■ TOP 　 ■ 概要 　 ■ 研究体制 　 ■ 研究成果 ■ アウトリーチ 　■ 解説講座 　 　■ リンク 　■ お問合せ

Ministry of Education, Culture, Sports, Science and Technology (MEXT)

Special Coordination Funds for Promoting Science and Technology - Research and Development Program for Resolving Critical Issues: “Transportation Safety”

 

Project Title:“Situation and Intention Recognition for Risk Finding and Avoidance:Human-Centered Technology for Transportation Safety”

Project Leader:Professor Toshiyuki INAGAKI, University of Tsukuba

 

Term: July 2004 - March 2007

Research budget: 585,000,000 JPY

 

1. Background

Various efforts have been devoted so far to improve car safety. Accidents rates, however, have not been reduced to a low enough level. Under the condition in which the number of cars grows fast, car accidents inevitably increase rapidly even though the accident rate is maintained at the current level. Mr. Jun’ichiro Koizumi, Prime Minister of the Japanese Government, made the statement in his speech in 2003 January that novel safety programs must be launched in order to reduce by half the fatalities of car accidents within the next 10 years.

The research project, entitled “Situation and intention recognition for risk finding and avoidance: Human-centered technology for transportation safety” was proposed in 2004 January to develop human-centered technology for achieving the goal in the Prime Minister’s speech. The proposal was accepted in 2004 May by the MEXT of Japanese Government as a research project in the Research and Development Program for Resolving Critical Issues, “Transportation Safety,” in the framework of the Special Coordination Funds for Promoting Science and Technology.

 

2. Purpose of the Project

It is often said that 70-80 % of the car accidents involve human errors. Note here, however, that some portions of human errors stem from mismatches among driver’s capabilities, vehicle functionalities, and traffic environment. As a matter of fact, controlling high-speed vehicles in a dense and dynamically changing environment is highly demanding for ordinary car drivers. Such factors sometimes impose drivers excessive claims on their abilities for situational recognition, decision-making, and action implementation. Proactive safety technology that finds those mismatches and avoid their associated risks is thus one of keys to automotive safety improvements and reduction of car accidents in an essential and drastic manner.

Our research project aims at developing a proactive safety technology, consisting of a system of advanced methods and techniques, that can (1) detect at an early stage possibletransitions of the driver’s psychological/physiological state into a risky condition that can lead to accident-prone driving conditions, and (2) provide the driver with appropriate countermeasure assistance and support in a situation-adaptive way.

In order to construct drivers’ behavior models, a multimedia database of drivers’ behavior in the real world is under construction. With the use of image processing methods and mathematical modeling techniques, a technological framework is developed for understanding traffic situations and drivers’ intentions. Augmented vision enhancement systems for supporting drivers’ risk perception are also developed within the framework. Multiple sensor fusion methods are developed for identifying driver’s psychological/physiological states to determine (a) whether drivers have lost situation awareness, (b) whether drivers’ intentions are inappropriate for the given situation, (c) whether drivers are inactive psychologically (e.g., due to inappropriate trust or complacency) or physiologically (e.g., due to fatigue).

With the above technologies and findings, adaptive automation shall be developed for commercial vehicles, in which authority of control is traded between a driver and automation dynamically depending on driver’s psychological and physiological states, and time-criticality and situational risks in the traffic environment. Characteristics of senior drivers are also investigated to make Advanced Driver Assistance Systems (ADAS) effective to drivers in the aging society: See, the attached figure, for an overview of our research project.

3. Project Teams

In order to achieve the project goal efficiently and effectively, 10 teams of researchers have been invited from 8 organizations (three universities and five research institutes), and 5 research aspects have been built with those teams:

Research Aspect 1: Adaptive function allocation between drivers and automation

(Aspect leader and Project leader: Prof. T. Inagaki, University of Tsukuba)

Research Aspect 2: IT and mathematics-based methods for understanding situation and intention (Aspect leaders: Dr. T. Kurita, National Institute of Advanced Industrial Science and Technology; Prof. Y. Ohta, University of Tsukuba)

Research Aspect 3: Drivers behavior modeling for risk evaluation (Aspect leaders: Dr. Akamatsu, National Institute of Advanced Industrial Science and Technology; Dr. Fukuto, National Maritime Research Institute)

Research Aspect 4: Chaotic speech voice analysis for cerebral activity evaluation (Aspect leaders: Dr. K. Shiomi, Electronic Navigation Research Institute; Dr. K. Sato, Railway Technical Research Institute; Prof. T. Fujimoto, TohokuUniversity)

Research Aspect 5: Human-interface and education for senior drivers (Aspect leaders: Dr. K. Morita, National Traffic Safety and Environment Laboratory; Prof. K. Tanaka, University of Electro-Communications)

 

For more detail, see next page.