An Elderly-Oriented Design of HMI in Autonomous Driving Cars Based on Rough Set Theory and Backpropagation Neural Network

Abstract

As the issues of social sustainability and the aging of population becomes increasingly severe, Autonomous Driving technology is increasingly being seen as an important issue for future travel. At present, the Human Machine Interface (HMI) of Autonomous Driving system has certain difficulties for middle-aged and elderly users, which further affects their perception of cars status and operation experience. Therefore, in order to design an elderly-oriented HMI design of Autonomous Driving cars that meets the Kansei needs of middle-aged and elderly users, a design flow based on Kansei Engineering/ Rough Set Theory/ Backpropagation Neural Network is proposed. The HMI of Autonomous Driving cars is taken as an example in this paper. Under the framework of Kansei Engineering research, Kansei intention analysis is carried out. Factor Analysis is used to reduce dimension and cluster the collected Kansei words. By using the morphological analysis method, the HMI samples are deconstructed into 14 different design features. The attribute reduction algorithm in Rough Set Theory is used to identify the key design features of HMI that have important influence on the elderly-oriented level. Backpropagation Neural Network is used to establish the mapping model between the Kansei intention of middle-aged and elderly users and the key design features of HMI. The mapping model demonstrates good fit as the errors between the predicted and actual values in the 4 types of kansei semantic evaluation tests are all less than 5%. So that it could meet the needs of middleaged and elderly users and obtain the design combination with the highest Kansei average value. Based on the experimental results, it is found that the optimal emotion of middle-aged and elderly users could be obtained by overall color type three, mode of speed display type three, font display of rotate speed type three, reminding color of turning type four and mode of fuel indicator display one in the morphological deconstruction table. The research results show that the elderly-oriented HMI design constructed by Kansei Engineering/ Rough Set Theory/ Backpropagation Neural Network can meet the sentimental needs of middleaged and elderly users and can provide a reference example for relevant elderly-oriented design.