2103 ドライバ距離学習モデルを用いた車間距離誤認識判定・修正システムの提案(要旨講演,メカニカルシステムとその知能化)

概要

論文の詳細を見る
To reduce the number of traffic accidents from driver's misunderstanding of headway distance, driver's subjective sense error correction method is proposed. First the subjective sense error is estimated using a back-propagation neural network. Then the method provides a warning based on the estimated driver's subjective sense of headway distance. A warning threshold value change mechanism is applied for reducing driver distraction. The threshold for starting providing warning is adjusted according to the vehicle's acceleration and speed relative to the car running just ahead of the driver. It was tested successfully on 7 people, reducing the subjective distance error up to 3%. The model is built in order to make it adaptable in real time to the changes in driver's subjective sense of headway distance. This proposed method can be implemented on conventional devices (such as a Head-Up Display, HUD), in order to reduce car collision accidents.
2007-03-19