Driver distraction state detection and identification method in mixed traffic environment

Abstract

The invention relates to a driver distraction state detection and identification method in a mixed traffic environment. The method comprises the steps of building a mixed driving scene, collecting human eye movement parameter data and driving performance data of a driver, processing the data and performing preferred screening, building an LSTM driver distraction recognition model based on a fusionattention mechanism, recognizing the distraction state of the driver and the like. According to the invention, the distraction driving identification modeling is carried out by using the LSTM algorithm fused with the attention mechanism; the accuracy and robustness of distraction recognition are improved, the blank of distraction recognition of the driver in the mixed traffic environment is filled, the technical problem of road traffic accidents caused by distraction driving of the driver in the driving process in the mixed traffic environment is solved, and the vehicle driving safety of thedriver is improved.

Description

technical field [0001] The invention relates to a driving state detection and recognition method, in particular to a driver distraction state detection and recognition method in a mixed traffic environment. Background technique [0002] Distracted driving is a leading cause of traffic accidents and constitutes a growing traffic safety problem, according to the U.S. Highway Safety Administration, 25% of traffic accidents are due to distracted driving, a survey based on 1,367 drivers The study found that the large-scale traffic accidents caused by driving distraction accounted for about 14%-33%, and with the gradual development of automatic driving and intelligent network technology, human-driven vehicles, The mixed driving of connected and autonomous driving vehicles will become the norm, and the mixed traffic environment will become more complex, and the factors leading to driving distraction will also increase significantly. Therefore, it is extremely important to recognize...

AI Technical Summary

Problems solved by technology

[0003] At present, most of the distracted driving detection methods are to identify the driver's distraction in the traditional driving environment. However, based on the inevitable mixed traffic environment, there are few studies on distracted identification. The method is generally based on the superposition processing of the driving posture picture and then used as input. The distracted driving recognition method based on the deep convolutional neural network recognizes the distracted driving state, which is based on the traditional driving environment, without considering the eye movement parameters, driving parameters, and driving conditions. Performance parameters, especially eye movement parameters, are important indicators for identifying distracted driving. If the impact indicators of driving distraction cannot be fully considered, it will have a certain impact on the accuracy of distracted driving

Images