Visual Coding Design Of Pedestrian Warning System Based On Augmented Reality Head-up Display

With the development of imaging and radar sensor technology,vehicle warning system can recognize and track pedestrians faster,wider and with higher accuracy.Compared with other warning methods,Augmented Reality Head-up Display(AR-HUD)can provide users with real-time and accurate warning information to reduce their driving risks,so it has strong application potential.However,the design standard of the AR-HUD early warning system has not been established yet,so it is of great significance to explore the early warning coding design of AR-HUD for its application.This study first explores the advantages of AR-HUD in attention distribution,behavioral performance and subjective evaluation in pedestrian warning scenarios by combining eye movement technology.Then,the influence of different coding designs on the early warning effectiveness of AR-HUD pedestrian warning information was evaluated from three aspects,and the optimal coding design of AR-HUD pedestrian warning was determined.Finally,shape specificity and dynamic design were adopted to further optimize the shape coding in the optimal coding design of AR-HUD.The purpose of this study is to provide reference for the design standards of AR-HUD warning in the future.This study consists of three experiments: In experiment 1,the Head up Display(HUD)was compared with the Augmented Reality head-up Display(HUD)in the simulated environment.The performance of AR-HUD in the three aspects of attention distribution,behavioral performance and subjective evaluation under the pedestrian warning scenario.Experiment 2: In a simulated driving environment,we explored the influence of the current mainstream color coding,shape coding,flashing coding and relevant multi-dimensional redundant coding designs on the warning effect of AR-HUD pedestrian warning information.Experiment 3: The shape coding in the design of optimal AR-HUD pedestrian warning coding obtained in experiment 2 was deeply optimized,and the influence of shape concretiveness and dynamic effect design on the warning effect was explored.The main conclusions of this study are as follows: 1)Compared with HUD warning,AR-HUD warning is better in terms of visual prominence and comprehensibility,showing faster response time,lower missed alarm rate,better warning score and lower occlusion score of driving environment.2)Color coding can enhance the visual prominence of AR-HUD warning information,showing low false alarm rate and short response time;Shape coding can reduce the comprehensibility of AR-HUD warning information,but can reduce the false alarm rate of the subjects on AR-HUD warning information.Scintillation coding can increase the false alarm rate and response time of AR-HUD warning information.Non-color coding using multi-dimensional redundant coding can increase the visual prominence of AR-HUD warning information,and subjectively it is considered that multi-dimensional redundant coding has stronger warning than single-dimensional coding.3)Increasing the shape specificity of AR-HUD shape coding can improve the comprehensibility of warning information and enhance the situational awareness of the driving environment...