Vehicle Drunk Driving Detection System Based On Multi-sensor Information Fusion

The phenomenon of drunk driving has been banned.In response to the demand for drunk driving,a multi-sensor information fusion vehicle-based drunk driving detection system was studied.Based on multi-sensor information fusion technology,a multi-sensor array was designed to construct a vehicle drunk driving detection system consisting of an air intake module,a control module and a data processing module.In order to improve the classification and recognition effect of the target gas on the system,the accuracy of alcohol detection.The sensor array was optimized and the experimental equipment was applied to the actual vehicle test.The specific research work of this paper is as follows:The multi-sensor information fusion technology was applied to the detection and identification of car drunk driving.For the special target gas of the driver's exhaled breath,the metal oxide semiconductor sensor with high commercialization level and mature technology is selected as the main body of the sensor array..The electric signal is converted into a digital signal by a data acquisition device,and together with a specific software,constitutes a vehicle drunk driving detection system.Using the built vehicle drunk driving detection system,the data acquisition experiment for the target gas was designed and implemented.In order to quantitatively evaluate the classification and recognition ability of the original gas sensor array for the target gas,the experimental data is processed by SVM.From the perspective of data dimension reduction and feature selection,the original sensor array is optimized by sensor array optimization algorithm based on principal component analysis,linear discriminant analysis and recursive feature elimination.The test set recognition rate obtained by the support vector machine and the number of sensors in the array are used as evaluation indicators.The performance of the optimized sensor array and its corresponding optimization algorithm is compared.Among the three sensor array optimization algorithms used,the array optimization algorithm based on linear discriminant analysis shows better overall performance.The actual vehicle simulation experiment was carried out by optimizing the algorithm to determine the type of sensor.