Safety

Recently pre-crash safety systems have become available to reduce the damage of accidents.
Also a camera based sensor to detect driver’s facing direction to improve the system response
have been developed to upgrade the system performance. When the system detects that a
driver is not facing forward, the timing of warnings are hastened so that the damage can be
reduced more than the ones without the sensor.
As the needs from the driving support systems is becoming more precise, monitoring the
conditions of drivers such as an eyelid position is another step.
A sensor which detects driver’s eyelid position to detect possible inattentive condition of a
driver who closes eyes for a certain period of time have been developed. System tests under
varieties of lighting conditions and test subjects are also performed.
Robustness of the system was confirmed to be a practical level for actual application to
mass-production cars.

AISIN SEIKI CO., LTD., Japan

TOYOTA MOTOR CORPORATION, Japan

TOYOTA CENTRAL R&D LABS., INC.

Presented at the 15th World Congress on Intelligent Transport Systems, November 16-20, 2008, New York, New York

We are studying a technology to monitor drivers based on a frequency analysis of their heart
rate variability. First, as a concrete target we decided to concentrate on developing a system to
detect driver sleepiness. Using a wheel sensor to get a heartbeat signal, we can detect the early
signs of drowsiness and cope with individual differences through a unique coordinated
analysis of the two axes denoted by the driver’s excitement level and awareness level.
According to our experiment using a driving simulator, our technology helped detect drivers’
sleepiness before critical situations arose.

Fujitsu Laboratories Ltd.

Presented at the 15th World Congress on Intelligent Transport Systems, November 16-20, 2008, New York, New York

In recent years, serious traffic accidents caused by drowsy driving have been big social issues
in Japan. Drowsy driving causes higher risk to driver to be involved in serious traffic
accidents by reducing the attention of driver at the wheel. Therefore, drowsiness detection and
its mitigation while driving are the highest prioritized challenges. Although many researchers
have studied detecting drowsiness at the wheel, the methodology of mitigating drowsiness
was not major area of study. In this research, we investigated the relationship between sound
pressure levels of music and driver’s drowsiness variation by using the electroencephalogram
(EEG), the electrocardiogram (ECG) and using images of the driver’s face taken with a video
camera. Evaluation was done by introducing the method of rating the subjective drowsiness
defined by NEDO (The New Energy and Industrial Technology Development Organization)
in Japan. The analytical results of data of gathered from EEG and ECG showed that the
autonomic nerve was activated when the subjects were listening to louder music under resting
state. The results of the subjective drowsiness by appearance showed that duration time inbetween
awake and drowsy state of the subjects was extended by a certain level while
listening to relatively loud music when operating the driving simulator. We ascertained that
high pressure sound level of music gave a certain effect on subjects to suppress increasing
their drowsiness and to set back initiation time of drowsiness.

Aichi Prefectural University

TOYOTA MOTOR CORPORATION

Presented at the 15th World Congress on Intelligent Transport Systems, November 16-20, 2008, New York, New York

To prevent accidents caused by human error, the authors aimed to develop a doze
prevention system. This paper presents a system built to estimate the driver
drowsiness state based on an individual drive behavior model. Using the driver’s
behavior data recorded from a driving simulator, authors constructed an individual
drive behavior model for every examinee using a Gaussian mixture model (GMM).
Based on the individual drive behavior model, the driver drowsiness state was
estimated. Only an awake model was used at first, and the result was not satisfactory
(at a precision of 71.3%, the recall was 60.5%). Three models were made next to
estimate the driver drowsiness state. As a result, when the precision of the
drowsiness state was 67.7%, the recall of the drowsiness state increased up to 94.5%.

EQUOS RESEARCH CO., LTD.

AISIN AW CO., LTD

NAGOYA UNIVERSITY

Presented at the 15th World Congress on Intelligent Transport Systems, November 16-20, 2008, New York, New York

Through the development of digital maps with updated speed limits and positioning
technology new possibilities are created for improving traffic safety and environment. One
example is incentives for safer driving behaviour through a new type of traffic insurance
schemes. This article argues that such a system would be a powerful tool to fight traffic
accidents. The system could be offered to drivers who voluntarily wish to obtain a
substantially lower premium when speed limits are followed to a level agreed upon with the
insurance company. Research show, that these kinds of incentives have a strong positive
impact on the adherence to speed limits and following also on the risk of accidents. The costs
for the society due to traffic accidents are high and societal benefits could be realised through
intelligent traffic insurances. This calls for a political initiative and a complementary or
alternative institutional framework.

VINNOVA

Presented at the 15th World Congress on Intelligent Transport Systems, November 16-20, 2008, New York, New York