JP2001101597A - Method and device for recognizing road shape and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately recognize a curve shape even in the case that the positional error of point data on a road is present. SOLUTION: By using a plurality of point data (a) arranged along the road, the shape of a curve provided on the road is recognized. Based on the plurality of point data, the rough shape of the curve, that is the approach line and parting line of the curve, is obtained (c). From rough shape information and the point data within the curve, the representative shape information of the curve is obtained. Preferably, circular arcs passing through each of points in the curve and inscribing the approach line and the parting line are individually obtained (d). From the plurality of circular arcs obtained in such a manner, the representative circular arc shape of the curve is obtained by a statistical processing (e).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for accurately recognizing a shape of a curve included in a road by using a plurality of point data arranged along the road.

[0002]

2. Description of the Related Art Conventionally, in the field of navigation devices, it has been proposed to notify a driver of information about a curve. For example, a curve ahead of the vehicle is detected using map data, and a curve warning is issued when the actual vehicle speed is higher than a speed suitable for the curve. In order to properly perform the curve notification, it is required to accurately recognize the curve shape.

[0003] Recognition of a curve shape is performed based on road map data. The road map data includes information (point data) indicating the positions of many points along the road. For example, a node in a link format map corresponds to point data.

[0004] In the conventional curve recognition, a combination of three points constituting one curve is selected, and the radius of a circle passing therethrough is calculated. A plurality of combinations of three points are created for one curve, and a radius is determined from each combination. Then, a plurality of radii are statistically processed, and, for example, a minimum value is obtained, thereby obtaining a representative radius of the curve. This kind of curve recognition method is disclosed in, for example, Japanese Patent Laid-Open No. 5-141.
No. 979. Also, JP-A-8-20
No. 2991 also discloses a method of recognizing a curve shape using an arc radius obtained from three points.

[0005]

However, in the prior art, as described below, there is a disadvantage that a recognition error is large because the accuracy of road map data on which curve recognition is based is strongly affected. .

[0006] Digital road map data is created by plotting roads in aerial photographs using, for example, a digitizing device. It is difficult to plot the centerline of a road accurately, even with fairly accurate equipment. Therefore, each point data contains some error.

[0007] A slight error in the point data greatly affects the curve recognition accuracy. This is because the radius of a circle passing through three points of data is determined in the conventional curve recognition technology. Even if the position of the point is slightly shifted, the radius of the circle changes greatly. For example, since one point is locally displaced, the radius of the circle related to the point becomes extremely small, and as a result, the radius of the curve may be recognized as being much smaller than it actually is.

Further, in digital road map data to be put to practical use, the number of points is reduced as much as possible in order to avoid an enormous amount of data. For example, points are roughly set after observing rules such as setting points at intersections. Instead of providing points at regular intervals, the distance between points is set longer for a straight line, and the distance between points is set shorter for a curve.

As described above, in general map data, points are set roughly and the distance between points is not constant. It is difficult to successfully apply the conventional method to such map data, and the radius of a circle passing through the three points greatly differs depending on the selected three points.

In the specific example of FIG. 22, a circle passing through three adjacent points is drawn at a curved portion on the road map data. Even within the same curve, the size of the circle varies greatly depending on the point selected. Therefore, it is difficult to accurately recognize the curve shape even by using these circles.

[0011] As described above, in the conventional road shape recognition method, due to the error of the point data, the size of the point interval, and the non-uniformity of the point interval. , The accuracy of curve recognition is reduced.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to provide a road shape recognition method capable of accurately determining a curve shape.

[0013]

SUMMARY OF THE INVENTION The present invention relates to a method for recognizing the shape of a curve included in a road by using a plurality of point data arranged along the road. In order to achieve the above object, a method of the present invention includes a step of obtaining coarse shape information representing a coarse shape of a curve based on a plurality of point data, and a representative shape of a curve from the rough shape information and the plurality of point data. Seeking information. Since the curve shape is obtained based on the rough shape information, the problem when using the conventional three-point passing circle is reduced. That is, the influence of the point position error on the calculated value of the curve shape can be reduced, and the accuracy of curve recognition can be improved.

Preferably, the rough shape information includes an approach line and an exit line of a curve. And preferably, the representative shape information of the curve is information indicating a representative inscribed arc inscribed in the approach line and the escape line, and the representative inscribed arc passes through each of a plurality of point data in the curve, the approach line and the inscribed line. This is a typical arc shape obtained from a plurality of inscribed arcs inscribed in the escape line.

According to the present invention, arcs passing through each point and inscribed into the approach line and the escape line are used. Since an inscribed circle is used, a curve shape can be obtained without being greatly affected by a point position error. This is because even if the point position is slightly shifted, the radius of the arc obtained from the point does not change much. Compared with the radius of a conventional arc passing through three points, the effect of the point shift is significantly reduced. As a result, a recognition error due to a point displacement can be significantly reduced. Similarly, even when the point interval is wide or the point-to-point distance is not constant, the calculation value of the curve radius does not largely depend on these conditions, so that the recognition error can be reduced.

In the present invention, a representative inscribed arc may be obtained after an inscribed arc is individually obtained for each point. Further, an appropriate arc that is in contact with the approach line and the escape line and that passes near the plurality of points may be directly obtained according to a principle such as the least square method.

In a preferred aspect of the present invention, a curve start point and a curve end point are determined from a plurality of point data, and the approach line and the exit line are determined based on the curve start point and the curve end point. As described above, the approach line and the exit line can be easily obtained by obtaining the curve start point and the curve end point.

[0018] Preferably, the curve start point and the curve end point are obtained based on the azimuth change amount and the point-to-point distance at each point. By considering the distance between the points, the start and end points of the curve can be accurately obtained.

Preferably, a filtering process is performed before obtaining a curve start point and the curve end point from a plurality of point data. The target of the filtering process is a point where both the azimuth change amount and the distance to the adjacent point are small, and the azimuth change amount is increased and corrected by the filter process based on the azimuth change amount of the adjacent point. As a result, it is possible to avoid erroneously recognizing a point whose azimuth change amount is unreasonably small due to a positional error as a curve end.

Preferably, of the point data constituting the straight line portion of the road, the point data other than both ends of the straight line are deleted as unnecessary in performing the curve recognition. This reduces the amount of data processing. Further, the start and end points of the curve can be easily determined, and the recognition accuracy can be improved.

[0021] Preferably, the plurality of point data is included in digital map data. In this digital map data, each point is set so as to represent necessary information on a road in accordance with a predetermined rule, and to have a low point density in a straight line portion and a high point density in a curved portion. In such map data, the setting of points is coarse and the distance between points is not uniform. When the method of the present invention is applied to this kind of map data, the effect of the present invention can be remarkably obtained.

Further, the present invention can take other forms for realizing the above road shape recognition method. For example, the present invention may be embodied in a form such as a road shape recognition device, a program recording medium, and a navigation device.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention (hereinafter, referred to as embodiments) will be described below with reference to the drawings.

First, the basic principle of the road shape recognition method according to the present invention will be described with reference to FIG. FIG. 1A schematically shows one curve portion of digital road map data used in a navigation device or the like. A method for performing curve recognition according to the present invention based on the road map data will be described below.

As shown in FIG. 1A, the road map data is composed of a plurality of point data arranged along the road. Each point is, for example, a node in a link type map. The point data preferably includes position coordinates, but the position of the point may be represented in other formats. Further, the point data is not provided in the map data from the beginning, and may be calculated from other elements in the map data.

Referring to FIG. 1B, in the present invention, first, a curve start point and a curve end point are obtained from a plurality of point data (specific processing will be described later in detail). Next, in FIG. 1C, an approach angle line (an approach line) as one form of the rough curve shape of the present invention.
And the escape angle line (escape line) is calculated. The coarse shape is information that roughly indicates the shape of the curve. The approach angle line is
This is a line that passes through the curve start point and is drawn along the curve approach direction. Similarly, the escape angle line is a line that passes through the curve end point and is drawn in the escape direction. The approximate shape of the curve is obtained by the entry angle line and the exit angle line.

In FIG. 1 (d) and FIG. 1 (e), the representative shape information of the curve is obtained from the above-mentioned entry and exit angle lines and the point data in the curve. First, FIG.
In this example, one arc that passes through one point in the curve and touches both the approach angle line and the escape angle line is determined. Similar inscribed arcs are determined individually for a plurality of points in the curve, preferably for all points in the curve. In FIG. 1E, a representative inscribed arc of the curve is determined from the plurality of inscribed arcs thus obtained.
Here, for example, an average value of the curve R is obtained, and for example, a maximum value or a minimum value of the curve R is obtained. Appropriate statistical processing may be employed depending on the subsequent use of the obtained curve shape.

In the above processing, a plurality of inscribed arcs are once obtained, and then a typical arc shape is obtained. On the other hand, an appropriate arc shape inscribed in the two lines and passing in the vicinity of a plurality of points in the curve may be directly calculated according to a principle such as a least square method.

FIG. 2 shows the effect of the present invention in comparison with the prior art. In the conventional technique of FIG. 2B, three points are selected from the points constituting the curve, and an arc passing through the three points is obtained. However, the size of the arc differs greatly depending on the three points to be selected. Therefore, even if these arcs are statistically processed, an appropriate representative curve shape cannot be obtained.

One of the causes of the variation in the arc shape in the prior art is the point position accuracy of the original road map data. Road map data is created by plotting points on the road from aerial photographs using a digitizing device,
It is created by calculating the road shape from the actual running trajectory of the vehicle, is created by performing local field measurement, and is created by connecting these data. In the digitizing apparatus, since the road is manually plotted using the original paper map or the aerial photograph, it is not easy to accurately plot the center line of the road. Further, when using the actual traveling locus, it is not easy to actually pass through a curve while tracing the same position (for example, the center) on the traveling lane.

Therefore, each point data of the road map data contains a certain degree of position error. On the other hand, from a geometrical point of view, the size of a circle passing through three points is
A slight movement of one of these three points will make a big difference. For this reason, in the prior art of FIG. 2B, a local position error of the point data greatly affects, and the size of the calculated curve arc varies greatly.

Originally, the conventional navigation road map data is created for the purpose of accurately calculating the route and the like, so that a slight displacement of the point data is not affected. Nevertheless, in the conventional method, the calculation process is performed on the assumption that the point data is accurate, and therefore, it is considered that the above-described problem occurs.

In view of such circumstances, in the present invention, a curve shape is obtained in consideration of a point position error so as not to be affected by the position error. That is, FIG.
In the present invention of (a), a rough shape of a curve, specifically, an approach line and an exit line are obtained, and an inscribed arc passing through each point and contacting the approach line and the exit line is obtained. The size of the arc thus obtained does not change significantly even if the position of the point slightly changes. Therefore, the position error of the point data does not significantly affect the calculated shape of the curved arc, and as a result, as shown in FIG.
The size of the arc passing through each point varies within a reasonable range. This is in contrast to the large variation in the conventional method of FIG. As described above, since the method of the present invention is not greatly affected by the point position error, the curve shape can be accurately obtained.

In the above, the effect of the present invention has been described in relation to the position error of the point data. Further, as described below, the present invention has an advantage that accurate recognition can be performed even when a point setting method at the time of map creation is considered.

In practical road map data, it is required to reduce the number of points as much as possible in order to avoid an enormous amount of data. In order to meet this demand, it is preferable to set points roughly while observing predetermined rules. Specifically, the predetermined rule is that points are set at intersections, changes in road types (from national roads to prefectural roads, etc.), and other landmarks.

In order to reduce the amount of data, it is general to set the distance between points longer for a straight line and shorter for a curve instead of setting points at regular intervals.

As described above, in practical road map data, points are set relatively coarsely, and the distance between points is not constant. Therefore, if you perform curve recognition on the assumption that points exist finely at regular intervals,
Accurate recognition may not be possible. However, according to the method of the present invention, even if the point setting is coarse or the point interval is not constant, the accuracy of curve recognition does not greatly change due to those factors, and therefore, the curve shape can be accurately recognized.

In the map illustrated in FIG. 3, the characteristics of the practical map described above are remarkably expressed, that is, the points are set roughly, and the distance between the points of the curved portion is short. Line P
0-P1-O is a curve entry line, and a line O-P4-P5 is a curve exit line. The arc in the figure is a curve shape obtained according to the method of the present invention. The radius of this arc is the average of the radii of the two inscribed arcs passing through P2 and P3. According to the present invention, a curve shape is accurately determined.

Next, a specific example of the above-described curve recognition processing according to the present invention will be described with reference to FIGS. FIG.
7 are flowcharts of the curve recognition processing according to the present embodiment. In the drawings after FIG. 8, the processing in each step is shown using a schematic diagram of map data. In FIG. 4, in S10, map data of a certain section is read. In S20, each point Pi in the map data is
Is calculated for the azimuth change θi. Referring to FIG. 8, the azimuth change amount is the amount by which the direction of the road changes at each point. Here, as the azimuth change amount θi, each point P
The angle between the road links on both sides of i is used.

<Deletion of Unnecessary Points> In S30, unnecessary point deletion processing is performed. This processing is to delete points other than both ends of the straight line portion. Here, points having an azimuth change amount equal to or smaller than the first predetermined threshold angle θt1 are deleted. This is because a point having a very small azimuth change amount is considered to be located in the middle of the straight line. As a result, points P1 and P7 shown in FIG. 8 are deleted. With this process,
Not only the amount of data to be calculated can be reduced, but also it becomes easier to determine a curve start point and a curve end point described later.

<Filtering> At step S40, filtering of the azimuth change is performed. This filtering process is shown in FIGS.
0 is shown. Referring to FIG. 9, a point to be filtered has an azimuth change amount (absolute value) smaller than a second predetermined threshold angle θt2, and a distance from the immediately preceding point is a predetermined threshold. The point is equal to or shorter than the distance dt. Here, the second predetermined threshold angle θt2 is set to be larger than the threshold angle θt1 in the unnecessary point deletion processing described above.

For the point Pi to be subjected to the filtering process, a value obtained by multiplying the azimuth change θi−1 of the immediately preceding point Pi−1 by the correction magnification k is given as a new azimuth change (the original azimuth). The change amount θi is ignored). The correction magnification k is, as shown in FIG.
It changes according to the distance d of i-1. The larger the distance between points, the smaller the correction magnification k is set.

In the example shown in FIG. 10, a filtering process is applied to the point P4. The original azimuth change amount θ4 of the point P4 is extremely small due to the position error of the point P4. Therefore, a value θ3 × k obtained by multiplying the azimuth change amount θ3 of the immediately preceding point P3 by the correction coefficient k is given to the point P4 as a new azimuth change amount θ4.

The reason why the above-described filtering process is provided will be described. As described above, the point position of the map data includes an error, and the distance between points is not constant.
Therefore, even in a single continuous curve, there is a point having a very small azimuth change amount, and the sign of the azimuth change amount may be reversed. Filter processing is for eliminating such absurdity. The filtering process gives an appropriate azimuth change amount based on the azimuth change amount of the immediately preceding point.

However, the azimuth change amount is a parameter to be used in the curve start and end point determination processing described below, and the purpose of the filter processing is to smoothly perform the determination processing. Therefore, the azimuth change amount is changed by the filter processing, but the position coordinates of each point are not changed.

In the above processing, a filter is applied according to the distance between points. As a result, it is possible to avoid the adverse effect of the filtering process at the portion where the two curves are continuous. Referring to FIG. 11, there is a long straight line portion between the two curves. At this time, it is not preferable to change the azimuth change amount θb of the start point Pb of the second curve based on the azimuth change amount θa of the end point Pa of the first curve. Therefore, a filter according to the distance between points is applied, and when the distance between points is large, the filtering process is prohibited.

<Curve start and end points> Returning to FIG. 4, when the filter processing in S40 is completed, a curve end point is calculated in S50, and a curve start point is calculated in S60. Here, a start point and an end point are selected from the point data in the map.
However, the actual curve start position and end position of the road do not always match the point installation location. Therefore, a correction process of the curve start and end positions is performed in a later process. Considering this, the start point and the end point selected here can be referred to as a temporary start point and a temporary end point, respectively.

(I) Calculation of Curve End Point First, the process of calculating the curve end point will be described with reference to FIG. In FIG. 12, points P1 and P7
Has already been deleted as an unnecessary point in S30. The azimuth change amount of each point is a value after the filtering process in S40. The linear distance between each point Pi and the next point Pi + 1 is defined as Li. In the end point calculation processing, the curve index Xi of each point is calculated according to the following equation.

[0049]

Xi = Li / (2 · sin (θi / 2)) The curve index Xi is an index used to determine whether the point Pi is a curve end point. At the end point of the curve, the distance to the next point becomes longer and the azimuth change amount becomes smaller. On the other hand, the curve index Xi also increases as the point-to-point distance increases, and increases as the azimuth change amount decreases. Therefore, it is possible to determine whether or not each point is the end point of the curve by using the curve index Xi.

Therefore, in this embodiment, the curve index Xi for all points from point P0 to point P8
Is calculated. When the curve index Xi exceeds a predetermined index threshold, it is determined that the corresponding point is a curve end point. In the example of FIG. 12, the point P6 is detected as a curve end point.

As described above, in the present embodiment, by using the curve index Xi, the curve end point can be accurately obtained in consideration of the azimuth change amount of each point and the distance between the points. For example, in a road map as shown in FIG. 13, two different curves C1 and C2 are continuous. The azimuth change amount θ5 of the final end point P5 of the first curve C1 and the azimuth change amount θ6 of the curve start point P6 of the second curve C2 are both relatively large. Therefore, it may not be understood that two different curves exist only by looking at the azimuth change amount. On the other hand, in the present invention, the distance between points is observed in addition to the azimuth change amount. Since the distance L5 between the point P5 and the point P6 is large, the curve index X5 of the point P5 increases, and it can be seen that this point P5 is the curve end point. Therefore, the two curves C1 and C2 are set to 1
It is possible to avoid erroneously recognizing two curves.

The reason why the distance between the points is longer at the end point of the curve is that, as described above, the points of the straight line portion are set coarser than the curve portion in a practical map. In the present embodiment, the curve end point is accurately detected using such characteristics of the map data. Further, in the present embodiment, the points of the straight line portion are thinned out by unnecessary point deletion processing as preprocessing. As a result, the distance between the points after the end of the curve is increased, and the end point can be easily detected.

By the way, referring to FIG. 14, the above-mentioned curve index Xi is actually similar to the conventional formula for calculating the curve R passing through three points. Focusing on the point P2, a virtual point P3 'is set on the line segment P2-P0. P2
Is set equal to the distance L2 between P2 and P3. The radius RX of the circle passing through the three points P3, P2, P3 'is
L2 / (2 · sin (θ / 2)), which is the same as the curve index Xi.

However, as can be seen from the setting of the virtual point P3 ', the curve index Xi is not the actual curve R. The conventional calculation formula is simply used as a determination index of the curve end point, and the radius R passing through three points is not necessarily obtained. And
Since the numerical value obtained by this calculation formula becomes larger as the azimuth change amount is smaller and the point-to-point distance is larger, the curve ending point is easily and accurately obtained by using this fact.

(Ii) Extraction of Stationary R Portion (Exclusion of Relaxation Curve Portion) In the present embodiment, it is preferable to extract a stationary R portion excluding the transition curve at the end of the curve by the following processing. FIG. 15 conceptually shows a change in the curve index Xi when the transition curve exists. The horizontal axis is the distance along the curve, and the vertical axis is the curve index Xi. The index threshold value Xt for determining the curve end point is determined based on the minimum value Xmin of the curve index. For example, the index threshold value Xt is set at a fixed ratio to the index minimum value Xmin (multiply Xmin by a predetermined numerical value).

As described above, in the present embodiment, the index threshold value Xt for determining the end point of the curve is determined according to the minimum value of the curve index, whereby the steady R portion excluding the transition curve portion is determined. The end point can be determined accurately.

In a specific process, the curve index Xi sequentially obtained for each point is compared with the curve index Xi-1 of the immediately preceding point. Newly calculated curve index Xi
Is smaller, the index threshold value Xt is updated based on the point index Xi. Newly calculated curve index X
If i is larger, the index threshold is not updated.
Thereby, the end point is determined using the index threshold value Xt obtained from the smallest curve index in the target curve.

FIG. 5 is a sub-flowchart of S50 in FIG. 4, and corresponds to the above-described process of extracting a steady curve portion. In S51, an appropriate initial value of the curve index threshold value is read. Next, a curve index Xi of the first target point is calculated (S52). In S53, the curve index threshold value is updated. That is, when the curve index calculated this time is smaller than the curve index for the immediately preceding point, a value obtained by multiplying the new curve index by a certain ratio is set as a new threshold value. In S54, it is determined whether or not the curve index of the target point is larger than the threshold. If NO in S54, the target point is moved to the next point in the forward direction in S55, and the process returns to S52.
If the determination in S54 is YES, the data is read with the target point as the curve end point (S56).

(Iii) Calculation of Curve Start Point Next, the process of calculating the curve start point (FIG. 4, S60) will be described. The process of calculating the curve start point is in principle exactly the same as the process of calculating the curve end point. However, the arithmetic processing is performed while following the points of the map data in the reverse direction.

Referring to FIG. 16, to determine the curve start point, (1) the azimuth change amount θi of each point Pi and (2) the difference between each point Pi and the immediately preceding point Pi-1 Is used. From these numerical values, the curve index Yi is obtained according to the following equation.

[0061]

Yi = Li−1 / (2 · sin (θi / 2)) For example, focusing on point P5, points P5 and P6
A virtual point P4 'is set on the line connecting. P4 'and P5
Is equal to the distance L4 between P5 and P4. The curve index Yi corresponds to the radius of a circle passing through the points P4, P5, and P4 '. However, as described above, this embodiment merely uses the radius calculation formula as a determination index of the curve start point. As the point distance increases, the curve index Y increases as the azimuth change amount decreases.
The property of the index Yi that i becomes large is used. Then, a curve index Yi of each point Pi is sequentially obtained from P8 to P0, and when the curve index Yi exceeds a threshold value, the point is set as a curve start point. FIG.
In the example of No. 6, point P2 is the curve start point.

FIG. 6 is a sub-flowchart of S60 in FIG. 4, and shows processing corresponding to the above-described method of extracting a steady curve portion. The processing in FIG. 6 is the same as the processing in FIG. 5 except that the update direction of the calculation target point is reversed, and a detailed description is omitted. In S61, the initial value of the curve index threshold value is read, and in S62, the curve index Yi of the first target point is calculated. In S63,
Until the curve index Yi reaches the minimum value, the index threshold is updated. In S64, it is determined whether or not the curve index of the target point is larger than the index threshold. If NO, the target point is updated in the reverse direction in S65, and the process returns to S62. If S64 is YES, the target point is read as a curve start point in S66.

<Calculation of Curve R> Returning to FIG.
After calculating the curve end point and the start point in S60, a curve R calculation process is performed in S70. The process of S70 is shown in the sub-flowchart of FIG. In FIG. 7, in S71, it is determined whether or not there is a point in the curve. Note that the points in the curve are points existing between the start point and the end point. In the map examples described so far, points within the curve exist. Therefore, the case where the determination in S71 is YES will be described first.

In S72, the curve entry angle line and the exit angle line are calculated using the curve start point and the curve end point. Referring to FIG. 17, an extension line P0-P2-O connecting the curve start point P2 and the immediately preceding point P0 is calculated as a curve approach angle line. Similarly, an extension line P8-P6-O of a line passing through the curve end point P6 and the point P8 immediately after the curve end point P6 is calculated as an escape angle line.

In S73, an arc corresponding to each point in the curve is calculated. The radii of three arcs tangent to the entry angle line and the exit angle line and passing through the respective points P3, P4, P5 are determined. Then, in S74, these three
From the two arc radii, the representative shape of the curve is determined. As the representative shape, a representative arc radius is calculated, and a contact point Ps between the arc and the entry angle line and a contact point Pe between the arc and the exit angle line are calculated.

In S74, an appropriate selection method may be used according to the method of using the curve recognition result. For example, when performing a curve warning, it is better to estimate the value of the curve small, so it is preferable to select the minimum value of the three arc radii. When simply transmitting the size of the curve R to the driver, the average value of the three curve radii may be used.

Next, the processing when the determination in S71 is NO will be described. S71 becomes NO when there is no point in the curve. In this case, the process proceeds to S75, and the curve R is calculated as shown in FIG.

In FIG. 18, both ends of the curve are Ps and Pe. Then, the central angle of the curve is set equal to the azimuth change amount of the corresponding point. Further, at both ends Ps and Pe of the curve, a straight line connecting the points and the actual road are separated by a predetermined distance D.
Suppose there is an error of Based on such an assumption, the curve radius value r is calculated according to the following equation.

[0069]

R = D / (1-cos θ2 / 2) <Calculation (correction) of curve start and end positions> Returning to FIG.
When the curve R calculation processing in S70 ends, in S80, the curve start position and the end position are calculated. This processing is performed in S5
This is a process for correcting the curve start and end points obtained in 0 and S60. In S50 and S60, the curve start and end points are selected from existing points. However, the existing points and the actual curve start and end positions often do not match. Therefore, the temporary start position and the temporary end position determined in S50 and S60. Therefore, in S80, as shown in FIG.
The curve start position and the end position are obtained again from the curve R obtained at 0.

In FIG. 19, a contact point Ps is a contact point between the curve obtained in S70 and the approach angle line. This contact Ps
And a point located on a straight line connecting the points. Here, the contact point P to the point connection line
The perpendicular foot from s (the point closest to the contact point Ps) is determined. The point obtained in this manner is defined as a curve start position S. Similar processing is performed for the curve end portion, and the obtained point is set as a curve end position E.

After the correction processing in S80, the target point is set as the curve end point (the point obtained in S50) in S90. Then, in S100, it is determined whether or not the above processing has been completed for all of the map data of the certain section read in the first S10. If S100 is NO, another curve may exist further. Therefore, the process proceeds to S110 to update the target point in the forward direction, and returns to S50. On the other hand, if the map data has ended in S100, the curve recognition processing ends.

The curve recognition method according to the present embodiment has been described above. Next, a road shape recognition device that realizes such a curve recognition method will be described. FIG. 20 is a block diagram illustrating a configuration of a navigation device including the road shape recognition device of the present embodiment. Navigation ECU
The control unit 10 controls the entire apparatus according to the navigation program and performs various navigation-related processes such as route calculation and route guidance. Navigation ECU1
0, the current position detector 12 for detecting the current position of the vehicle.
And a map data storage unit 14 storing road map data
And a display 16 and a speaker 18 as output means are connected. The navigation ECU 10 is provided with a recognition unit 20 for recognizing a curve shape by the above method. The recognizing unit 20 reads, from the storage unit 14, map data in a certain range from the current position to a predetermined distance ahead based on the current position of the vehicle detected by the current position detecting unit 12. Then, a curve existing on the read map data is searched.

The navigation ECU 10 further has a notifying unit 22, which notifies the driver of information on a curve ahead of the vehicle based on the curve shape recognized by the recognizing unit 20. The notification unit 22 performs, for example, a curve warning. When the vehicle reaches a predetermined distance before the curve (start position), an appropriate curve approach speed is obtained from the curve R. This approach speed is compared with the actual vehicle speed, and if the actual vehicle speed is higher, a curve warning is issued. For example, speaker 18
Is used to prompt the driver to decelerate by voice, and the display 16 is used to prompt deceleration.

In the above processing, the curve shape was recognized for the road near the current position of the vehicle. On the other hand, the curve recognition process may be performed on a wider range of road map or all road maps at once. For example, it is assumed that a medium such as a CD-ROM on which map data is recorded is mounted on the navigation device. At this time, a curve shape recognition process is performed on all map data. The recognized curve shape is stored in a readable and writable recording medium such as a hard disk. The curve shape data is read and used as needed in the subsequent navigation processing.

The curve recognition device and the navigation device as preferred embodiments of the present invention have been described above.
Still another embodiment of the present invention is a recording medium storing a program for causing a computer to execute the above-described curve recognition method. The recording medium may be any medium that can be read by electrical, electromagnetic, optical, or other means. For example, a CD-ROM, DVD, hard disk, flash R
AM and the like. This recording medium may record a road shape recognition program as a part of a larger program such as a navigation program. The recording medium may record the program according to the present invention together with a map database and the like.

Another embodiment of the present invention will be described below. Here, an apparatus or method for creating map data used in a navigation device or a curve alarm device is provided. When creating map data, it is preferable to reduce the overall capacity by thinning out points from the original map. The present invention is used for this decimation. That is, the curve shape of each curve is calculated by the above-described recognition method. Then, for each curve, the contact points Ps, Pe
And three points consisting of a point on the curve arc (the center point and the like) are recorded in the map data. The original point is deleted. When actually using this map for curve calculation on a vehicle, an arc passing through the three recorded points is calculated. According to such a map data creation method, it is not necessary to add a curve-specific data item to the map, and the conventional map data format can be used as it is. In addition, curve information such as curve R, curve start point, and curve end point can be accurately provided with a very small number of data.
And the total amount of map data can be reduced.

According to still another aspect of the present invention, a suitable map matching technique for a navigation system or the like is provided. The map matching is a process of correcting a current position with a characteristic portion such as a turning angle or a curve by comparing a traveling locus with road shape data. In this embodiment, the method of the present invention is used to calculate the trajectory of the vehicle.

Referring to the example of FIG.
It is assumed that the vehicle travels along the locus shown in (a), and as a result, a large number of point data shown in the drawing are obtained. The point data is obtained by using a current position detecting device such as a GPS.
Using these point data, a curve shape is obtained according to the present invention, as shown in FIG. In the example of FIG. 21, the shape of the curve is accurately obtained even though the vehicle travels straight on the curve.

As described above, according to the present invention, when calculating the traveling locus, the shape of the curve on which the vehicle traveled can be accurately obtained. The traveling locus thus determined is used for map matching, and the traveling locus can be compared with the road map data. Therefore, the accuracy of map matching can be improved.

[0080]

As described above, according to the present invention, even when there is an error in the points of the road map data, when the point setting is coarse, or when the distance between the points is not constant, such a case is obtained. The curve shape can be accurately recognized using the point data. Then, by using an accurate curve shape, a curve notification can be appropriately performed. Further, the curve shape recognized by the present invention can be suitably used for creating map data and the like.

[Brief description of the drawings]

FIG. 1 is a diagram schematically showing a road shape recognition method of the present invention.

FIG. 2 is a diagram showing a curve shape recognized by the method of the present invention in comparison with a conventional method.

FIG. 3 is a diagram showing a curve shape recognized by the method of the present invention.

FIG. 4 is a flowchart of a road shape recognition process according to a preferred embodiment of the present invention.

FIG. 5 is a flowchart showing a process of calculating a curve end point in S50 of FIG. 4;

FIG. 6 is a flowchart showing a process of calculating a curve start point in S60 of FIG. 4;

FIG. 7 is a flowchart showing a curve R calculation process in S70 of FIG. 4;

FIG. 8 is a schematic diagram of an example of map data, showing an unnecessary point deletion process.

FIG. 9 is a diagram illustrating a filtering process of an azimuth change amount of each point in the map data.

FIG. 10 is a diagram showing an example of map data on which the filtering process of FIG. 9 has been performed.

FIG. 11 is a diagram for explaining the effect of the filtering process of FIG. 9;

FIG. 12 is a diagram showing a process of calculating a curve end point.

13 shows a map of a portion where two curves are continuous, and is a diagram for explaining an advantage of the processing of FIG. 12;

FIG. 14 is a diagram for describing a curve index used in the processing of FIG. 12;

FIG. 15 is a diagram illustrating a process for extracting a steady R portion of a curve.

FIG. 16 is a diagram illustrating a process of calculating a curve start point.

FIG. 17 is a diagram illustrating a process of calculating a curve R using a curve start point and a curve end point.

FIG. 18: Curve R when there is no point in the curve
It is a figure showing a calculation process.

FIG. 19 is a diagram illustrating a process of calculating a curve start position.

FIG. 20 is a block diagram illustrating a configuration of a navigation device including the road shape recognition device of the present invention.

FIG. 21 is a diagram illustrating a technique for improving the accuracy of map matching using the road shape recognition method of the present invention.

FIG. 22 is a diagram illustrating a conventional technique.

[Explanation of symbols]

10 navigation ECU, 12 current position detecting section,
14 map data storage unit, 16 display, 18
Speaker, 20 recognition unit, 22 notification unit.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G09B 29/10 G09B 29/10 A (72) Inventor Motoki Ishiguro 1 Pioneer Corporation Kawagoe Factory (72) Inventor Kazushige Kaneko 25-25 Nishimachi, Yamada-shi, Kawagoe-shi, Saitama 1 Pioneer Corporation Kawagoe Factory F-term (reference) 2C032 HB06 HB11 HC31 HD29 2F029 AA02 AB12 AC02 AC14 AC18 AC19 AD01 5H180 AA01 BB13 CC30 EE02 FF03 FF25 FF27 LL01 LL02 LL07 LL08 LL15

Claims (12)

[Claims] 1. A road shape recognition method for recognizing a shape of a curve included in a road by using a plurality of point data arranged along a road, the method comprising the steps of: A road shape recognition method, comprising: obtaining rough shape information representing a shape; and obtaining representative shape information of a curve from the rough shape information and the plurality of point data. 2. The road shape recognition method according to claim 1, wherein the rough shape information includes an approach line and an exit line of a curve. 3. The road shape recognizing method according to claim 2, wherein the representative shape information of the curve is information indicating a representative inscribed arc inscribed in the approach line and the exit line. Shape recognition method. 4. The road shape recognition method according to claim 3, wherein the representative inscribed arc includes a plurality of inscribed arcs passing through each of a plurality of point data in the curve and inscribed in the approach line and the escape line. A road shape recognition method, characterized in that the road shape is a representative arc shape obtained from the above. 5. The road shape recognition method according to claim 2, wherein a curve start point and a curve end point are obtained from the plurality of point data, and the approach line and the curve end point are determined based on the curve start point and the curve end point. A road shape recognition method, wherein the escape line is obtained. 6. The road shape recognition method according to claim 5, wherein the curve start point and the curve end point are obtained based on an azimuth change amount and a point-to-point distance at each point. Shape recognition method. 7. The road shape recognition method according to claim 6, wherein before the curve start point and the curve end point are obtained from the plurality of point data, a point having a small azimuth change amount and a distance to an adjacent point. A method of recognizing a road shape, comprising: performing a filter process for increasing and correcting the azimuth change amount based on the azimuth change amount of the adjacent point. 8. The road shape recognition method according to claim 1, wherein point data other than both ends of the straight line out of the point data forming the straight line portion of the road is deleted as unnecessary in performing the curve recognition. A road shape recognition method characterized by the following. 9. The road shape recognition method according to claim 1, wherein the plurality of point data are included in digital map data, and the digital map data represents necessary information on a road according to a predetermined rule. A road shape recognition method characterized in that each point is set such that the point density is small in a straight line portion and the point density is high in a curved portion. 10. A road shape recognition device for recognizing a shape of a curve included in a road by using a plurality of point data arranged along a road, wherein the rough shape of the curve is determined based on the plurality of point data. A road shape recognition apparatus, comprising: means for obtaining rough shape information representing a shape; and means for obtaining representative shape information of a curve from the rough shape information and the plurality of point data. 11. A recording medium recording a program for recognizing a shape of a curve included in a road by using a plurality of point data arranged along a road, wherein the recording medium is based on the plurality of point data. A step of obtaining coarse shape information representing a coarse shape of the curve; and a step of obtaining representative shape information of the curve from the rough shape information and the plurality of point data. 12. A map data acquiring means for acquiring map data composed of a plurality of point data arranged along a road, and a curve recognizing means for recognizing a shape of a curve included in the road based on the map data. And a notifying means for notifying a driver of information about a curve ahead of the vehicle based on the recognized curve shape.In a navigation device having a curve notification function, the curve recognition means includes: A navigation device, comprising: means for obtaining rough shape information representing a rough shape of a curve based on the information; and means for obtaining representative shape information of the curve from the rough shape information and the plurality of point data.