Research On Ethical Decision Making Of Autonomous Car Intelligence

In order to improve the safety and efficiency of road traffic,autonomous cars driving has become the inevitable trend of future road traffic development.Autonomous cars will replace drivers to realize intelligent driving operations such as making driving decisions,controlling vehicles and ensuring safe driving.In order to ensure the safe landing use of autonomous cars,it is necessary to ensure that autonomous cars can make safe driving decisions and conform to human ethics,especially under the dilemma of decision-making.Therefore,That is great research value and practical significance to carry out intelligent ethical decision-making research on the ethical dilemma of autonomous cars.This paper is supported by the outstanding youth fund of Jiangxi province,key research and development projects of Jiangxi province and other related scientific research projects.It takes autonomous cars as the research object and aims to realize the ethical decision-making of autonomous cars.On the basis of reviewing and summarizing the research status at home and abroad,systematic and in-depth analysis and research are carried out respectively from the aspects of unmanned decision-making analysis and ethical decision-making modeling,and on this basis,the GUI for ethical decision-making of autonomous cars is designed.Its main research contents and conclusions are as follows:(1)Taking "trolley dilemma" as the prototype of ethical dilemma,this paper analyzes various versions of trolley dilemma and current ethical dilemma scenarios of autonomous cars,and obtains the main characteristics of ethical decision-making dilemma of autonomous cars.The four-stage ethical decision-making process theory is used to make a deep and systematic analysis of the ethical decision-making process of autonomous cars,and the influencing factors of the ethical decision-making process are analyzed respectively from individual factors,organizational factors and the decision-making problem itself,and the two decision-making characteristics of moral strength and decision-making risk degree are summarized.(2)By analyzing the characteristics of the moral machine test data set and extracting the characteristics of six decision-making dilemma scenarios,and combining the meaning of moral strength and decision-making risk degree,the quantitative calculation formula of moral strength and decision-making risk degree is constructed;The main purpose of this paper is to explore ethical decision-making solutions for autonomous cars,so the calculation is simplified in the form of quantifying eigenvalues only by the number of attributes.The source data sets were classified according to the characteristics of the six decision dilemma scenarios,and the characteristic values of the classified data sets is calculated by using the two quantitative calculation formulas of the decision characteristics.The characteristic values is used as the coordinate to make the distribution diagram of the decision characteristics under the six scenarios,and the decision rules is summarized.(3)This paper selects Bayesian learning as the basic modeling method based on the comprehensive characteristics of decision feature distribution and the performance requirements of autonomous cars decision-making,that is a classical pattern recognition method.According to the basic principles of Bayesian learning,the theoretical model of ethical decision making for autonomous cars is established.The model of the minimum error rate decision making is obtained from the model training of the data of six scenarios,and the feasibility of the model is further verified by using the test data set.In order to adapt to the actual driving decision,the decision judgment is modified.(4)According to the required functions,the MATLAB/GUI tool is used to design the GUI decision-making interface and parameter interface for the ethical decision-making of autonomous cars.In addition,the ethical decision-making model is embedded by writing control functions to construct the GUI for the ethical decision-making of autonomous cars.The authenticity of GUI is verified by constructing ethical decision scenarios from sampled data.