ATR Prevents Traffic Accidents and Uses Sensors to Improve Driving Skills

How can we reduce the number of fatalities in traffic accidents?

The representative answer is probably the vehicle system that can protect the occupant's airbag system at the time of collision, and the vehicle structure that can reduce the impact of the collision. The number of deaths in traffic accidents in Japan has been decreasing year by year, and it has been reduced to less than 5,000 in 2009. It is recognized as a contribution to the progress of the automotive safety system. In January 2006, the IT strategy headquarters of the then Prime Minister Junichiro Koizumi issued a declaration stating that he will aim at "the world's safest transportation society" and "reduce the number of traffic accident deaths to 5,000 in 2012. The following ", in fact, the goal has been achieved significantly ahead of schedule. In the future, with the advancement of preventive collision safety systems that can reduce the impact of accidents as a result of prior inspections, and the expansion of Intelligent Transportation Systems (ITS) that can prevent collisions and rear-end collisions, the number of traffic accident deaths is expected to decrease further.

But is there only a car safety system to reduce the number of fatalities in traffic accidents? In fact, the number of traffic accidents exceeds 700,000. Although it has decreased temporarily, it is still at a high level. The fatality rate (the number of deaths divided by the total number of casualties) was 0.54% and did not continue to decline in recent years. If you do not reduce the number of traffic accidents, it may be very difficult to drastically reduce the number of fatalities in traffic accidents. Although the popularization and expansion of ITS is expected to prevent traffic accidents, if the driver's driving skills are further improved, traffic accidents should be less.

Under such circumstances, the International Institute of Basic Telecommunications Technology (ATR) has carried out an activity to quantitatively evaluate the driver's technology using electronic technology, and to improve the driver's awareness of prevention safety through feedback evaluation results. The numerical values ​​include the certainty of the left and right when determining the right and left turns (angle and time of the neck), the speed at the time of turning, the timing when the accelerator pedal and the brake pedal are depressed, and the behavior of the driver when driving normally. And find the driver's driving habits from it, and make clear its deviation from the ideal driving. By making the driver aware of the results of the objective evaluation, the potential danger in driving "customs" is exposed, thereby eliminating the traffic accident in the "embryonic" state. Recently, the reporter has continuously expanded the scope of application since the practical use in 2008, and has been conducting interviews with ATR systems that collect practical verification data and explored its effect on preventing safety.

Using the sensor of the head and right foot to grasp the driver's condition The ATR system uses a sensor module equipped with a 3-axis acceleration sensor, a 3-axis gyro sensor, and a 3-axis geomagnetic sensor because of different performance specifications. The driver wears a hat fitted with a sensor module and the same module is mounted on the right foot (Figure 1). The sensor module of the hat is responsible for detecting data such as the angle of the face toward the left and right, and the sensor module of the right foot is for detecting the data that the foot steps on the accelerator pedal or the brake pedal. The external dimensions of the module are only 39mm x 44mm x 12mm and weigh only 20g, so it does not interfere with driving. In addition to the module, the system uses a GPS module that determines the position of the vehicle and collects small devices that hold the data of these modules. The data detected by the sensor module is transmitted via Bluetooth to the small device.

In August 2008, it was put into practical use. The training system for professional drivers was first started with the basic research of ATR as a prototype. ATR and Yamasaka Motorsport were jointly developed in 2008 (the business was carried out by ATR-Sensetech). The system was used as an “Objet system” that automatically evaluated the driver’s driving skills and provided suggestions for improvement. In August of the same year, the system was used in the training of professional drivers for the Yamashiro Motorsport.

The test data shows the right foot position (positioned on the accelerator pedal, brake pedal, or instability), the neck twist angle, and the vehicle speed, which are expressed as changes in time (Fig. 2). Allegedly using only three districts, it is possible to know whether the neck torsion angle during turning is sufficient to confirm the left and right, the confirmation time is sufficient, whether the vehicle deceleration is sufficient, and whether the timing of changing the accelerator pedal is appropriate. The scores (out of 100 points) or five grades are evaluated for the overall driving situation based on the evaluation of individual conditions such as straight ahead and right and left turns.

If the neck is less than 10 degrees, it is not enough to confirm the left and right sides. Even if the neck rotation is large, if the time around the confirmation is short, it is not sufficient (Figure 3). If the driver hesitates to "walk or stop", there will be data on the unstable position of the foot. Just looking at the waveform, not only the driving habits of the driver, but also “can't see the shadows, don’t pay too much attention when you turn,” and “the driver’s thoughts can also be seen” (ATR Intelligent Robotics Research Institute of Environmental Intelligence Nomae Chunsheng). Including the speed at the time of turning, and the degree of deceleration at the time when the visibility is poor, the data on individual differences can be known truthfully (Mountain Driving School refers to this speed varying from person to person as the individual's "holding speed").

In addition to the test data, the features seen from the test data will also be presented to the driver in a concise text (Figure 4). By numerically digitizing the driver's driving habits and mental state, the driver is presented with an easy-to-understand short comment so that the driver can easily notice his own danger. The test data evaluation method is to prepare a variety of driver safe driving behavior patterns as a reference, by comparing it with the detection data, extracting the characteristics of the driver. The short commentary can give a detailed evaluation according to the degree of danger of "slight attention", "attention", and "need to improve." Yamashiro Motors said that the driver has developed a safe driving method by using the Objet system to perform a driving evaluation every 3 to 6 months.

Draw attention to the driver's blind confidence in his “good technology” The application of the Objet system has been expanding since it became practical in 2008. The Objet system has evaluated more than 3,000 drivers.

The number of driving schools adopted has increased to 16 in 11 counties. A taxi company uses its training taxi drivers, and the Truck Association of Kyoto Prefecture and Nara Prefecture uses it to provide training services to member companies. The City Transportation Bureau and the Nagoya City Transportation Bureau also used the system in the training of bus drivers, and the JR West Japan Bus applied it to train long-distance bus drivers. In addition to training professional drivers, Kyoto Prefecture and Gifu Prefecture have identified the training using the Objet system as a certification program for the elderly.

It is said that the effect is very obvious. For example, after a taxi company used the Objet system to train drivers who had accidents, the number of drivers who received the second training (ie, those who received training again after an accident due to training) decreased by about 20%. This shows that for professional drivers with confident driving skills, when they point out their shortcomings, it is very effective to use such objective data as numerical values.

In terms of pointing out the shortcomings of confident drivers, it is equally effective in training for the elderly. ATR said that in the evaluation of his driving skills divided into 5 levels (5 is the best), the middle-aged is slightly lower than 4, while the elderly is slightly higher than 4. The Objet system's actual evaluation of 314 elderly people showed that in the five-grade evaluation, the proportion of the worst skills accounted for about 20% of the total, which was more than three times that of the average driver. ATR said that although elderly people are driving at a safe speed and are observing temporary parking, the timing for confirming safety when turning right or left is also accurate. However, the turning angle and confirmation time are not sufficient when turning left and right. This situation is clearly reflected in the test data. By showing them practical skills, many elderly people have discovered the gap between their actual driving skills and their subjective perceptions.

It can also be used to design safe roads. In addition to cars, the system can also be used to implement bicycle safety training and optimize the way ITS provides information based on numerical driver behavior. It can even be used to design roads that are not prone to traffic accidents.

The application in ITS has been verified using the Hanshin Expressway in the validation test of the road car coordination ITS information provision service “King Hanshin Area Intelligent Road Verification Experiment”. In the road car coordination service, the system can send information such as “traffic jam in front, attention to rear-end”, “merging vehicles on the right, and attention” according to road conditions, and the information is displayed on the car navigation system to remind the driver to pay attention. To prevent accidents. ATR and Osaka University conducted a detailed examination of the actions taken by the driver after obtaining the information in a verification experiment.

Subjects were 32 20-50 year old drivers who were openly solicited. They wore a hat with a sensor module installed and the same module was mounted on the right foot and investigated the behavior after seeing the traffic information. In the point where the side road vehicles merged into the main line from the right side, the experimental results showed that after the driver saw the message “There is a confluence vehicle on the right, please note”, 0.23 seconds before the throttle was throttled (average). With the action in the right direction. When the traffic jams in the current direction, the driver relaxes the accelerator pedal after an average of 1.77 seconds after seeing traffic jams. Based on these results, the ATR will investigate the influence of the timing, display position, and display contents of the sent information on the driver's sight line movement and pedal operation in the future, thereby making the ITS information providing service more effective.

Regarding road design, ATR teamed up with construction consultants engaged in road design. The Objet system was used to investigate whether the driver's hands, feet, and neck were adapted to road conditions such as road structure and traffic rules.

The verification was performed at the point where the upper road of the ordinary highway in Matsukachi, Mie Prefecture merged with the main line. The behavior of the vehicle where the driver wearing the Objet system merged from the side track to the trunk line was investigated. As a result, it was revealed that many drivers did not completely stop at the “temporary stop” place. Although there was an action to confirm the right side (main line direction), the throttle was The pedal is not released and the right foot moves from the brake pedal to the accelerator pedal sooner. Based on this result, it is possible to take measures such as widening the parking position, setting the road piles, adjusting the signs on the side road converging roads, etc. to solve the problem of incomplete temporary parking and insufficient left and right confirmation.

From "individual" to "social"

ATR plans to hold a seminar on October 28th, 2011 to further expand the application scope of the Objet system. It is hoped that the users and researchers of the Objet system will create a platform for communication and widely solicit comments to promote the development of the Objet system. ATR is studying strategies to promote the system to ordinary drivers in the future. For example, as the visibility of the Objet system increases, if insurance companies are interested, they will launch voluntary insurance linked to driving technology. In that case, the opportunity for ordinary drivers to use the Objet system is expected to increase.

Moreover, the evaluation system also needs to develop. Although the existing system is small in size, a sensor module must be installed every time driving. This inconvenience may deter users. This problem can be solved by using optical technology and camera technology to detect neck and right foot behavior. If a large number of vehicles equipped with an evaluation system are on the road, it is estimated that the driver's behavior and changes in road conditions can be read. Consideration will also be given to identifying on the map the use of roads that are difficult for the driver to drive and that may cause panic. “The Objet system has already done a fairly good evaluation for everyone. The next stage is to collect a large number of drivers and apply them to social services”