Abnormal driving detection model establishment method and device and storage medium

Abstract

The invention discloses an abnormal driving detection model establishment method, and the method comprises the steps: obtaining a plurality of first samples in a normal driving state, and enabling each first sample to comprise first modal data, collected by a plurality of collectors, in the same normal driving time period; Inputting the first modal data of the first sample into a first abnormal driving state detection model, and outputting a first prediction error value corresponding to the first sample; Adjusting parameters of the first abnormal driving state detection model by using a back propagation algorithm according to the first prediction error values corresponding to the plurality of first samples to form a second abnormal driving detection model; And setting a second abnormal driving detection model to judge a prediction error threshold value of the abnormal driving state. The invention further discloses an abnormal driving detection model establishing device and a storage medium. According to the method, the detection model of the abnormal driving behavior can be established by integrating the multi-modal data, and the detection precision and robustness of the abnormal driving detection model are improved.

Description

technical field [0001] The present invention relates to the field of communication technology, in particular to a method, device and storage medium for establishing an abnormal driving detection model. Background technique [0002] In recent years, with the rapid development of social economy and large-scale construction of urban roads, the number of cars has increased exponentially, and traffic accidents have also continued to increase. Traffic accidents caused by distracted driving, emotional driving, and sudden illnesses are also gradually increasing. Therefore, monitoring the driving behavior of the driver and giving an alert to the abnormal driving behavior is of great significance to improving the driver's driving ability and reducing his driving load, and essentially reducing the occurrence of traffic accidents. [0003] The current abnormal driving detection model establishment methods include detection methods based on wearable sensors, detection methods based on v...

AI Technical Summary

Problems solved by technology

[0003] The current abnormal driving detection model establishment methods include detection methods based on wearable sensors, detection methods based on visual information, and detection methods based on vehicle driving status. The detection methods based on wearable sensors include motion sensors and physiological sensors. Researchers use motion sensors to collect characteristics such as changes in the driver's body posture, body movements, and operating behaviors, and use physiological sensors to collect characteristics such as the driver's respiratory rate, heart rate, heart rate interval, and skin electricity, and integrate multi-information features to detect various abnormal driving conditions. , but due to the redundancy and a lot of noise in the multivariate time series data, the generalization ability of the model is reduced, making the detection accuracy of this method poor; the detection method based on visual information usually uses machine vision technology or sensor technology to detect the facial features of the driver , such as eye features, pupil diameter changes, gaze direction changes, and mouth status, etc. to study driver fatigue, distraction, emotional changes, etc., but this method is largely affected by the environment and camera shooting in the real-time detection process. The influence of the angle, such as when the face is blocked or the light changes strongly, will affect the effectiveness and accuracy of the detection results; the abnormal driving state detection method based on the driving state of the vehicle is based on the existing equipment of the vehicle, collecting the vehicle itself Speed, lateral acceleration, lateral displacement, lane departure, vehicle trajectory changes and other characteristics, but limited by the specific model of the vehicle, the specific conditions of the road, and the driver's personal driving habits and driving experience, this detection method is not universal

[0004] The same abnormal driving behavior has changes in the modal data of the three detection methods, but the above three detection models are established based on a single modal, and the detection of a single modal data is adopted without fusing the modal data method, not only cannot avoid the problems of the limitations of the single mode itself, but also has low support for the judgment and detection of abnormal driving behaviors, resulting in low detection accuracy and poor robustness of abnormal driving behaviors. Poor detection of behavior

Images