Research On Switching And Evaluation Method Of Intelligent Vehicle’s Driving Control Considering Driving Style

In today’s world,the emerging industries represented by mobile Internet,big data and cloud computing are developing vigorously,and the traditional manufacturing industry represented by automobile is also gradually moving towards "intelligent manufacturing" for transformation and upgrading.Automobiles,which were simple combination of mechanical structure in the early days,have now embarked on the road of electrification,intelligence,networking and sharing.After the realization of intelligent vehicle,its safety problems haven’t disappeared,but shifted in the form of existence from the driver’s wrong operation,mechanical system failure to the imperfection of intelligent system and the incompletion of humancomputer interaction.Among those unsafe factors,the switching of driving control,especially the driver’s taking over behavior,can easily lead to dangerous working conditions.Scholars from home and abroad are still in their infancy in the research on the switching of driving control authority,mainly from the perspectives of the definition,classification,strategies,and influencing factors of the switching.This paper is going to start from the perspective of driving style,and explore the differences between drivers of different styles in the face of control switching and takeover through the co-simulation of Carsim and Simulink software.Then,based on the simulation data,this paper builds a comprehensive evaluation model,and makes a reasonable quantitative evaluation of the driving control switching behavior with reference to multiple levels of indicators.The main contents of this paper are as follows:(1)Driving data collection and driving style classification: Based on the experimental vehicle and data collection device of the research group,an experimental platform was built,and different drivers were driving on the actual road in order to collect driving data.Thirteen characteristic parameters that can characterize the driving style were selected.Firstly,principal component analysis is used to reduce the dimension of the data,and five principal components are retained.Then K-means clustering algorithm is used to classify the data,and the driving style is divided into two types: aggressive and robust.(2)Simulation experiment of driving control switching under typical working conditions: This paper first analyzed the scenes that need to switch the driving control.First,there are functional problems,that is,the operating conditions that the vehicle faces exceed the operational design domain(ODD)of the intelligent driving assistance system;second,there are functional safety problems,that is,the intelligent driving assistance system causes unexpected failures due to malfunctions.Then,this paper used Carsim and Simulink software for co-simulation to simulate that under different working conditions,drivers of two driving styles respectively take over the vehicle,and brake to avoid collision.(3)Building a comprehensive evaluation model of driving control switching: This paper considered the risk of working conditions,operational redundancy,occupant comfort and feasibility,selected the corresponding evaluation indexes,and established a comprehensive evaluation model of driving control right switching based on Analytic Hierarchy Process,which can effectively evaluate the switching of driving control,and provide some reference for the design of intelligent vehicle’s human-computer interaction and driving control switching.