JP3253760B2 - Vehicle position detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle position detecting device for detecting a vehicle position on a map.

[0002]

2. Description of the Related Art FIG. 9 shows, for example, Japanese Patent Application Laid-Open No. 63-314411.
FIG. 1 is a block diagram showing a conventional vehicle position detecting device disclosed in Japanese Patent Application Laid-Open Publication No. H10-260, in which 11 is a distance sensor that detects the traveling distance of the vehicle, 12 is an azimuth sensor that detects the direction of the vehicle, and 13 is road network data The storage memory stores various data relating to each node and each traveling link in a road network in which a plurality of node points in a region including the traveling area of the vehicle and a traveling link that is a traveling route between each node point are set. I have.

[0003] Reference numeral 14 denotes a data processor, which comprises a traveling locus calculating means 15, a locus pattern comparing means 16, and a data transmitting means 17. The traveling locus calculation means 15 includes:
The traveling locus of the vehicle is calculated based on the detection data from the distance sensor 11 and the detection data from the direction sensor 12.
The trajectory pattern comparison means 16 compares the travel trajectory calculated by the travel trajectory calculation means 15 with the travel link read from the road network data storage memory 13 and selects a match or an approximate one. Data sending means 1
7 sends out position data representing the traveling link selected by the trajectory pattern comparing means 16.

A display unit 18 is connected to the data processor 14 and displays a road map in the traveling area of the vehicle on the basis of the road network data read from the road network data storage memory 13. The position of the vehicle based on the position data sent from 17 is displayed on the road map. Reference numeral 19 denotes a data storage memory, which stores the position data transmitted from the data transmission means 17 in a readable manner.

Next, the operation will be described. The pattern comparison between the traveling locus obtained by the traveling locus calculating means 15 of the data processor 14 and the traveling link is performed by the locus pattern comparing means 16. As a result of the pattern comparison, a match or an approximate one is selected, and the data sending means 1
7, the travel link and the position data selected by the pattern comparison are transmitted. This pattern comparison is performed at a timing such as when a change in the vehicle direction such as a right or left turn is detected at a certain angle or more, and the position of the vehicle is sequentially detected. The vehicle position detected in this way is supplied to the display unit 18 by the data transmission unit 17 and is also stored in a data storage memory 19 connected to the data processor 14 in a readable manner.

[0006]

Since the conventional position detecting device is configured as described above, there is a risk that an erroneous traveling link may be selected due to an error in local sensor data. In addition, storing a long-distance travel locus or storing map data over a wide range requires an enormous memory, so that the shape of the travel locus cannot be subjected to pattern comparison as it is. For this reason, there has been a problem that the vehicle position detection accuracy is poor when a plurality of similar road patterns exist in the target range.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a vehicle position detecting device capable of improving the vehicle position detecting accuracy and reducing the memory. Aim.

[0008]

According to a first aspect of the present invention, there is provided a vehicle position detecting apparatus for developing a traveling locus, which is calculated based on a vehicle direction and a traveling distance, into a one-dimensional symbol sequence and storing the developed traveling locus. Means, a feature pattern extraction means for extracting a feature pattern from the travel trajectory represented by the one-dimensional symbol sequence calculated by the travel trajectory calculation means, and a correspondence corresponding to the feature pattern extracted by the feature pattern extraction means Corresponding feature pattern search means for searching a feature pattern from an image map data storage means storing map data;
A position specifying unit that specifies the vehicle position by evaluating the characteristic pattern and a corresponding characteristic pattern corresponding to the characteristic pattern.

According to a second aspect of the present invention, there is provided a vehicle position detecting apparatus including a position specifying means for specifying a vehicle position by performing pattern matching between a characteristic pattern and a corresponding characteristic pattern.

According to a third aspect of the present invention, there is provided a vehicle position detecting device including a position specifying means for specifying a vehicle position by evaluating an arrangement relationship between a plurality of characteristic patterns and corresponding corresponding characteristic patterns. It is.

[0011] The vehicle position detecting apparatus according to the invention of claim 4, the corresponding feature patterns of feature pattern Hough (Houg
h) It is provided with a corresponding feature pattern search means for searching using conversion.

According to a fifth aspect of the present invention, there is provided a vehicle position detecting apparatus including a corresponding feature pattern search unit for generating a plurality of feature patterns by rotating a feature pattern within an error range.

According to a sixth aspect of the present invention, there is provided a vehicle position detecting device for extracting a characteristic pattern by dividing a traveling locus represented by a one-dimensional symbol sequence into a predetermined number and examining an angle distribution with an adjacent section. It is provided with.

According to a seventh aspect of the present invention, there is provided a vehicle position detecting device including a traveling trajectory calculating means for dividing the azimuth of the vehicle into a plurality of 360 degrees and assigning a code to each of the divided directions to express the traveling trajectory. .

The vehicle position detecting device according to the invention of claim 8 is provided with an image map data storage means expressed by allocating one bit to one dot.

According to a ninth aspect of the present invention, there is provided a vehicle position detecting apparatus including a hierarchical image map data storage unit.

[0017]

According to the first aspect of the present invention, there is provided a vehicle position detecting device comprising:
Expressing travel trajectories with one-dimensional series symbols reduces the amount of data, allows long-distance travel trajectories to be stored, and improves detection accuracy by identifying vehicle positions from feature patterns and corresponding feature patterns. I do.

In the vehicle position detecting device according to the second aspect of the present invention, the detection accuracy is improved by using pattern matching when the traveling locus is a loop or the like.

According to a third aspect of the present invention, the vehicle position detecting device obtains a vehicle position from a positional relationship with each corresponding characteristic pattern when there are a plurality of characteristic patterns.
The detection accuracy is improved.

The vehicle position detecting device according to the fourth aspect of the present invention can easily search a corresponding feature pattern from map data by using the Hough transform.

According to a fifth aspect of the present invention, the vehicle position detecting device can cope with an azimuth error of the vehicle by rotating the characteristic pattern to obtain a plurality of characteristic patterns.
The search rate improves.

The vehicle position detecting device according to the sixth aspect of the present invention can easily extract a characteristic pattern by dividing a traveling locus into a plurality of parts and examining an angular distribution of an adjacent section.

In the vehicle position detecting device according to the present invention, the vehicle travels by dividing 360 degrees and assigning a code in each direction.
The Rukoto to store locus, can reduce the amount of data.

The vehicle position detecting apparatus according to the eighth aspect of the present invention draws a map by using one dot as one bit.
A wide range of map data can be stored.

In the vehicle position detecting apparatus according to the ninth aspect of the present invention, the map data is hierarchized, for example, at a high speed or on a national road, so that the search speed of the corresponding feature pattern can be increased.

[0026]

【Example】

Embodiment 1 FIG. An embodiment of the present invention will be described below with reference to the drawings. In FIG. 1, reference numeral 1 denotes an azimuth sensor for detecting the azimuth of the vehicle, and 2 denotes a distance sensor for detecting the traveling distance of the vehicle. Numeral 3 denotes a running locus calculating means for calculating a running locus based on the detection data of the direction sensor 1 and the distance sensor 2, and
Stored in a. In the present embodiment, it is assumed that the traveling locus is stored in a form developed into a chain code sequence in eight directions described later. Reference numeral 4 denotes a position detection processing unit, which is the center of the apparatus, and includes a characteristic pattern extracting unit 41, a corresponding characteristic pattern searching unit 42, and a position specifying unit 43.

The characteristic pattern extraction means 41 examines the angular change distribution of the travel locus (by examining the angular change distribution of adjacent sections of a predetermined number each time since it is expressed by a chain code), Is extracted. The corresponding feature pattern search unit 42 creates a template based on the feature pattern extracted by the feature pattern extraction unit 41, reads out the map data of the corresponding area from the image map data storage unit 5 described later, and searches this map data. In this manner, a corresponding feature pattern corresponding to the feature pattern represented by the template is extracted.
The position specifying unit 43 specifies the position of the vehicle by evaluating the corresponding feature pattern searched by the corresponding feature pattern searching unit 42.

5 is an image map data storage means,
Data is read from map data 6 described later, developed into an image, and held. In this embodiment, it is assumed that the data is developed into a bit image and held. Reference numeral 6 denotes map data, which is a road map including coordinate information and the like used in a general navigation device. Reference numeral 7 denotes a display means for displaying a map, a current location, and the like.

Next, the operation will be described. The operation in FIG. 1 will be described with reference to the operation flowchart in FIG. In FIG. 1, a travel locus is calculated by a travel locus calculating means 3 based on detection data from an orientation sensor 1 and a distance sensor 2 (step ST201), and a chain code is calculated based on the calculated travel locus by a method shown in FIG. It is developed into a series and stored in the running locus storage means (step ST202).

In FIG. 3A, the direction of the vehicle is set to eight directions, and the direction numbers are set to 0-7. Based on this, FIG.
As shown in (b), a chain code sequence is formed by sequentially assigning each direction number every 10 m. Hereinafter, it is assumed that the description of the traveling locus is all represented by a chain code sequence.

Next, a characteristic pattern is extracted from the running locus (step ST203). The feature pattern referred to here is a right / left turn, refraction, curvature, loop, a long straight line of about 2 km, and a composite feature pattern combining these (when features are short and should be expressed as one feature pattern collectively). ). The angular change distribution for each fixed number in the running locus represented by the chain code sequence is examined.

Next, a method of extracting a characteristic pattern will be specifically described with reference to FIG. The left / right turn, refraction, and the like have a shape that rapidly changes in a short section as shown in FIG. As shown in FIG. 4B, the shape of the curve, loop, or the like changes in a certain section and does not change for a while, and in the case of a loop, the total angle change is 360 degrees or more. In addition, straight running and a curve having a large turning radius are shown in FIG.
The shape is such that almost no change can be seen.

When a feature is extracted by the above method (Yes in step ST204), the extracted feature pattern (here, the loop of FIG. 4B) is binarized as shown in FIG. Create a bit image template. Next, a map developed into a bit image of a corresponding area as shown in FIG. 6 is read from the created template and image map data storage means 5, and a corresponding feature pattern is searched from this map (step ST205).
As a search method, for example, there is a method of finding a superposition position where the number of dots overlapping each other is maximized by shifting the template on a map while shifting the template.

Next, it is checked whether the characteristic pattern is a loop or a composite characteristic pattern (step ST206). Here, the feature pattern is a loop (Ye in step ST206).
s), evaluate the shape of the feature pattern (step ST2)
07). When the ratio of the number of overlapping dots is equal to or larger than the threshold value and the number of the corresponding feature patterns is one (step ST2).
07, Yes), the end point of the corresponding feature pattern (● in FIG. 6)
Is used to specify the position. If No in step ST207, the position is not specified here, and the process proceeds to step ST20.
Return to 1 and wait for a new feature pattern.

In the above description, a loop is used as an extracted feature pattern. However, when a feature pattern is a composite feature pattern in which a plurality of feature patterns are extracted within a short distance, there is a possibility that a similar feature pattern exists. When the number is small, the position can be specified by evaluating the shape in step ST207 in the same manner as described above.

In the first embodiment, a method of specifying a position by shape evaluation in the case where the possibility of existence of a similar feature pattern in the vicinity such as a loop and a complex feature pattern is extremely low is shown. A long straight running or a curve having a large radius of rotation are characteristic patterns in which similar patterns are unlikely to be present in the vicinity, and evaluation by shape can be used as in the above-described embodiment. However, a long straight running or a large curvature of the turning radius can determine the traveling road of the vehicle, but it is difficult to specify a position on the traveling road. Therefore, in this case, not determination of the position, but determination of the traveling road of the vehicle is performed. Then, a new feature pattern is waited, and depending on the determined feature pattern obtained on the traveling road, the evaluation based on the shape is performed by using the above embodiment, or step ST2 described in a second embodiment described later.
08 may be used to try to identify the position.

In the first embodiment, the corresponding feature pattern is searched by superimposing the template and the image map. However, as another search method, as shown in claim 4, FIG. By setting the center of the loop as a representative point on the template and holding the radius of the loop as a parameter, the corresponding feature pattern search may be performed by the Hough transform used in the image processing.

Embodiment 2 FIG. Here, a position specifying method in a case where similar feature patterns such as right / left turn, refraction, and curvature with a small radius of gyration are relatively frequently seen as feature patterns will be described. In the case of a feature pattern such as a right / left turn, a bend, and a curvature having a small radius of rotation, the only corresponding feature pattern cannot be retrieved from the image map by the evaluation using one feature pattern as described in the first embodiment. Therefore, the position is specified using the correspondence between the arrangement relationship of a plurality of feature patterns and the arrangement relationship of a plurality of corresponding feature patterns described below.

In the operation flow chart of FIG. 2, since many similar feature patterns such as right / left turn, refraction and curvature with a small radius of rotation are often found here, the result is No in step ST206, and in this case, The arrangement relation of the characteristic patterns is evaluated (step ST208). Here, for example, an example of an evaluation method of an arrangement relation in a case where three right / left turns are obtained as a characteristic pattern as shown in FIG. 7 will be described below. The arrangement coincidence is calculated using two parameters of the distance error and the azimuth error.

Distance error de1 = 1− | pq | / | ab | de2 = 1− | qr | / | bc | azimuth error ae1 = 1− (pq · ab) / | pq || ab | ae2 = 1− ( <qr | bc) / | qr || bc | <condition> (1) | de1 |, | de2 | <threshold 1 (2) | ae1 |, | ae2 | <threshold 2 (3) | de1- de2 | <threshold 3

When all of the conditions (1), (2) and (3) are satisfied, the arrangement coincidence PS is calculated. PS = w1 · ( | ae1 | + | ae2 | ) + w2 | de1-de2 | w1; azimuth error coefficient w2; distance error coefficient

In this case, since the value of the arrangement coincidence PS is lower as the coincidence is higher, the arrangement coincidence PS has a value equal to or less than the threshold value and 1
At this time (Yes in step ST208), the position is specified using the position on the map corresponding to the end point of the final feature pattern (feature 3 in FIG. 7) (● in FIG. 7). Step ST208
If No in step ST201, the process returns to step ST201 and waits for a new feature pattern to be obtained.

In the second embodiment, the search for the corresponding feature pattern is performed by superimposing the template and the image map.
In the case of the feature pattern as shown in FIG. 9A, a middle point (O) between the start point and the end point as a representative point and an auxiliary point (P
_{1 to} P ₃ ) are similarly set on the template as shown in FIG. 8B (O is O ₁ , P _{1 to} P ₃ is P _{11 to} P _3).
31 ), a similar corresponding feature pattern search can be performed by the Hough transform used in the image processing.

In the second embodiment, the evaluation is performed using three characteristic patterns in which similar patterns are relatively large. However, the evaluation may be performed using four or more characteristic patterns in the same manner.

Also, in combination with the evaluation based on the shape of the first embodiment, a feature having a small number of similar patterns such as a loop or a complex feature pattern and a feature having a large number of similar patterns such as a right / left turn are evaluated for two or more arrangements. This may be performed to improve the position identification detection accuracy. This is particularly effective when a single corresponding feature pattern cannot be determined in the evaluation based on the shape as in the first embodiment.

Also, since it is extremely likely that a left / right turn, a bend, etc., is an intersection, image map data consisting of only intersection nodes is prepared, and a search for a corresponding feature pattern is performed using the intersection nodes as clues. By selecting the image map data to be searched according to the type of the feature pattern and performing the search with priority, the corresponding feature pattern search speed can be increased.

In the first and second embodiments, a feature pattern obtained from the calculated traveling distance is searched for a corresponding feature pattern by using a template having no rotation error. As described above, by generating templates in which several types of rotation errors assumed for the obtained feature patterns are added, and searching for the corresponding feature patterns using the method described in the first and second embodiments, The corresponding feature pattern search rate can be improved.

For a certain number for examining the angle change in the first and second embodiments, a short section is suitable depending on the type of the characteristic pattern (turn right / left, refraction, curvature having a turning radius of about 30 m, etc.). There are some feature patterns, but on the other hand, there are feature patterns that can only be extracted in a relatively long section (such as a curve with a large turning radius).
Left or right turn by set several different number examining angular change distribution, refractive, and can be extracted feature patterns to large curvature such as a small curvature of radius of rotation.

Also, since the vehicle is moving every moment during the execution of the first and second embodiments, the running trajectory being processed is stored and stored from the vehicle position determined in each embodiment. For example, Japanese Patent Laid-Open Publication No.
A new vehicle position can be presented by the method described in Japanese Unexamined Patent Application Publication No. 1 (1999).

Further, as shown in claim 9, in the first and second embodiments, the image map is hierarchized in advance (for example,
Main roads such as expressways, national roads and prefectural roads, and all roads
(E.g., a hierarchy), and by selecting according to the type of the feature pattern, the corresponding feature pattern search speed can be increased.

[0051]

As described above, according to the first aspect of the present invention, the traveling locus is represented by a one-dimensional symbol sequence, and the vehicle position is specified from the characteristic pattern and the corresponding characteristic pattern. In addition, the memory can be reduced, and the vehicle position can be detected with high accuracy.

According to the second aspect of the present invention, the position of the vehicle is specified by pattern matching between the characteristic pattern and the corresponding characteristic pattern. Therefore, any shape such as right / left turn, refraction, curve, loop, long straight travel, etc. The place of the running trajectory
Also in this case, there is an effect that the detection accuracy can be improved.

According to the third aspect of the present invention, the position of the vehicle is specified based on the positional relationship between the plurality of feature patterns and the corresponding feature patterns. There is an effect that detection accuracy can be improved when many similar feature patterns are seen.

According to the fourth aspect of the present invention, since the corresponding feature pattern is searched by the Hough transform, there is an effect that the corresponding feature pattern can be easily searched from the map data.

According to the fifth aspect of the present invention, since a plurality of characteristic patterns are obtained by rotating the characteristic pattern, it is possible to cope with a case where the traveling locus has an azimuth error. This has the effect of improving the search rate.

According to the sixth aspect of the present invention, the running trajectory represented by the one-dimensional symbol sequence is divided into predetermined numbers, and the characteristic pattern is extracted from the angular distribution with the adjacent section. There is an effect that a characteristic pattern can be extracted.

According to the seventh aspect of the present invention, the vehicle travels by dividing 360 degrees into a plurality of parts and assigning a code in each direction.
Since the trajectory is configured to be expressed, the data amount can be reduced and the memory can be reduced.

According to the eighth aspect of the present invention, since image map data drawn by assigning one bit to one dot is used, there is an effect that a wide range of map data can be stored.

According to the ninth aspect of the present invention, since the hierarchical image map data is used, the speed of searching for the corresponding feature pattern can be increased.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a position detecting device according to an embodiment of the present invention.

FIG. 2 is a flowchart for explaining an operation in the embodiment of FIG. 1;

FIG. 3 is a configuration diagram showing a running locus represented by a chain code sequence in the embodiment of FIG. 1;

FIG. 4 is a configuration diagram showing an angle change of a traveling locus in the embodiment of FIG. 1;

FIG. 5 is a configuration diagram illustrating an example of a template when a loop is detected as a feature pattern.

FIG. 6 is a configuration diagram showing an image map developed into a bit image.

FIG. 7 is a configuration diagram for explaining an example of an arrangement evaluation method according to a second embodiment of FIG. 1;

FIG. 8 is a configuration diagram showing a method for searching for a corresponding feature pattern by Hough transform.

FIG. 9 is a block diagram showing a conventional vehicle position detecting device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Direction sensor 2 Distance sensor 3 Travel locus calculating means 5 Image map data storing means 41 Feature pattern extracting means 42 Corresponding feature pattern searching means 43 Position specifying means

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-4-345190 (JP, A) JP-A-3-77014 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01C 21/00 G08G 1/0969

Claims (9)

(57) [Claims] 1. An azimuth sensor for detecting an azimuth of a vehicle, a distance sensor for detecting a traveling distance of the vehicle, and a traveling locus of the vehicle based on the detection of the azimuth sensor and the distance sensor. A travel locus calculation means for developing and storing the travel locus into a one-dimensional symbol sequence, an image map data storage means for storing map data, and a one-dimensional symbol sequence calculated by the travel locus calculation means. Feature pattern extraction means for extracting a feature pattern from a running locus; corresponding feature pattern search means for searching a corresponding feature pattern corresponding to the feature pattern extracted by the feature pattern extraction means from the image map data storage means;
A vehicle position detecting device comprising: a position specifying unit that specifies the vehicle position by evaluating the characteristic pattern and the corresponding characteristic pattern. 2. The method according to claim 1, wherein the position specifying unit specifies the vehicle position by performing pattern matching between the characteristic pattern extracted by the characteristic pattern extracting unit and the corresponding characteristic pattern searched by the corresponding characteristic pattern searching unit. The vehicle position detection device according to claim 1, wherein 3. The method according to claim 1, wherein the position specifying unit is configured to determine an arrangement relationship between the plurality of feature patterns extracted by the feature pattern extracting unit and a correspondence corresponding to each of the plurality of feature patterns searched by the corresponding feature pattern searching unit. The vehicle position detecting device according to claim 1, wherein the vehicle position is specified by evaluating an arrangement relationship with a characteristic pattern. Wherein said corresponding feature pattern retrieval means, according to claim 1, wherein the searching using the corresponding feature pattern Hough (Hough) conversion corresponding to the characteristic pattern extracted by the feature pattern extraction means Vehicle position detecting device. 5. The vehicle position detecting device according to claim 1, wherein said corresponding characteristic pattern search means obtains several types of characteristic patterns by rotating the extracted characteristic patterns within a prediction error range. 6. The characteristic pattern extracting means examines an angular change distribution with respect to an adjacent section obtained by dividing the traveling locus represented by the one-dimensional symbol sequence calculated by the traveling locus calculating means into a predetermined number. 2. The vehicle position detecting device according to claim 1, wherein a characteristic pattern of the traveling locus is extracted. 7. The running trajectory calculating means divides 360 degrees as the one-dimensional symbol sequence into a plurality of directions and calculates a running trajectory expressed by assigning a code to each of the divided directions. The vehicle position detecting device according to claim 1. 8. The image map data storage means includes:
The vehicle position detecting device according to claim 1, wherein map data in which one bit is assigned to each dot is stored. 9. The vehicle position detecting device according to claim 1, wherein said image map data storage means stores image map data hierarchized in advance.