Research On Test And Evaluation Method Of Human-machine Co-piloting System For Intelligent Vehicles

Human-machine Co-piloting System for Intelligent Vehicles has the function of sharing the driving right of the vehicle with the driver and cooperating with the driver to complete the driving task.It needs to have the autopilot ability of L2 or higher as specified by SAE.Due to current policies,regulations and technical reasons,it is not realistic to achieve full automation of the L5 level in the short term.Therefore,in the foreseeable future,smart cars will be market-oriented in the form of Human-machine Co-piloting.The research of driving system has great practical value and significance.It has been agreed by experts and scholars at home and abroad now.The development of Human-machine Co-piloting systems for smart cars follows the "V" development process.The significance of test and evaluation plays a role in product acceptance and optimization.However,the current evaluation of the Human-machine Co-piloting System measurement area is often limited.The single switching function lacks the evaluation of the overall functional test of the system;its takeover test scenario is relatively single,limited to one or two operating conditions,and the considered test scenarios are not comprehensive;at the same time,the study of the driver's subjective feeling evaluation in the takeover test Lack of research.Based on the national key research and development plan(2016YFB0100904-1)and the National Natural Science Foundation of China(51775235),this paper conducts a research on the evaluation method of intelligent vehicle man-machine base driving system measurement range.It is mainly carried out from the following aspects:(1)Analysis of testing requirements for Human-machine Co-piloting systemsThis article starts with the concept of Human-machine Co-piloting,and analyzes the basic composition and classification of Human-machine Co-piloting,and divides Human-machine Co-piloting into switching and sharing types.At the same time,the test requirements for switching Human-machine Co-piloting are based on the system.The function and test content are described in two aspects.Among them,the switching Human-machine Co-piloting is divided according to the conversion process of the vehicle control right in the functional analysis,and it is divided into 2 types of hosting and taking over,and these 2 types are further divided into 5 types of triggering conditions.The functions required for the system are determined by the trigger conditions.After that,the test content of the Human-machine Co-piloting system is summarized,including the three aspects of functionality,takeover characteristics,and interactive experience,which will lay the foundation for the subsequent determination of specific test conditions.(2)Construction of Human-machine Co-piloting system test platformDesigning the overall architecture of the Human-machine Co-piloting system test platform,and building a switchable Human-machine Co-piloting system test platform that includes a driver's condition monitoring system and a driving simulator.The driver status monitoring system includes a facial information monitoring module,an EEG information monitoring module,and an eye movement information monitoring module;the driving simulator includes a test vehicle module and a traffic vehicle module.(3)Analysis of test content of Human-machine Co-piloting systemA set of test and test methods including the functionality,takeover characteristics,and interactive experience of the Human-machine Co-piloting system is proposed.Among them,the functional tests include normal driving test,fatigue driving test,and cruise control.The takeover characteristics test includes tests on highway exit scenes,highway construction scenes,traffic accident scenes,and unstructured road scenes yield test results on takeover time and takeover quality;the interactive experience is mainly subjective evaluation by the driver acceptance,reliability,Driver's subjective scoring in four dimensions of driving load and user experience.(4)Test and analysis of Human-machine Co-piloting systemSelecting testers to test according to the test content and analyzing the test results.Among them,the relevant analysis method is selected to obtain a correlation between the subjective evaluation method and the objective takeover quality index.The principal component analysis method is used to select valuable evaluations.Indicators,and the corresponding test conclusions are obtained according to the results of the test analysis.(5)Construction and verification of Human-machine Co-piloting system evaluation systemAccording to the test indicators obtained from the analysis of the test results and the reasonable value range of each indicator,the test indicators of each layer are weighted by the analytic hierarchy process,and a reasonable index scoring standard is established.A Human-machine Co-piloting system is established according to the fuzzy comprehensive evaluation method The evaluation system selects two sets of systems containing Human-machine Co-piloting functions to verify the proposed test method and evaluation system,which provides a valuable reference for the development of Human-machine Co-piloting system test and evaluation standards.