JP4950587B2 - Traffic information providing apparatus, traffic information providing system, anomaly data detection method and data collection method - Google Patents

Description

  The present invention relates to a technology of a traffic information providing apparatus, a traffic information providing system, an abnormal data detection method, and a data collection method for obtaining traffic information from road sensors and vehicles and providing traffic information to an in-vehicle terminal.

VICS (Vehicle Information and Communication System: registered trademark), which currently provides road traffic information services for car navigation systems, collects road traffic information from road sensors installed on the road and edits it. Through FM (Frequency Modulation) multiplex broadcasting and beacons, traffic congestion information such as traffic information and travel time information representing travel time is provided.
However, VICS (registered trademark) has the disadvantage that infrastructure is expensive, such as installation of road sensors, and that abnormal data is sometimes included due to failure of the road sensors or inadequate adjustment.

There is also known a probe information collection system that acquires road traffic information from a traveling vehicle (probe car). In the probe information collection system, since the vehicle itself is a sensor, it is possible to collect traffic information that does not depend on road infrastructure like VICS (registered trademark), and in principle it is possible to provide traffic information on all roads. is there.
However, since the traveling position and timing of the vehicle, which is a sensor, is probabilistic, its information quality differs from the continuous data collected by the road sensors described above, and is lost in space and time. .

Therefore, when the data acquired by the road sensor and the data acquired by the probe car are both present, the road traffic information service described above is used to select and adopt highly reliable data. It was the most important design item for the traffic information provision system.
By the way, conventionally, there is a method of distributing traffic information using data acquired by a probe car at a place where there is an abnormality in sensor information acquired by a road sensor, and there have been many traffic information providing devices for that purpose. Proposed.

For example, a technique is known in which a degree of change specific to a road section is calculated from sensor information detected by a road sensor, and based on the degree of change, a valid period of probe information collected by a probe car is dynamically determined. (For example, refer to Patent Document 1).
Japanese Patent Laying-Open No. 2004-286533 (paragraphs “0011” to “0013”, FIG. 1)

However, the technique disclosed in Patent Document 1 described above is a technique for determining the validity period of probe information for which utilization of probe information is permitted. Data collected by road sensors and probe cars are collected. It is not a technique for judging the deviation from the data.
Therefore, there is no description about how highly reliable data can be selected and used when both the data acquired by the road sensor and the data acquired by the probe car are present.

  The present invention has been made based on the circumstances described above, and when an abnormality is included in the data acquired by the road sensor, the link in which the abnormality is detected is regarded as an object to be uploaded with priority given to probe data. It is an object of the present invention to provide a traffic information providing apparatus, a traffic information providing system, an abnormal data detection method, and a data collection method that provide a highly reliable traffic information providing service.

  In order to solve the above-described problem, the traffic information providing apparatus of the present invention filters reference data generated by filtering link data acquired from a road sensor with normal link data and analyzing the filtered sensor data as a principal component. When the content rate of the reference pattern generated by calculating the inner product of the current link data acquired from the road sensor and the current link data is less than the predetermined value, the link data acquired from the road sensor is regarded as abnormal and the probe data has priority. It was set as the structure output as a link. The in-vehicle terminal is configured to upload the probe data of the link to the traffic information providing apparatus and preferentially display the probe data of the link on the screen.

  According to the present invention, when an abnormality is included in the data acquired by the road sensor, the link in which the abnormality is detected is regarded as a link to be uploaded with priority given to probe data. Traffic information can be provided using only high link data.

(First embodiment)
FIG. 1 is a block diagram illustrating an internal configuration of the traffic information providing apparatus according to the first embodiment.
The traffic information providing apparatus 1 includes a road sensor normal link list (normal link list) 11, a filtering calculation unit 12, a normal data history DB 13, a principal component analysis calculation unit 14, a reference pattern DB 15, a data history DB 16, The reference pattern content rate calculation unit 17 and the probe data priority link list 18 are configured. The filtering calculation unit 12 and the data history DB 16 of the traffic information providing apparatus 1 are connected to the road sensor 2 via a communication infrastructure. The road sensor 2 is installed on the road, and is provided to a traffic information center (traffic information providing device) via VICS (registered trademark) or the like.

The road sensor normal link list 11 has already been acquired from the road sensor, for example, based on an empirical rule reflecting the day type (day of week, fifty days, weekdays / holidays, large holidays, school holidays, etc.) A list of link numbers of links determined to be artificially normal is stored. That is, the links listed in the road sensor normal link list 11 are determined in advance to be empirically normal.
The road sensor normal link list is generated manually or by scoring by inner product with reference data simulating morning traffic jam, evening traffic jam, and the like. The manual method is a more realistic method because it can create a reference pattern that more closely conforms to actual measurement, and it is only necessary to extract representative ones. Details regarding the method of generating by scoring by the inner product with the reference data will be described later. The road sensor normal link list 11 is simply a list of link numbers such as normal link number 1, normal link number 2,..., Normal link number n. However, this is not limited to nationwide, for example, divided into meshes, mesh number 1; normal link number 1, normal link number 2,..., Mesh number 2; normal link number 1, normal link number 2,. The structure is also conceivable.
The filtering calculation unit 12 acquires road sensor data (first link data) including a link number, a travel time, and traffic volume data indicating the degree of traffic congestion transmitted from the road sensor 2 via the communication infrastructure. Then, only the road sensor data corresponding to the link number stored in the road sensor normal link list 11 (verified to be normal) is passed and output to the normal data history DB 13. The normal data history DB 13 stores the road sensor data as normal road sensor data (second link data). That is, normal road sensor data is stored in the normal data history DB 13 as a set having a corresponding link number and traffic data. The road sensor data transmitted from the road sensor 2 via the communication infrastructure is accumulated in the data history DB 16 in a time series. This time series data length is accumulated over a span of several days to several months.

The principal component analysis calculation unit 14 reads out a predetermined number of time-sequential normal road sensor data stored in the normal data history DB 13, and performs principal component analysis on the traffic data included in the read road sensor data. To get the base. Here, the basis obtained by principal component analysis is a component that changes with correlation among a plurality of variables (traffic volume for each link), and when expressed as a vector (basis vector), a plurality of basis vectors. The space spanned by (the space with the base vector as the coordinate axis) is the feature space. For example, if there is time-series data composed of 100 variables, the basis is how the 100 variables change with correlation as time passes.
For example, if variable 1, variable 2, variable 3,..., Variable 100 changes at a ratio of 0.1, 0.5, 0.3,. There will be a basis vector having 100 elements 0.3,..., 1, 2]. Since the correlation component included in the 100 variables is not the only one, there are a plurality of bases.

The principal component analysis method described above is disclosed in, for example, Japanese Patent Application Laid-Open No. 2006-85511 (FIGS. 5 to 9) , Information Processing Society of Japan research report “Advanced Traffic System”, Masatoshi Kumagai, No. 20, disclosed in detail in March 2005.

Returning to FIG. The basis of the traffic volume data generated by the principal component analysis processing by the principal component analysis calculation unit 14 is stored in the reference pattern DB 15 as a reference pattern indicating normal road sensor data.
The reference pattern DB 15 stores reference patterns normalized for each pattern. The data length of each reference pattern is, for example, one day. For example, assuming data at intervals of 5 minutes, the number of blocks is 288 blocks in 24 hours. In addition, the data history DB 16 stores road sensor data, that is, a link number and a traffic volume data as a set, similar to the normal data history DB 13 described above.

On the other hand, the reference pattern content rate calculation unit 17 reads a predetermined number of road sensor data accumulated in the data history DB 16, and the reference pattern content rate (reference pattern content rate) read from the reference pattern DB 15 with respect to this road sensor data. ). Here, the content rate is a ratio of the reference pattern included in the traffic data of the road sensor data.
The reference pattern content rate calculation unit 17 compares the content rate of the output reference pattern with a preset determination threshold value, and stores the result in the probe data priority link list 18. Here, the road sensor data (traffic volume data) having a content rate lower than the determination threshold is used as a link for preferentially uploading probe data acquired from the in-vehicle terminal 5 as abnormal road sensor data. . The probe data priority link list 18 is a list in which, for example, probe data priority link number 1, probe data priority link number 2,... In addition, when setting the reliability of the road sensor data, reliability data represented by 0.0 to 1.0 is assigned to each link number, for example. Details will be described later.

Next, the basic principle of the abnormal road sensor data detection method will be described with reference to FIGS. 2 and 3 with reference to FIG.
2 and 3 are diagrams showing the basic principle of the abnormal road sensor data detection method in the traffic information providing system in the first embodiment. Specifically, the reference pattern output by the pattern content rate calculation unit 17 is shown in FIG. The algorithm for content rate calculation of is shown.
2 and 3, the horizontal axis represents time, and the vertical axis represents travel time.

FIG. 2 is a diagram illustrating the basic principle of the abnormal road sensor data detection method in the traffic information providing system according to the first embodiment, and FIG. 2A is a diagram for explaining normalized traffic data. b) is a diagram illustrating a reference pattern.
The normalized traffic data in FIG. 2 (a) shows the transition of travel time (traffic situation) according to time on a certain road. Here, the time zone of approximately 6:00 to 9:00 is a time zone generally referred to as a morning peak, since traffic conditions vary greatly. In addition, there are some fluctuations in the evening from 17:00 to 19:00, and the traffic situation is stable after 19:00. In other words, the travel time is short and the road is stable.
In addition, the reference pattern of FIG. 2B filters the traffic data acquired from the road sensor 2 based on normal road sensor data verified to be normal by the filtering calculation unit 12, and results of the filtering. This is a reference pattern obtained by subjecting the traffic data of the output normal road sensor data to principal component analysis by the principal component analysis calculation unit. Here, a reference pattern representing a morning peak is shown.

Here, if the traffic volume data stored in the reference pattern DB 15 or the data history DB 16 is expressed as a vector and the traffic volume data is acquired at intervals of 5 minutes for 24 hours, the data of 288-dimensional vectors are time-series respectively. Will be stored.
Here, when the component expressing the reference pattern as a vector is Ri and the component expressing the traffic data of the road sensor data as vector is Xi (i = 1 to 288), the traffic data of the reference pattern and the road sensor data. In order to determine the similarity to, an inner product operation of two vectors is performed, and an inner product score (X1 × R1 + X2 × R2 +... + X288 × R288) is calculated.

FIG. 3 is a diagram showing the basic principle of the abnormal data detection method in the traffic information providing system according to the first embodiment. (A) shows a morning peak as a reference pattern, and (b) and (c) Both represent abnormal traffic congestion at night.
3A to 3C show an example in which the traffic data is shown on the left, the reference pattern is shown on the right, and the similarity is determined based on the size of the inner product score. (A) represents a morning peak as a reference pattern, and (b) and (c) both represent an abnormal traffic jam at night.
Since the result of the inner product calculation is a dimensionless number, it cannot be associated with a concrete expression such as travel time or congestion, but here we normalize the reference pattern and the traffic data of road sensor data. The inner product is in the range of -1.0 to 1.0, 0.5 is considered as the similarity determination threshold, less than 0.5 is determined as no similarity, and it is determined that the sensor data is abnormal road.

Returning to FIG. The traffic information providing apparatus 1 described above is installed in a traffic information center, and is connected to an in-vehicle terminal 5 and a probe car 3 that provide a traffic information providing service via a mobile communication network 4.
The in-vehicle terminal 5 always obtains road sensor data from the road sensor 2 and probe data from the probe car via the communication infrastructure. In addition, the traffic information provided by the traffic information providing device 1 is displayed on a display device (not shown) via the mobile communication network 4. Then, on the link determined to be abnormal based on the same acquired probe data priority link list, the probe data is preferentially displayed. The probe data is actively acquired and uploaded to the traffic information providing apparatus 1 via the mobile communication network 4.

Next, the operation of the traffic information providing apparatus 1 according to the first embodiment will be described along FIG. 4 with reference to FIG.
FIG. 4 is a flowchart showing the operation of the traffic information providing apparatus according to the first embodiment, in which (a) is a reference pattern generation process performed by a filtering operation unit and a principal component analysis operation unit, and (b) is a reference operation. The calculation process of the content rate of the reference | standard pattern by a pattern content rate calculation part is shown.
Hereinafter, the operation of the traffic information providing apparatus according to the first embodiment shown in FIG. 1 will be described in detail with reference to FIGS. 4 (a) and 4 (b).

First, the filtering calculation unit 12 acquires the road sensor normal link list 11 from the normal link data stored in the road sensor normal link list 11, searches the road sensor normal link list 11 (step S201), and detects the road sensor. By referring to the list of link numbers stored in the normal link list 11 and the link number of road sensor data (first link data) acquired from the road sensor 2, the link to which the road sensor data is sent can be obtained. When filtering is performed by determining whether or not the link (normal link) registered in the road sensor normal link list 11 is matched (step S202), the filter is matched with the normal link (step S202 "Yes"). ), The road sensor data is a pair of link number and traffic volume data. As a normal road sensor data is data (second link data) accumulated in the normal data history DB 13 (step S203).
When it is determined in step S202 that the link does not match the normal link (step S202 “No”), the road sensor data is not used, and the traffic information data corresponding to the link is collected by the probe car. Provide traffic information using probe data.
Subsequently, the principal component analysis calculation unit 14 reads out a predetermined number of normal road sensor data from the normal data history DB 13 and performs principal component analysis on the traffic data included in the normal road sensor data (step S204). ) And the basis and feature amount of the traffic data generated as a result are stored in the reference pattern DB 15 as a reference pattern.

  On the other hand, road sensor data (third link data) acquired from the road sensor 2 is always accumulated in the data history DB 16 via the communication infrastructure (step S211). For this reason, the reference pattern content rate calculation unit 17 reads out the reference pattern stored in the reference pattern DB 15 and the road sensor data stored in the data history DB 16 and executes normalization processing, and then performs the inner product calculation shown in FIG. Is executed to calculate the reference pattern content (reference pattern content) (step S212). If the road sensor data and the reference pattern are normalized, the inner product is in the range of -1.0 to 1.0. Therefore, here, it is synonymous with obtaining the content rate of the reference pattern.

Subsequently, the reference pattern content rate calculation unit 17 compares the calculated inner product with a determination threshold value to determine whether or not the inner product with the reference pattern is equal to or less than the determination threshold value (step S213). When the inner product value is equal to or smaller than the determination threshold value (step S213 “Yes”), the road sensor data is regarded as abnormal, and is registered in the probe data priority link list 18 as a probe data priority link. Moreover, the probe data priority link list | wrist 18 is transmitted with respect to the vehicle-mounted terminal 5 at the timing which received the request | requirement of traffic information from the vehicle-mounted terminal 5 (step S214).
In the process of step S213, if it is determined that the inner product with the reference pattern exceeds the determination threshold value (step S213 "No"), the road sensor data is determined to be normal and corresponds to the link. Traffic information is provided to the in-vehicle terminal 5 using road sensor data.
Here, the probe data priority link is a link that should be preferentially uploaded and acquired from the probe car.

FIG. 5 is a diagram illustrating a system configuration of the traffic information providing system according to the first embodiment.
As shown in FIG. 5, the traffic information providing apparatus 1 is further connected to a probe data DB 20 in the configuration of the traffic information providing apparatus shown in FIG. 1. Here, the road sensor normal link list 11, the filtering operation unit 12, the normal data history DB 13, and the data history DB 16 are not shown.

The in-vehicle terminal 5 is connected to the traffic information providing apparatus 1 through the mobile communication network 4 and acquires the current position information of the vehicle by the in-vehicle sensor 8 such as GPS (Global Positioning System) and The link number included in the acquired road sensor data matches the link number listed in the probe data priority link list 18 transmitted from the traffic information providing apparatus 1 via the mobile communication network 4. In addition to displaying the acquired probe data with priority, the probe data is transmitted to the traffic information providing apparatus 1. The traffic information providing apparatus 1 accumulates the received probe data in the probe data DB 20.
As a result, the traffic information providing apparatus 1 can provide road information to the in-vehicle terminal 5 by using road sensor data and probe data in combination and using only highly reliable data. Moreover, the vehicle-mounted terminal 5 can display the road sensor data and probe data with high reliability in combination.

FIG. 6 is a flowchart showing the operation of the in-vehicle terminal.
In FIG. 6, the in-vehicle terminal 5 always receives the probe data and the road sensor data from the probe car and the road sensor, while receiving the probe data priority link list 18 from the traffic information providing apparatus 1. (Step S300). Therefore, the probe data priority link list 18 received from the traffic information providing device 1 is first searched (step S301), and whether or not the acquired road sensor data is abnormal, that is, the ring number included in the road sensor data is determined. Then, it is determined whether or not the link numbers listed in the probe data priority link list 18 are matched (step S302).
If they match (step S302 “Yes”), the acquired probe data is preferentially displayed on a screen (not shown), and the probe data is uploaded to the traffic information providing apparatus 1 (step S303). If they do not match (step S302 “No”), the road sensor data is displayed on a screen (not shown) (step S304).

An example of a screen configuration displayed on the in-vehicle terminal 5 will be described along FIG. 7 with reference to FIG.
FIG. 7 is a diagram illustrating an example of a screen configuration displayed on the in-vehicle terminal. Here, the display data reliability setting window 10 can be displayed as well as simply displaying probe data and road sensor data in combination with map information.
The reliability setting window 10 is a mechanism for adjusting a determination threshold when the traffic information providing apparatus 1 performs a similarity determination using a GUI (Graphical User Interface) of the in-vehicle terminal 5.

That is, on the user side, there is a request to display road sensor data acquired from the road sensor 2 even if it is uncertain, or to display only reliable link data (probe data), depending on the usage environment of the in-vehicle terminal 5.
At this time, the user displays the reliability setting window 10 of the display data by the GUI on the display device screen of the in-vehicle terminal 5 using the GUI, and slides the slide bar on the display screen in the horizontal direction, for example. Can be adjusted. This information is transmitted to the traffic information providing apparatus 1 via the mobile communication network 4, and the traffic information providing apparatus 1 uses the determination threshold value used when determining the similarity between the reference pattern and the road sensor data based on this information. To control. As described above, the reliability for each link calculated by the traffic information providing apparatus 1 is stored in the probe data priority link list 18, and the in-vehicle terminal 5 that has received the probe data priority link list 18 is designated by the slider. The road sensor data of the link having the reliability equal to or higher than the threshold value is displayed as reliable. The reliability for each link is obtained by the traffic information providing apparatus 1 calculating an average inner product value, or by multiplying this average inner product value by some coefficient, and distributing it to the in-vehicle terminal 5 as information for each link. To do. Here, as described above, the average inner product value performs the inner product calculation for the reference pattern and the normalized road sensor data. This calculation is repeatedly performed during the period for evaluating the road sensor data in units of one day. The average of the inner product values calculated for each day during the evaluation period.

According to the first embodiment described above, when the road sensor data includes an abnormality, the link in which the abnormal road sensor data is detected is set as a target for uploading the probe data with priority. Providing high traffic information. Further, the in-vehicle terminal 5 can collect the probe data by effectively using the band, and can display the VICS (registered trademark) and the probe data together.
Furthermore, it is possible to update the display range of probe data or road sensor data by providing the user with a determination threshold setting mechanism using a GUI, thereby providing convenience to the user and providing flexibility and flexibility. A high traffic information provision system can be constructed.

(Second Embodiment)
FIG. 8 is a block diagram showing an internal configuration of the traffic information providing apparatus according to the second embodiment.
The traffic information providing apparatus 1 ′ according to the second embodiment is different from the first embodiment shown in FIG. 1 in that a GUI providing unit 19 is added. The GUI providing unit 19 takes in a reference pattern that is artificially set and input via the center terminal 7 and calculates a reference pattern content rate based on the reference pattern. For this reason, the center terminal 7 is connected to the traffic information providing apparatus 1 ′. Others are the same as those of the first embodiment shown in FIG.

  FIG. 9 is a diagram showing an example of the display screen configuration of the center terminal. (A) is a result of calculation of the reference pattern content rate based on the reference pattern shown in (b), resulting in congestion in the morning and evening. This is an example in which a road link is detected and displayed, and FIG.

The actual screen operation at the center terminal 7 will be described with reference to FIG.
As shown in FIG. 9B, the reference pattern content rate calculation unit 17 reads a reference pattern registered in a file in a storage device (not shown) in advance through a screen provided by the GUI providing unit 19 or is artificially generated. On the basis of the reference pattern drawn in the above, the content rate of the reference pattern with respect to the road sensor data stored in the data history DB 16 is calculated to calculate the similarity. The result is displayed at the lower right of the screen by a color bar representing the level of similarity.
For example, an artificially drawn reference pattern (handwritten input pattern) is replaced with coordinate data on the screen inside the center terminal 7, and travel time series data (speed, traffic volume, Traffic jam length). The travel time series data converted here is handled in the same manner as the reference pattern (travel time series data) stored in the reference pattern DB 15 described above. Then, as shown in FIG. 2, by calculating the inner product of the travel time series data converted as described above and the normalized road sensor data, the reference pattern content rate is calculated, and the calculation result is obtained. A color bar display indicating the degree of similarity is performed according to the reference pattern content ratio.
Here, not only the traffic information of the entire area is simply displayed, but the road links that are congested in the morning and evening are displayed in a filtered manner in distinction from other roads. The traffic information providing apparatus 1 described above can be applied as a road sensor 2 or a probe data monitoring system in a traffic information center.

In the first embodiment and the second embodiment described above, road sensor data acquired from the road sensor 2, probe data acquired from the probe car 3, and probe data priority link list distributed by the traffic information providing device 1 Although 18 is all link data, it is expressed by being distinguished by its origin.
Further, in the first embodiment and the second embodiment described above, the filtering calculation unit 12, the principal component analysis calculation unit 14, and the reference pattern content rate calculation unit shown in FIG. 1, FIG. 5, and FIG. 17, the probe data priority link list 18, and the functions of the GUI providing unit 19 are programmed and stored in a storage device (not shown), and the computer sequentially reads and executes the program stored in the storage device (not shown). It goes without saying that the same effect can be obtained by this. At this time, the road sensor normal link list 11, the normal data history DB 13, the reference pattern DB 15, the data history DB 16, and the probe data priority link list 18 are allocated and stored in the storage device.

It is a block diagram which shows the internal structure of the traffic information provision apparatus which concerns on 1st Embodiment. It is a figure which shows the basic principle of the abnormal road sensor data detection method in the traffic information provision system in 1st Embodiment, (a) is a figure explaining the normalized traffic data, (b) It is a figure explaining a reference pattern. It is a figure which shows the basic principle of the abnormality data detection method in the traffic information provision system in 1st Embodiment, (a) represents a morning peak as a reference pattern, (b), (c) is both night abnormality It represents a traffic jam. It is a flowchart which shows operation | movement of the traffic information provision apparatus which concerns on 1st Embodiment, (a) is a reference pattern production | generation process by a filtering calculating part and a principal component analysis calculating part, (b) is reference pattern content rate calculation. The calculation process of the content rate of the reference | standard pattern by a part is shown. It is a figure which shows the system configuration | structure of the traffic information provision system which concerns on 1st Embodiment. It is a flowchart which shows operation | movement of a vehicle-mounted terminal. It is a figure which shows an example of the screen structure displayed on a vehicle-mounted terminal. It is a block diagram which shows the internal structure of the traffic information provision apparatus which concerns on 2nd Embodiment. It is a figure which shows an example of the display screen structure of a center terminal, (a) is the road link which is crowded in the morning and evening as a result of calculating a reference pattern content rate based on the reference pattern shown in (b) This is an example in which is detected and displayed, and (b) is a diagram showing generation of a reference pattern.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,1 'Traffic information provision apparatus 2 Road sensor 3 Probe car 4 Mobile communication network 5 Car-mounted terminal 7 Center terminal 8 Car-mounted sensor 10 Reliability setting window 11 Road sensor normal link list 12 Filtering calculation part 13 Normal data history DB
14 Principal component analysis calculation unit 15 Reference pattern DB
16 Data history DB
17 Reference pattern content rate calculation unit 18 Probe data priority link list 19 GUI providing unit 20 Probe data DB

Claims (8)

The first link data obtained from a road sensor installed on the road , which is information of a set of a link number and traffic data, is input, and the input first link data may be normal by the user. A filtering operation unit that outputs only the first link data registered in the normal link list as the second link data by comparing with the normal link list that lists the link numbers of the verified links ;
A principal component analysis calculation unit for generating a reference pattern which the base and element and principal component analysis a predetermined number of the second link data output by the filtering operation unit,
The is predetermined number acquired from the road sensor, the reference pattern generated by the third link data Contact and the principal component analysis arithmetic unit is a set of information between the link number and traffic amount data, respectively normalized, the when the third link data after normalized, the content of is that standards pattern obtained with the reference pattern and the inner product calculation after being the normalization is less than a predetermined value, the road sensor the third link data regarded as abnormal, the appropriate link to the third link data deemed the abnormality, preferentially up Rhodes should probe data from the probe cars to the traffic information providing device itself obtained from and the link is a probe data priority linked, the reference pattern content calculation unit for outputting information relating to the probe data priority link,
Traffic information providing apparatus characterized by Ru with a have. The reference pattern content rate calculation unit is:
Claim 1, characterized in that said by the predetermined value determined by the number of bases is generated and output by the principal component analysis calculation unit, abnormality determination of said third link data acquired from the road sensor The traffic information providing device described in 1. Content of the reference pattern by artificial setting captures the input, the reference pattern content rate calculating unit by using the reference pattern generated by the principal component analysis calculation unit on the basis of the captured set input of the reference pattern A user interface part that executes the calculation of
Traffic information providing device according to claim 1, characterized in that Ru provided with have. The first link data obtained from a road sensor installed on the road , which is information of a set of a link number and traffic data, is input, and the input first link data may be normal by the user. A filtering operation unit that outputs only the first link data registered in the normal link list as second link data by comparing the link numbers of verified links with the normal link list, and the filtering operation principal component analysis calculation unit a predetermined number of the second link data outputted by principal component analysis to generate a reference pattern that the base and element by parts, and obtained a predetermined number from the road sensor, the link number it and the reference pattern generated by the third link data and the principal component analysis arithmetic unit is a set of information with traffic data Normalized, and the third link data after the normalization, when the content of is that standards pattern obtained with the reference pattern and the inner product calculation after being the normalization is less than a predetermined value the regarded third link data acquired from the road sensor abnormal, the appropriate link to the third link data deemed the abnormality, preferentially upload from probe cars to the traffic information providing device itself a probe data preferred link is a link probe data to a traffic information providing device having a reference pattern content calculation unit for transmitting information relating to the probe data priority link to the vehicle-mounted terminal,
Wherein the traffic information providing device connected via a mobile communication network, road sensor data from the path sensor, and the acquired probe data from the probe cars, the obtained road sensor data, via the mobile communication network When it corresponds to a probe data priority link transmitted from the traffic information providing device, the acquired probe data is displayed with priority, and an in- vehicle terminal that uploads the probe data to the traffic information providing device ;
Traffic information providing system characterized in that Ru provided with have. The in-vehicle terminal is
5. The acquired probe data is uploaded to the traffic information providing apparatus via the mobile communication network when the acquired road sensor data corresponds to the probe data priority link. Traffic information provision system. The in-vehicle terminal is
In order to change the predetermined value, the traffic information is acquired based on the predetermined value by capturing the predetermined value via a user interface unit and transmitting the predetermined value to the traffic information providing apparatus via the mobile communication network. traffic information providing system according to claim 4 for receiving the probe data priority links generated by the providing apparatus, to display the probe data of the probe data priority link, characterized by. A method for detecting abnormal data in a traffic information providing system that obtains traffic information from road sensors and vehicles installed on the road and provides traffic information to an in-vehicle terminal,
Traffic information providing device
The first link data, which is a set of link number and traffic data, obtained from the road sensor is used as an input, and the input first link data is verified to be normal by the user. By comparing with the normal link list in which the link numbers are listed, filtering is performed to output only the first link data registered in the normal link list as the second link data ,
Generates a reference pattern that the base and element a predetermined number of the second link data resulting from the said filtering by principal component analysis,
The is predetermined number acquired from the road sensor, turned into a regular and a third link data with the generated reference pattern is a set of information between the link number and traffic amount data, respectively,
A third link data after the normalization, the a reference pattern after being normalized by the inner product computation to calculate the content of the standards pattern,
When the content rate of the reference pattern is less than a predetermined value, the third link data acquired from the road sensor is regarded as abnormal, and a link corresponding to the third link data regarded as abnormal is probed. It is registered as a probe data priority link that is a link of probe data to be preferentially uploaded from the car to the traffic information providing device itself, and information related to the probe data priority link is transmitted to the in-vehicle terminal . For detecting abnormal data in a traffic information providing system. The traffic information providing apparatus and the traffic information providing apparatus have an in-vehicle terminal connected via a mobile communication network, and the traffic information providing apparatus acquires traffic information from a road sensor and a vehicle installed on the road. And a data collection method in a traffic information providing system for providing traffic information to the in-vehicle terminal,
The traffic information providing device includes:
The first link data, which is a set of link number and traffic data, obtained from the road sensor is used as an input, and the input first link data is verified to be normal by the user. By comparing with the normal link list in which the link numbers are listed, filtering is performed to output only the first link data registered in the normal link list as the second link data ,
Generates a reference pattern that the base and element a predetermined number of the second link data resulting from the said filtering by principal component analysis,
The is predetermined number acquired from the road sensor, turned into a regular and a third link data with the generated reference pattern is a set of information between the link number and traffic amount data, respectively,
A third link data after the normalization with the reference pattern after being the normalization, and inner product operation to calculate the content of the standards pattern,
When the content rate of the reference pattern is less than a predetermined value, the third link data acquired from the road sensor is regarded as abnormal, and a link corresponding to the third link data regarded as abnormal is probed. While registering as a probe data priority link that is a link of probe data to be preferentially uploaded from the car to the traffic information providing device itself, and transmitting information related to the probe data priority link to the in- vehicle terminal,
The in-vehicle terminal is
Obtain road sensor data from the road sensor, obtain probe data from a probe car, and obtain information on the probe data priority link from the traffic information providing device ,
When the acquired road sensor data corresponds to the probe data priority link , the acquired probe data is displayed with priority, and the probe data is uploaded to the traffic information providing device . Data collection method in providing system.

Images