抄録
Advanced Driver Assistance Systems (ADAS) is receiving increased attention throughout the world for avoiding traffic accident. The authors' research focuses on "Mutual Assistance" paradigm which drivers will mutually assist each other for safer automobile society. The authors' previous research examined the effectiveness of the mutual assistance system from both the recipient and the assister side with a driving simulator. However, in condition of experiment, examinee did recipient-side and assister-side separately. Therefore, examinee have to do recipient and assister simultaneously to obtain data in a more realistic situation. This paper reveals effectiveness of mutual assistance system for system user in realistic situation. As evaluation items, success rate of warning, mental workload, attitude change rating and collision rate were used in analysis. These experimental results indicate that fully automatic system(it is written automatic level IV in the text) kept high success rate of warning and decrease mental workload. On the other hand, semi-automatic system(these are written automatic level II or III in the text) induce positive attitude changes, and guided feature semi-automatic system(it is written automatic level III in the text) is most effective for assister's collision rate decreasing. These experimental results indicate that guided feature semi-automatic system is the most proper automatic level for mutual assistance system and has potential of restrain risk compensation behavior.
