Study On User Behavior Dynamic Pattern Mining And Application Based On Intelligent Networked Vehicle Multi-source Data

As time goes on,the states of various elements in the human-vehicle-environmental transportation system will constantly change,such as the improvement of driving skills,the change of living habits and the improvement of road facilities,etc.,and the user behavior pattern will change dynamically accordingly,such as the driver’s preference for medium-low speed driving to high-speed driving.However,the traditional research on user behavior patterns mainly focuses on static pattern recognition,ignoring the mining of dynamic patterns.For automobile enterprises,whether they can timely identify the dynamic changes of user behavior patterns and quickly adjust the direction of product and service innovation based on the insight of user demand changes directly determines the success or failure of commercial competition.The mining of dynamic user behavior patterns has many problems,such as staggered behavior levels,heterogeneous data sources and high complexity.In this paper,the car snatched and intelligent environment product service innovation as the background,the integrated use of data science,behavior science,traffic engineering and design theory,method and technology,in the multi-source data from intelligent made cars to obtain the dynamic changes of the user behavior model based on dynamic model for product service innovation as the goal.The method and application path of user behavior dynamic pattern mining are proposed.The main research contents and achievements include the following aspects:1.A dynamic pattern mining system of user behavior in human-vehicle-environment system is constructed,and the "why" and "what" of dynamic pattern are explored.This paper explores the theoretical basis of the dynamic change of user behavior pattern in human-vehicleenvironment system,and analyzes it from the perspectives of philosophy,psychology,system science,data science and biology.The concepts of user behavior at micro,medium and macro levels of human-vehicle-environment system are systematically discussed,and the connotations of user behavior patterns at micro,medium and macro levels are summarized.Based on the continuity of behavior,the user behavior is divided into continuous behavior and discrete behavior,and the formal definition and solution of dynamic pattern mining problem of continuous behavior and discrete behavior are constructed respectively.2.From the micro perspective,driver emotion is selected as the research object,and a driver dynamic emotion recognition method based on intelligent networked vehicle image sequence data is proposed to explore the problem of "how to recognize" the dynamic pattern of user behavior at the micro level.Facial expression and head posture were considered as multimodal input of driver emotion recognition to improve the accuracy of emotion recognition.An image sequence data processing algorithm based on deep transfer learning is designed to improve the robustness of the algorithm.The driving emotion data collection method based on emotion induction and real vehicle condition is constructed to solve the problem of insufficient driving emotion data set.Finally,it is found that drivers have different emotional fluctuations in different driving situations through experiments,which proves that intelligent networked vehicle image sequence data can effectively identify emotional changes in the process of user and vehicle interaction.3.From the meso perspective,the driving behavior evaluation is selected as the research object,and the calculation method of dynamic evaluation of driving behavior based on the operation data of intelligent connected vehicles is proposed,and the problem of "how to identify" the dynamic pattern of user behavior at the meso level is explored.The driving behavior evaluation content composed of driving style and driving score is defined,which solves the problem that each driving journey cannot be qualitatively and quantitatively evaluated.Driving style recognition method based on clustering analysis and driving score calculation method based on analytic hierarchy process are designed,and the selection and calculation method of characteristic parameters of pattern recognition are discussed in detail.Driving Evaluation Volatility Index(DPVI)was constructed as the Evaluation Index of dynamic changes of user Driving behavior Evaluation results,and the calculation method of DPVI was introduced in detail.Finally,through the actual vehicle operation data set of a certain vehicle enterprise user,it is verified that the above method can effectively identify the dynamic changes of user driving evaluation results.4.From the macro perspective,the travel behavior is selected as the research object,and the dynamic pattern recognition method of travel behavior based on the trajectory data of intelligent connected vehicles is proposed,and the problem of "how to recognize" the dynamic pattern of user behavior at the macro level is explored.A dynamic pattern recognition process based on function fitting and data tagging is designed to analyze the dynamic changes of travel mileage and travel frequency of users.A dynamic pattern recognition process based on density clustering and improved top-K frequent sequence mining was constructed to analyze the dynamic changes of user travel sequences.Finally,through the actual trajectory data set,it is verified that the proposed method can effectively identify the dynamic changes of user travel behavior patterns.5.Proposed the innovation method of intelligent connected vehicle products and services based on dynamic pattern perception and behavior design,and explored the problem of "how to use" the dynamic pattern of user behavior.A product and service innovation process based on "behavior focus-behavior understanding-innovative design-behavior reward-product iteration" was constructed to solve the problem of product and service innovation without theoretical process and excessive reliance on experts and leadership experience.5 product service innovation paths are designed based on behavior reinforcement and behavior inhibition according to user’s behavior capability and characteristics of intelligent connected vehicles.Finally,the results are applied to the interior optimization design and charging pile layout optimization design of an automobile manufacturer,and the effectiveness of the method is proved.