean square deviation, the autocorrelation coefficient, and the waveform factor of the alcohol concentration value curve obtained in real time as inputs, outputting whether the passenger drinks alcohol or not, and completing a classification, and determining the identity of the drinker.

Assignees

• Univ Tianjin

Inventors

• Liu Xiulong
• Li Keqiu
• Ma Liyuan
• Zhang Jiuwu

Classifications

• B60K28/06 Primary

  responsive to incapacity of driver · CPC title

• B60W2050/143

  Alarm means · CPC title

• G01S13/931

  of land vehicles · CPC title

• B60W50/14

  Means for informing the driver, warning the driver or prompting a driver intervention · CPC title

• B60W2040/0836

  due to alcohol · CPC title