<Abstract>
Today I would like to introduce a research and development project, which is currently in progress and being carried out at the Yokosuka ITS Research Center of the National Institute of Information and Communications Technology (NICT). Let me briefly introduce an outline of this project before taking up the main subject. One of the core technologies of the system we are currently trying to develop at NICT is an active electronic tag. When a pedestrian or a cyclist with this RFID tag enters the perimeter of a position-detection area in which a Low Frequency (LF) band set over the street is used, the LF signal activates the tag and the tag receives location-specific information from the location marker. The RFID tag adds the attribute information of the tag holder to the location information, and then sends the information to approaching vehicles using UHF waves from the RFID tag. This will enable drivers of approaching vehicles to obtain detailed location information of pedestrians, cyclists, or anyone else through the tag before they come into the driver's field of view, especially at intersections. Drivers can also use the information pertaining to his/her own vehicle. With all of the information on the pedestrian and the vehicle, the driver can prepare for the possibility of, for example, someone rushing into the street which looks like it may occur in the next few seconds, and take action to avoid an accident. We are expecting that this will result in a reduction of traffic accidents involving vulnerable road users such as pedestrians and cyclists. The research and development project, which is currently being carried out at the NICT, is to establish a demonstration model related to the Safe Driving Support System.
First, let me inform you of the present status of traffic accidents involving vulnerable road users. According to the 2006 white paper on traffic safety in Japan, the total number of accidents, fatalities, and injuries has, unfortunately, remained the same. The total number of fatalities has decreased, but the number of fatalities involving elderly people (over 65 of age) has remained the same, which contrasts starkly with that of youth which has dramatically decreased. Thus, to reduce traffic accidents as a whole, it is necessary to take measures to protect elderly people from traffic accidents.
Accidents that involved pedestrians and cyclists account for about 25% of total traffic accidents. The number of fatalities in accidents involving pedestrians and cyclists also account for about 40% of the total fatalities in traffic accidents. Furthermore, more than 80% of accidents involving pedestrians and cyclists are due to the delay in locating them. Combined with error in judgment, the percentage goes up to 98%. That is why we believe that traffic accidents can be reduced by letting drivers know of approaching danger in advance.
In accordance with the report on traffic accidents I have just mentioned, the NICT has been promoting a research and development project on applied ITS technology using RFID tags. As I just said a moment ago, more than 80% of accidents involving pedestrians and cyclists are due to the delay in finding them. In general, this type of accident often occurs because some unexpected object suddenly comes into sight from outside of the driver's field of view. One of the measures against this type of accident is to let drivers know the location of vulnerable road users in advance. In other words, a system that can provide information on pedestrians and cyclists to drivers before they come into the drivers' sight is one of the solutions.
The goal of this research and development is to accurately detect the location of tags by combining 1 mW/950 MHz band active electronic tag (RFID) technology and 125 kHz LF signal device technology that acts as a location marker, which will enable the tag detection area to be set flexibly. The detected data will be sent to vehicular devices via roadside communication devices. Upon receiving the information, the vehicular device judges the possibility of an accident using the tag speed and direction, and let's the driver know of that possibility. In doing so, we are aiming for the development of a system that will help reduce common traffic accidents involving vulnerable road users, for example, head-on collisions and collisions with pedestrians or cyclists when a vehicle makes a turn.
One of the technologies in progress that we are focusing on other than the core technologies I have already mentioned is a method to use roadside repeaters as a way of sending information from a tag to a vehicular device. As a means to complement this system, we are studying a method to send information from a tag to a repeater and from a vehicle to another vehicle (a vehicle-to-vehicle communication method) as well. Meanwhile, as research on the direction of the communication, we are mainly examining the transmission of pedestrian or cyclist information from an RFID tag to a vehicular device, but at the same time, we are also examining a 2-way function that sends information on approaching vehicles to an RFID tag.
We are also considering a way of detecting traveling speed and direction by analyzing the tag's travel history which includes how the tag traveled among several LF markers. If a vehicle can get the location information of vulnerable road users and then combine that information with its own traveling speed and direction, it will be possible to predict the possible risks and categorize them into risk types. In addition, we are studying relatively easy algorithms that can be used for predicting collisions or near-collisions based on the positional relationship between the tag and the vehicle.
In order to protect personal information, we are considering the preparation of variable IDs. There are two methods, a random number system and a preconfigured system. Since we are trying to detect tag traveling speed and direction using the tag and more than one LF marker, we are examining the idea of a system that does not allow IDs to change within a small area.
Our prototype device consists of a 950 Hz RFID tag, a 125 kHz LF signal generator, antenna coils that produce a magnetic field, a repeater, and a vehicular reader that relays the wireless information sent from the tag to the next reader. We use the same frequency for the repeater and vehicular reader, and both of their hardware configurations are nearly identical. These devices are connected to a car navigation device to display the information.
Since the results of our field experiment showed that the LF signal intensity would change depending on the angle of the tag, we have examined some solutions to that. We had significant reception intensity change when using a small coiled antenna as a LF signal reception antenna, but we found that we could ensure stable reception conditions by using three coiled antennas arranged in the three orthogonal directions.
For LF antennas that act as position markers, since they will be laid underground, we need to examine the influence from changes in the environment, for example, the influence of being placed underground, or from being exposed to rainwater. To address this issue, we have performed some experiments in which we set an LF antenna under some concrete blocks and in a water bowl. These experiments showed that the formation of the detection area that activates RFID tags is unaffected even if we change the thickness of the blocks or depth of the water.
In consideration of measures to prevent accidents caused by someone suddenly appearing from an unseen residential road onto a business road or bigger residential road, we are planning to perform some field experiments using LF antenna distribution models for pedestrians and cyclists. One of the key points of the experiment will be how to ensure the required time (3 seconds) and travel distance to avoid the accident for the driver when someone rushes onto the street, and how to transmit the information to him/her in advance. In addition to these experiments, we are also carrying out research for accidents at large blind intersections.
According to the required range of ASV communication to ensure safety, the detection distance between the RFID tag and the vehicle should be about 200 meters. We are also conducting an experiment on transmitting information from a tag to a vehicle more than 200 meters away.
We are also preparing a public experiment in the second half of 2007. In this experiment, we will use roads in the Yokosuka Research Park and aim at preventing traffic accidents based on the assumption that some pedestrians or cyclists will run out onto the street from a blind intersection. The challenges that remain for the practical application of this system include examining some other frequency bands that could increase transmission rates, how to spread the use of vehicular devices and RFID tags, how to establish the infrastructure, whether or not the system will be generally accepted by itself, collaborations with other systems, and its functional supplementation.