Structural Equation Modeling for Quantifying Riding Performance of Motorcycle Rider using Real-time Measurable Indexes

Abstract

Motorcycle riders' fatality is four times that of four-wheeled vehicle drivers. Previous studies have shown the effect of the Advanced Rider Assistance System (ARAS) is different depending on the user's driving style. To realize personally optimized ARAS, it needs to keep track of riding performance and emotional state. Most studies define one index as driving performance to control the onset timing of ARAS. In this study, we designed a structural equation model to identify the driving behavior indexes that are directly related to the risk of traffic accidents from the emotional state and driving behavior. We investigated the driving behaviors of 23 test subjects using a riding simulator by inducing various emotional states in different conditions of driving scenery, traffic volume, and music. As a result, this model suggests that arousal level, valence level, carelessness, lateral instability, steering instability, and driving style are related to riding performance.