Recognition Method Of Driving Propensity Based On Amap Navigation Data

With the continuous increase of car ownership,the problem of traffic safety has become increasingly prominent.Frequent traffic accidents have caused a large number of casualties and property losses.Existing research results show that the driver’s own factors are important factors causing traffic accidents,and its influence on traffic accidents is mainly manifested as driving propensity.At present,active safety assistance system of automobiles is an effective means to prevent human traffic accidents.However,intent recognition,as the core part of the system,often ignores the influence of driving propensity,which leads to the low effectiveness and accuracy of system warning.Therefore,it is of great significance to improve the accuracy of safety warning and reduce the occurrence of traffic accidents by refining the research on driving propensity of different drivers,deeply studying the method of driving propensity identification,and introducing the driving propensity identification into the active safety driving assistance system of automobiles.In order to realize the identification of driving propensity,the following research work is carried out in this paper:（1）The deficiencies of the data collection methods in previous researches on driving propensity are analyzed,and a dynamic data collection method for Amap navigation is proposed.Nine characteristics data of driving propensity deduced from time,speed and acceleration in the process of navigation are obtained through the construction of dynamic data acquisition application and the design of data acquisition algorithm,which laid a foundation for the research of driving propensity identification based on navigation dynamic data in the next step.（2）The dynamic data deduced by time,speed and acceleration of drivers with different driving propensity during navigation driving are obtained through physiological,psychological and driving experiments,and the collected experimental data are analyzed and processed.The principal component analysis（PCA）technique is used to process the experimental data to extract the characteristic variables of driving propensity.（3）The fruit fly optimization algorithm（FOA）and the generalized regression neural network（GRNN）is integrated to establish a high-precision FOA-GRNN model for driving propensity prediction,which is further trained and verified using collected data.The verification results show that the overall accuracy of the identification model is 94.17%.Compared with the verification results of the GRNN and BPNN models,the FOA-GRNN model has higher identification accuracy and better stability.（4）An identification system of driving propensity is constructed.First,the functional requirements of the system are analyzed to determine the overall framework and construction method of the system.Then,the functional modules in the system are designed and implemented,so that the system can realize the real-time collection,processing and storage of the driver’s data and accurate identification of the driving propensity during the process of navigation.Finally,a real vehicle test experiment is carried out to identify the performance of the system.The overall accuracy of the system’s driving propensity recognition can reach 92.5%,and the recognition precision,recall and F1 score of the three types of aggressive,common and conservative types are all over 90%.The test results show that the overall identification accuracy of the system is high,and it has good performance in precision,recall and F1-Score.In practical application,the system can accurately identify the driver’s driving propensity,which can provide a convenient and new method for the establishment of personalized intelligent driving assistance system.