JP4696974B2 - Road shape estimation apparatus and method - Google Patents

Description

  The present invention relates to an apparatus and an estimation method for estimating the shape of a road.

A method of estimating a road shape based on position information of a plurality of nodes constituting road map data is known (see, for example, cited document 1).
In this estimation method, three positions of a plurality of nodes constituting the road map data and the own vehicle position on the road are selected, and the first and second positions of the selected three positions are connected. A plurality of vehicle turning amounts θ / L are calculated based on the intersection angle θ formed by the link and the link connecting the second and third positions and at least the link length L between the second and third positions. The road shape is predicted based on an average value or a sum total of a plurality of vehicle turning amounts θ / L.

Prior art documents related to the invention of this application include the following.
JP 2000-321086 A

  By the way, in order to avoid an enormous amount of data, nodes representing road shapes stored in advance are set so that the node density is small in the straight line portion and the node density is high in the curve portion. There are many cases.

  As described above, when the vehicle turning amount θ / L is calculated by the above-described prediction method based on map information whose link length is not constant, there is a problem that a correct road shape cannot be recognized. In addition, since it is inevitable that the position information of the node includes a slight error, there is a problem that when the road shape is predicted based on the position information, the result varies.

(1) After dividing the x-axis and y-axis coordinates of a road node represented by nodes and links into an x-axis coordinate data string and a y-axis coordinate data string in the order of the nodes, and filtering each data string The road shape is estimated based on the x-axis coordinate data sequence of the filter calculation result and the y-axis coordinate data sequence of the filter calculation result.
(2) Interpolation points are set between nodes of roads represented by nodes and links, and the x-axis and y-axis coordinates of the nodes and interpolation points are arranged in the order of the nodes and interpolation points. After dividing each data string and performing a filter operation on each data string, a road shape is estimated based on the x-axis coordinate data string of the filter calculation result and the y-axis coordinate data string of the filter calculation result.

  According to the present invention, it is possible to reduce the influence of errors included in node position information, and to estimate an accurate road shape.

  An embodiment in which the road shape estimation device of the present invention is applied to a curve curvature radius estimation device to estimate the curve curvature radius of a curved road will be described.

  FIG. 1 shows a configuration of a road shape estimation apparatus according to an embodiment. The navigation device 1 stores road map data in a storage device (not shown) such as an IC card, a CD-ROM, or a magneto-optical disk, and an optimum route to a destination set by a passenger based on the road map data. The GPS receiver receives radio waves from GPS satellites, detects the vehicle position by satellite navigation, and displays the current position of the vehicle and the optimal route to the destination on the road map around the vehicle position. To guide.

  The navigation device 1 outputs map information and own vehicle position information within a predetermined range ahead of the host vehicle as shown in FIG. The map information provided from the navigation device 1 includes the position coordinates of the host vehicle, the position coordinates of each node on the road ahead of the host vehicle, the road type to which each node belongs, the link type, the road width, the speed limit, and the like. The types of roads include expressways, toll roads, national roads, prefectural roads, major local roads, and general roads. The link types include main line links, crossover links between main lines, ramp links, side road links, and SA side road links.

  The curve curvature radius estimation device 2 obtains map information and own vehicle position information from the navigation device 1, and estimates the curve curvature radius of the road ahead of the own vehicle based on these information. The curve curvature radius estimation device 2 includes a microcomputer, a ROM, a RAM, and the like (not shown). Depending on the software form of the microcomputer, a link angle calculation unit 21, an estimated path setting unit 22, a filter characteristic and setting unit 23, a filter and a curve curvature. A radius calculation unit 24 and the like are provided.

  The link angle calculation unit 21 calculates a link angle formed by a link with an adjacent node based on the map information from the navigation device 1. The estimated route setting unit 22 estimates a route along which the host vehicle travels based on the link angle calculated by the link angle calculating unit 21. The filter characteristics and setting unit 23 sets the characteristics of the filter based on the map information from the navigation device 1. The filter and curve curvature radius calculation unit 24 performs filter calculation processing and calculates a curve curvature radius based on the map information and the estimated route set by the estimated route setting unit 22.

  FIG. 3 is a flowchart showing a curve curvature radius estimation process according to an embodiment. The operation of the embodiment will be described with reference to this flowchart. In step 1, the position information of the own vehicle and map information within a predetermined range in front of the own vehicle are read from the navigation device 1. In the subsequent step 2, the link angle is calculated based on the position coordinates of each node included in the predetermined range ahead of the host vehicle.

As shown in FIG. 2, the coordinates of the nodes on the road ahead of the host vehicle are set in the order of P (Xk-1, Yk-1), Q (Xk, Yk), R (Xk + 1, Yk + 1) from the front in the traveling direction. Then, the link angle is defined as the intersection angle formed by the link PQ connecting the nodes P and Q and the link QR connecting the nodes Q and R. Therefore, since the following expressions (1) to (4) hold, the link angle θk can be uniquely determined from the expressions (3) and (4) in the range of [−π, π].

  In step 3, the estimated route setting process shown in FIG. 4 is performed to set a route on which the host vehicle will travel on the road ahead of the host vehicle. The route setting process will be described with reference to FIG. In step 11, it is confirmed whether or not the destination is set in the navigation device 1. When the destination is set, the process proceeds to step 25, and when the destination is not set, the process proceeds to step 12. When the destination is not set, in step 12, based on the map information from the navigation device 1, the number k is assigned in order from the node closest to the host vehicle on the road ahead of the host vehicle. Specifically, the node number k is initialized to 1, and the first node P1 closest to the host vehicle is stored in a predetermined storage area on the RAM as the first node S1 of the estimated route.

  Next, in step 13, it is confirmed whether or not the k-th node Sk of the estimated route is the end of the route, that is, whether or not the node is located at the end of the obtained map information of the road ahead of the host vehicle. . When the node Sk is not the node at the end of the map information ahead of the host vehicle, the process proceeds to step 14 to check whether or not the node Sk has a branch path. The determination as to whether or not there is a branch path is made based on, for example, whether there are a plurality of adjacent nodes in front of the node Sk based on the map information obtained from the navigation device 1.

  If the node Sk does not have a branch path, the process proceeds to step 15, and the node Pi + 1,1 next to Sk (= Pi) is set as the k + 1-th node Sk + 1 of the estimated path. Note that, in the node Pi + 1,1, the subscript i + 1 represents a node adjacent to the front of the node Pi, and the subscript 1 represents the first node of the node adjacent to the front of the node Pi. . In the subsequent step 16, k is updated to k = k + 1, i is updated to i = i + 1, and the process returns to step 13.

  On the other hand, if there is a branch path in the node Sk in step 14, the process proceeds to step 21 to determine the priority path according to the priority order corresponding to the link type shown in FIG. In the example of the priority order shown in FIG. 5, the priority order is highest when the link type is a main line link. Hereinafter, the side roads such as the crossover link, the ramp link, the side road link, and the service area (SA) between the main lines. The order of priority is set to be lower in the order of links and others. In this example, the priority order of the side road link and the side road link such as the service area is the same.

  By setting priorities in this order according to the link type, for example, it is unlikely that the host vehicle will travel to the side road while traveling on the main line, so the host vehicle will advance even if no destination is set. Predict the road that will be.

  If there are a plurality of links with the same priority in the process of step 21 and a priority route cannot be set based on the priority shown in FIG. 5, the process proceeds to step 22, and the road type shown in FIG. The priority route is determined according to the priority order. In the example of priority shown in FIG. 6, the priority is highest when the road type is an expressway or a toll road, and the priority is lowered in the following order: national or prefectural road, main local road, general road, and so on. Is set to In this example, the national road, the prefectural road, and the main local road are set to the same priority.

  By setting such priorities according to the road type, for example, when the front road is a national road and a general road at a three-way road, the vehicle is likely to travel on the national road, so even if the destination is not set It is possible to predict the road on which the vehicle will travel.

  If there are a plurality of roads having the same priority order ahead in the process of step 22 and the priority road cannot be set according to the priority order shown in FIG. 6, the process proceeds to step 23, and each of the roads calculated in step 2 of FIG. A priority path is determined based on the link angle θi, J at the node position. That is, the node Pi + 1, j having the smallest absolute value of the link angle θi, J is determined as the priority path, and the process proceeds to step 24. Note that, in the node Pi + 1, j, the subscript i + 1 represents a node adjacent to the front of the node Pi, and the subscript j represents the jth node of the node adjacent to the front of the node Pi. .

  When the priority route can be determined according to the link type at step 21 or when the priority route can be determined according to the road type at step 22, the process proceeds to step 24 as it is. In step 24, the estimated route Sk + 1 is set. That is, when the priority route is determined according to the link type in step 21 or when the priority route is determined according to the road type in step 22, the node located next to the node Pi determined as the priority route Among them, the node Pi + 1, j having the highest priority is set as the (k + 1) th node Sk + 1 of the estimated route.

  When the priority path is determined based on the link angle in step 23, the process proceeds to step 24, and the node Pi + 1, j having the smallest absolute value of the link angle determined as the priority path is set as the k + 1-th node of the estimated path. Set to Sk + 1. Then, after storing the set node Sk + 1 in a predetermined storage area of the RAM, the process proceeds to step 16 where k and i are updated.

  The process from step 13 to step 24 is repeated until the node Sk becomes a node located at the end of the map information ahead of the host vehicle. When the node Sk becomes a node located at the end, the process proceeds from step 13 to step 25 to estimate. Set the route.

  Specifically, when the destination is set, that is, when the process proceeds from step 11 to step 25 as it is, the optimum route to the destination obtained from the navigation device 1 is set as the estimated route. On the other hand, if the destination is not set and the process proceeds from step 11 to step 25 via step 13, the node Sk (k = 1, 2, 3) stored in a predetermined storage area of the RAM. , ...) is set as the estimated route. This completes the estimated route setting process.

  When the setting of the estimated route is completed, the process returns to step 4 in FIG. 3 to set the filter characteristics according to the road type in the map information obtained from the navigation device 1.

  In this embodiment, the linearity of the road is determined according to the road type shown in FIG. 6, and the filter characteristic is set. Here, a road with higher priority is determined to have higher linearity and a strong characteristic filter is set, and a weak characteristic filter is set for a road with lower priority.

  For roads with high linearity such as expressways, the filter is set to have strong characteristics so as to minimize the effect of errors, and conversely, for roads with frequent road changes such as mountain roads. Therefore, the filter can be set to a weak characteristic in order to improve the reproducibility of the road shape, and the optimum filter characteristic can be set according to the characteristics of the road. This makes it possible to accurately estimate the road shape.

  Here, the strong characteristic filter is a filter having a strong characteristic for removing noise and a weak characteristic for following a node point on the map. On the other hand, a weak filter is a filter that has a weak characteristic of removing noise and a strong characteristic of following a node point on a map.

  Note that the filter characteristics may be set according to the own vehicle speed, the speed limit of the road, the road width, the number of lanes, the yaw rate of the own vehicle, the lateral acceleration of the own vehicle, the distance between the node points, and the like. For example, when the vehicle speed is high, when the speed limit is high, when the road width is wide, when the number of lanes is large, when the yaw rate is small, when the lateral acceleration is small, when the distance between node points is long, the comparison is not made The same effect can be obtained by setting a strong filter.

  When the filter characteristics are set in this way, the process proceeds from step 4 to step 5 to execute filter calculation processing and curve curvature radius calculation processing. The filter calculation process and the curve curvature radius calculation process are performed according to the flowchart shown in FIG.

  Here, in the estimated route estimated in the process of step 3, as shown in FIG. 8, the node coordinates (having x-axis coordinates and y-axis coordinates) on the road ahead are in the order of the traveling direction Pf [0],. , Pf [i],..., Pf [Nfront_node]. Pf [0] represents a node immediately before the own vehicle. In addition, the node coordinates on the road through which the vehicle has passed are Pr [0], ..., Pr [i], ..., Pr [Nrear_node] in order of the traveling direction. Pr [0] represents the node point immediately behind the vehicle.

  Here, without performing the process of step 100, Pr [Nrear_node],..., Pr [i],..., Pr [0], Pf [0] ],..., Pf [i],..., Pf [Nfront_node], and after performing the processing of step 105, these points may be directly filtered.

  In step 100, as shown in FIG. 9, interpolation points are set at a predetermined distance interval ds on a curve (bending line) connecting the front and rear node points with straight lines. Interpolation point coordinates are set to Qh [0] (x [0], y [0]), ..., Qh [i] (x [i], y [i]), ..., Qh [N] (x [ N], y [N]).

  In the subsequent step 105, a range for performing the filter operation is set. As shown in FIG. 10, the filter calculation range is a range obtained by adding a predetermined distance before and after the range where the road turning radius is to be calculated. In this embodiment, the range in which the road turning radius is to be calculated is the range ahead of D1 to D2 [m] ahead of the vehicle position, and the range in which the filter calculation is performed is set as (D1−ΔD) to (D2 + ΔD). To do. Here, the value of ΔD is set to a larger value as the filter has stronger characteristics. Of the interpolation points obtained in step 100, the interpolation point closest to D1-ΔD and the interpolation point closest to D2 + ΔD are calculated as the Ns-th interpolation point and the Nf-th interpolation point, respectively, and filter calculation is performed within this range. Do.

In step 110, i is iteratively calculated from Ns to Nf by the following arithmetic expressions (5) and (6), thereby performing a filter operation for each of the x coordinate and the y coordinate.
xf [i] = b0.x [i] + b1.x [i-1] + b2.x [i-2] -a1.xf [i-1] -a2.xf [i-2] (5 ),
yf [i] = b0.y [i] + b1.y [i-1] + b2.y [i-2] -a1.yf [i-1] -a2.yf [i-2] (6 )
Here, b 0, b 1, b 2, a 1, a 2 are parameters for the filter operation, and xf [i], yf [i] are the results of the filter operation.

  In this embodiment, an example using a secondary filter is shown, but a filter of another order may be used. The road shape can also be estimated by performing only the filter calculation process in step 110 and calculating the road turning radius in step 130. In this embodiment, after the filter operation is further performed in the reverse direction in step 120, the road turning radius is calculated in step 130.

In step 120, i is repeatedly performed from Nf to Ns on the basis of the following arithmetic expressions (7) and (8) for xf [i] and yf [i] obtained in step 110, as opposed to step 110. , X-coordinate and y-coordinate filter operations.
xff [i] = b0.xf [i] + b1.xf [i-1] + b2.xf [i-2] -a1.xff [i-1] -a2.xff [i-2] (7 ),
yff [i] = b0.yf [i] + b1.yf [i-1] + b2.yf [i-2] -a1.yff [i-1] -a2.yff [i-2] (8 )

  As described above, when a road shape near a node point is estimated by performing a filter operation on the x-axis data and the y-axis data at the node point position or the interpolation point position, there is an error in the coordinate value of the node point. Even if it is included, it is possible to calculate a shape that smoothly connects the node points in the predetermined range before and after the node point, and to reduce the influence of the error included in the coordinate value of the node point. The road shape can be estimated.

  Further, in step 110 of FIG. 7, a filter operation from Ns to Nf is performed, and a filter operation is performed in the reverse direction from Nf to Ns in step 120 on the x-axis data and the y-axis data of the filter operation result. It is possible to prevent the road shape from being estimated in a distorted form. That is, when estimating the road shape in the vicinity of an i-th node point, the forward filter operation in step 110 calculates the coordinates of a point (Ns to i-th point) before the i-th node point. Calculated based on On the other hand, the filter operation in the reverse direction in step 120 is performed based on the coordinates of points deeper than a certain i-th node point (i-th point to Nf). By performing the filter operation in two steps, the forward direction and the reverse direction, a road shape near a certain i-th node point can be obtained by smoothly connecting points in both the front and back directions.

  Note that a filter that performs forward processing and backward processing as described above is referred to as a “filter having a zero phase characteristic”.

  In step 130, the road turning radius R is obtained from the following equation using the x, y coordinates xff [i], yff [i] after the filter processing obtained in step 120. In this embodiment, since the range in which the road turning radius is to be calculated is the range ahead of D1 to D2 [m] ahead of the host vehicle position, as shown in FIG. The nearest interpolation point to D1 and the nearest interpolation point to D2 are calculated to be the NRsth interpolation point and the NRfth interpolation point, and the turning radius R is calculated within this range.

First, i is repeated from NRs to NRf and the following calculation is performed to obtain the first-order difference and the second-order difference for each xy coordinate, and the turning radius R is obtained.
dx [i] = xff [i] −xff [i−1],
dy [i] = yff [i] −yff [i−1],
ddx [i] = dx [i] −dx [i−1],
ddy [i] = dy [i] −dy [i−1],
R [i] = {sqrt (dx [i] ² + dy [i] ² )} ³ / (dx [i] · ddy [i]-ddx [i] · dy [i]) (9)
A road turning radius R is obtained for each predetermined interval ds by the calculation in steps 100 to 130.

  Thus, according to one embodiment, interpolation points are set between nodes of roads represented by nodes and links, and the coordinates of the nodes and the x- and y-axes of the interpolation points are arranged in the order of arrangement of the nodes and the interpolation points. It is divided into an x-axis coordinate data string and a y-axis coordinate data string, each data string is subjected to a filter operation, and the road shape is determined based on the x-axis coordinate data string of the filter operation result and the y-axis coordinate data string of the filter operation result. Since the estimation is performed, the influence of the error included in the node position information can be reduced, and an accurate road shape can be estimated. In addition, by setting the filter characteristics to the optimum characteristics, it is possible to achieve both road shape reproducibility and performance of removing the influence of errors, and to estimate an accurate road shape. Furthermore, since the filter characteristics can always be set to the optimum characteristics regardless of the density of the nodes, an accurate road shape can be estimated.

  In the above-described embodiment, an example in which an interpolation point is set between nodes has been described. However, the interpolation point is not set, and the coordinates of the x-axis and y-axis of the road node represented by the node and the link are set as the node. Are divided into an x-axis coordinate data string and a y-axis coordinate data string in the order in which they are arranged, and each data string is subjected to a filter operation, and then converted into an x-axis coordinate data string as a filter calculation result and a y-axis coordinate data string as a filter calculation result. The road shape may be estimated based on this. Thereby, the influence of the error included in the position information of the node can be reduced, and an accurate road shape can be estimated. In addition, by setting the filter characteristics to the optimum characteristics, it is possible to achieve both road shape reproducibility and performance of removing the influence of errors, and to estimate an accurate road shape.

  According to one embodiment, since the filter operation is performed using a filter having zero phase characteristics, a road shape without distortion can be estimated, and an accurate road shape can be estimated.

  According to one embodiment, the filter calculation is performed on a road range that is longer by a predetermined distance than the range in which the road shape is estimated, so the place where the road shape is estimated is far from the vehicle position. Even when it is close to the case, the accurate road shape can be estimated.

  According to an embodiment, since the predetermined distance is increased as the filter characteristic is stronger, the range for performing the filter calculation can be set to an optimum range, and the calculation amount is unnecessarily increased. Can be prevented.

  According to one embodiment, the higher the linearity of the road, the stronger the filter is set, so that the optimum filter characteristic can be set according to the characteristics of the road. Can be estimated.

  The correspondence between the constituent elements of the claims and the constituent elements of the embodiment is as follows. That is, the curve curvature determination estimation apparatus 2 constitutes an interpolation means, a filter processing means, a shape estimation means, and a linearity determination means. The above description is merely an example, and when interpreting the invention, the correspondence between the items described in the above embodiment and the items described in the claims is not limited or restricted.

It is a figure which shows the structure of the curve curvature radius estimation apparatus of one Embodiment. It is a figure which shows an example of the map information ahead of the own vehicle. It is a flowchart which shows the curve curvature radius calculation process of one Embodiment. It is a flowchart which shows the presumed path | route setting process of one Embodiment. It is a figure which shows a response | compatibility with a link classification | category and a priority. It is a figure which shows a response | compatibility with a road classification and priority. It is a figure which shows a response | compatibility with road classification, road width, the number of lanes, speed limit, and link length maximum value. It is a figure for demonstrating the relationship between a node and the own vehicle position. It is a figure for demonstrating the setting method of an interpolation point. It is a figure for demonstrating the setting method of a filter calculation range. It is explanatory drawing for demonstrating the setting method of the calculation range of a road turning radius.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Navigation apparatus 2 Curve curvature radius estimation apparatus 21 Link angle calculation part 22 Estimated path setting part 23 Filter characteristic and setting part 24 Filter and curve curvature radius calculation part

Claims (9)

And x-axis coordinate filtering means for performing a full Iruta operation on x-axis coordinate data column of the node of the road in terms of nodes and links,
Y-axis coordinate filter processing means for performing a filter operation on the y-axis coordinate data string of the node;
A road shape estimation apparatus comprising: a shape estimation unit that estimates a road shape based on an x-axis coordinate data string as a filter calculation result and a y-axis coordinate data string as a filter calculation result. And interpolating means for setting the interpolation point between the nodes of the road in terms of nodes and links,
And x-axis coordinate filtering means for performing a full Iruta operation on x-axis coordinate data strings of the interpolation point and the node,
Y-axis coordinate filter processing means for performing a filter operation on the y-axis coordinate data string of the node and the interpolation point;
A road shape estimation apparatus comprising: a shape estimation unit that estimates a road shape based on an x-axis coordinate data string as a filter calculation result and a y-axis coordinate data string as a filter calculation result. In the road shape estimation device according to claim 1 or 2,
The x-axis coordinate filter processing means and the y-axis coordinate filter processing means use a filter having a zero phase characteristic. In the road shape estimation device according to claim 1 or 2,
The road shape estimation apparatus, wherein the x-axis coordinate filter processing means and the y-axis coordinate filter processing means perform a filter operation on a road range that is longer by a predetermined distance than the range in which the road shape is estimated. The road shape estimation apparatus according to claim 4,
The x-axis coordinate filter processing means and the y-axis coordinate filter processing means increase the predetermined distance as the filter characteristic is stronger. In the road shape estimation device according to claim 1 or 2,
It is equipped with a straight line determination means for determining the straightness of the road,
The road shape estimation device, wherein the x-axis coordinate filter processing means and the y-axis coordinate filter processing means set a stronger characteristic to the filter as the road linearity is higher. The road shape estimation apparatus according to claim 6,
The straightness determination means uses at least one of the own vehicle speed, the speed limit of the road, the road width, the number of lanes, the type of road, the yaw rate of the own vehicle, the lateral acceleration of the own vehicle, and the distance between the nodes. A road shape estimation device characterized by determining the nature. And x-axis coordinate data strings facilities to x-axis coordinate filter the full Iruta calculation step of the nodes of the road in terms of nodes and links,
A y-axis coordinate filtering step of performing a filter operation on the y-axis coordinate data string of the node ;
A road shape estimation method comprising: an estimation step of estimating a road shape based on an x-axis coordinate data string as a filter calculation result and a y-axis coordinate data string as a filter calculation result. An interpolation step of setting the interpolation point between the nodes of the road in terms of nodes and links,
And facilities to the x-axis coordinate filtering step the full Iruta operation on x-axis coordinate data strings of the interpolation point and the node,
A y-axis coordinate filter processing step of performing a filter operation on the y-axis coordinate data string of the node and the interpolation point ;
A road shape estimation method comprising: an estimation step of estimating a road shape based on an x-axis coordinate data string as a filter calculation result and a y-axis coordinate data string as a filter calculation result.

Images