Virtual-Reality-Based Training System for the Improvement of the Driver’s Ability to Predict Dangers in Driving Situations

摘要

Improving the danger prediction during driving can significantly reduce the risk of accidents.However,previous danger prediction training systems had not been sufficiently effective owing to the lack of realism.In this study,we propose an immersive training system for danger prediction training using virtual reality(VR)technology.This system provides drivers with a highly realistic training environment with 360°videos viewed with VR goggles.Users can practice various dangerous scenarios in an environment that simulates a real-driving situation.In addition,we introduced a system to select dangerous spots with a controller and implement training schemes on a voluntary basis.This setup enables them to train in a highly interactive state.In addition,we proposed a method to express multiple indices numerically so that users can understand the training effect.We tested the effect of the system on the danger prediction abilities of various users with two experiments by using this approach.These results show that our system was more effective in improving the driver’s danger prediction ability than previous systems.