JP6303810B2 - Vehicle position estimation apparatus and vehicle position estimation method - Google Patents

Description

  The present invention relates to a vehicle position estimation device and a vehicle position estimation method.

Conventionally, as a technique of a vehicle position estimation apparatus, there exists a technique of patent document 1, for example.
In the prior art described in Patent Document 1, when an image around a vehicle is photographed, and when three or more landmarks are detected from the photographed image, the interval between the selected landmarks among the three or more landmarks detected is detected. The relative angle of is calculated. Based on the calculated relative angle and the landmark position data stored in advance, the current position of the vehicle is estimated. Thereby, in the prior art described in Patent Document 1, the current position of the vehicle can be estimated even when a GPS (Global Positioning System) signal cannot be received.

JP 2001-142532 A

By the way, in the technique described in Patent Literature 1, the detection accuracy of the current position of the vehicle changes depending on the positional relationship between the selected landmark and the vehicle. Therefore, when the positional relationship between the selected landmark and the vehicle changes with the progress of the vehicle, and the detection accuracy of the current position of the vehicle decreases, it is necessary to switch the selected landmark. However, in the technique described in Patent Document 1, when the selected landmark is switched, the positional relationship between the selected landmark and the vehicle changes between immediately before the switching and immediately after the switching. Could fluctuate. Therefore, for example, when the vehicle steering control is performed using the estimation result of the current position of the vehicle, there is a possibility that the steered state fluctuates with the switching of the landmark.
The present invention provides a vehicle position estimation device and a vehicle position estimation method that can suppress fluctuations in the estimation result of the current position of the vehicle, which occurs with the switching of landmarks, focusing on the above points. Is an issue.

  In order to solve the above-described problem, in one embodiment of the present invention, the distance and direction of a plurality of landmarks existing around a vehicle are detected. Subsequently, the positions of a plurality of landmarks with respect to the vehicle are detected based on the detected distance and direction. Subsequently, the detected positions of the plurality of landmarks are collated with the positions of the landmarks on the map. Subsequently, a pair of landmarks are selected from the plurality of landmarks. Subsequently, the current position of the vehicle is estimated on the basis of a collation result between at least the position of the pair of landmarks and the position of the landmark on the map. In this case, a pair of landmarks is based on a relative angle that is an angle formed by straight lines extending from the vehicle to each of the two landmarks and a relative distance that is a distance along the vehicle traveling direction between the two landmarks. Select.

  In one aspect of the present invention, in consideration of both the relative angle between two landmarks and the relative distance in the vehicle traveling direction, a pair of landmarks having both relatively large relative angles and relative distances can be selected. Therefore, it is possible to suppress a change in the relative angle between the selected landmarks as the vehicle travels. Thereby, the same landmark can be used for a long time, and the fluctuation | variation of the estimation accuracy of the present position of a vehicle which generate | occur | produces with switching of the selected landmark can be suppressed.

It is a conceptual diagram showing schematic structure of the vehicle position estimation apparatus 1 which concerns on 1st Embodiment. It is a flowchart showing a present position calculation process. It is a figure for demonstrating a map area | region. 4 is a diagram for explaining a method for calculating a current position of a vehicle A. FIG. It is a flowchart showing a priority order setting process. It is a figure showing operation | movement of the combination exclusion part 4aa. It is a figure showing operation | movement of the vehicle position estimation apparatus 1. FIG. It is a figure showing operation | movement of the vehicle position estimation apparatus 1. FIG. It is a flowchart showing the present position calculation process of the modification of 1st Embodiment. It is a figure showing operation | movement of the vehicle position estimation apparatus 1 which concerns on 2nd Embodiment. It is a figure showing operation | movement of the vehicle position estimation apparatus 1. FIG.

Embodiments of the present invention will be described below with reference to the drawings.
(Constitution)
As shown in FIG. 1, the vehicle position estimation device 1 is mounted on a vehicle A. The vehicle position estimation device 1 includes a map information storage unit 2, a distance detection unit 3, and a controller 4.
The map information storage unit 2 stores map information. As map information, for example, the position of the road (latitude / longitude), shape (curvature), number of lanes, position of landmark B in the world coordinate system (latitude / longitude) on the map (hereinafter also referred to as “world coordinate system”). Some include longitude), shape (planar shape), and the like. Examples of the landmark B include a utility pole, a streetlight, and a traffic light. The world coordinate system includes a predetermined absolute coordinate system such as a geographic coordinate system.

The distance detection unit 3 detects the distance between the target around the vehicle A and the vehicle A and the direction of the target with respect to the vehicle A. Then, the distance detection unit 3 outputs the detection result to the controller 4. As the distance detection unit 3, for example, a laser range finder or radar that emits laser light to the surroundings and calculates the distance and direction of the target using reflected light of the emitted laser light can be employed.
The controller 4 includes an A / D (Analog to Digital) conversion circuit, a D / A (Digital to Analog) conversion circuit, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. Integrated circuit. The ROM stores one or more programs that realize various processes. Then, the controller 4 performs various processes (for example, a current position calculation process described later) according to one or more programs stored in the ROM based on the stored contents of the map information storage unit 2 and the detection results of the distance detection unit 3. Execute. The controller 4 also includes a landmark selection unit 4a, a distance direction detection unit 4b, a landmark position detection unit 4c, a landmark collation unit 4d, and a current position estimation unit 4e.

  Based on the map information stored in the map information storage unit 2, the landmark selection unit 4 a selects a pair of landmarks B (hereinafter, “estimation landmarks” from among a plurality of landmarks B present around the vehicle A. B1, B2 ") is selected. The estimation landmarks B1 and B2 are selected in accordance with the angle formed by the straight lines extending from the vehicle A to each of the two landmarks B (hereinafter also referred to as “relative angle θ”) and the vehicle travel between the two landmarks B. This is performed based on the distance along the direction (hereinafter also referred to as “relative distance L”). For example, the selection is made from a combination group of a pair of landmarks B that can be selected from a plurality of landmarks B. Then, the landmark selection unit 4a outputs the selection result to the landmark position detection unit 4c.

Specifically, the landmark selection unit 4a includes a combination exclusion unit 4aa, a first selection unit 4ab, a second selection unit 4ac, a selection execution unit 4ad, and a priority order setting unit 4ae.
In the combination excluding unit 4aa, the relative distance L between the two landmarks B in the combination group is a predetermined set distance (for example, a distance shorter than a general arrangement interval 30 [m] of utility poles. 20 [m]. ) Less than landmark B combinations. Thereby, the combination exclusion unit 4aa excludes the estimation landmarks B1 and B2 from the selection targets.

  The first selection unit 4ab includes a combination of a landmark B having a large relative angle θ between two landmarks B and a combination of a small landmark B out of a combination group (a combination group after exclusion by the combination exclusion unit 4aa). And a combination of landmarks B having a large relative angle θ is selected. Specifically, the first selection unit 4ab selects a combination that is equal to or greater than a predetermined set angle (for example, 30 °) from the combination group after exclusion by the combination exclusion unit 4aa.

  The second selection unit 4ac distinguishes the combination of the landmark B having the long relative distance L and the combination of the short landmark B from the combination group (the combination group after the exclusion by the combination excluding unit 4aa), and the relative distance. A combination of landmarks B with a long L is selected. Specifically, the second selection unit 4ac has a predetermined set distance (for example, a general arrangement interval of utility poles and street lamps 30 [m]) or more from the combination group after exclusion by the combination exclusion unit 4aa. Select a combination.

The priority order setting unit 4ae executes a priority order setting process for setting a priority order for the combination of landmarks B used for estimating the current position of the vehicle A. Details of the priority setting process will be described later.
The selection execution unit 4ad is a combination of the landmarks B selected by both the first selection unit 4ab and the second selection unit 4ac, that is, two landmarks having a relatively large relative angle θ and a relatively long relative distance L. From a combination of B, a pair of landmarks B (estimation landmarks B1 and B2) having the highest priority ("1") is selected. Then, the selection execution unit 4ad outputs the selection result to the landmark position detection unit 4c. If the landmark B of the combination with the highest priority (“1”) cannot be detected by the distance detection unit 3 due to an obstacle or the like, the landmark B with the priority “2” is detected as a pair of landmarks B (estimated). Landmarks B1, B2).

  Based on the detection result output by the distance detection unit 3, the distance direction detection unit 4 b determines the distance between the vehicle A and each of the plurality of landmarks B existing around the vehicle A (hereinafter, “landmark distance”). And a direction with respect to the vehicle A (hereinafter also referred to as “landmark direction”). A pair of landmarks B (estimation landmarks B1 and B2) selected by the landmark selection unit 4a are included. The distance direction detection unit 4b outputs the detection result to the landmark position detection unit 4c. Of the plurality of landmarks B, only a pair of landmarks B (estimation landmarks B1 and B2) selected by the landmark selection unit 4a may be detected.

  The landmark position detection unit 4c detects the positions of a plurality of landmarks B with respect to the vehicle A (for example, positions in the moving coordinate system) based on the landmark distance and the landmark direction detected by the landmark position detection unit 4c. As the moving coordinate system, for example, there is a relative coordinate system having the front end of the vehicle A as an origin. The landmark position detection unit 4c outputs the detection result to the landmark collation unit 4d. Of the plurality of landmarks B, only the position of the pair of landmarks B (estimation landmarks B1 and B2) selected by the landmark selection unit 4a may be detected.

  The landmark collation unit 4d collates the positions of the plurality of landmarks B detected by the landmark position detection unit 4c with the positions of the landmarks B on the map stored in the map information storage unit 2. Then, the landmark collation unit 4d outputs the collation result to the current position estimation unit 4e. Of the plurality of landmarks B, only a pair of landmarks B (estimation landmarks B1 and B2) selected by the landmark selection unit 4a may be collated.

  The current position estimation unit 4e includes at least the positions of the estimation landmarks B1 and B2 and the landmark B (estimated on the map) stored in the map information storage unit 2 among the plurality of landmarks B by the landmark matching unit 4d. The current position of the vehicle A is estimated based on the result of collation with the positions of the landmarks B1, B2). Note that if there is a landmark B that can be collated in addition to the estimation landmarks B1 and B2 by the landmark collating unit 4d, the current position of the vehicle A may also be used for the estimation.

(Current position calculation process)
Next, the current position calculation process executed by the controller 4 will be described.
As shown in FIG. 2, in step S <b> 101, the controller 4 reads map information of an area where the vehicle A is expected to travel from the map information storage unit 2. As an area where the vehicle A is expected to travel, for example, when the vehicle A includes a navigation device that presents a travel route from the current position of the vehicle A to the destination, an area around the presented travel route is employed. it can. Further, for example, when the vehicle A includes an automatic driving device that automatically operates from the current position of the vehicle A to the destination, an area around the driving route of the automatic driving can be adopted.

  Subsequently, the process proceeds to step S102, where the controller 4 (landmark selection unit 4a), as shown in FIG. (Hereinafter also referred to as “map area”). The length of the map area in the vehicle traveling direction is the vehicle traveling direction length d of the detection range of the distance detection unit 3. In addition, the length of the map area in the direction orthogonal to the vehicle traveling direction is the sum of the width of the road and the width of the sidewalk.

  Subsequently, the process proceeds to step S103, and the controller 4 (landmark selection unit 4a) selects a map area including the current position of the vehicle A and a map area that goes ahead from the map areas divided in step S102. As the map area including the current position of the vehicle A, for example, when the determination in step S109 described later is performed during the execution of the current position calculation process, the map area determined in step S109 is selected. Further, as the map area including the current position of the vehicle A, for example, when the determination in step S109 is not performed, a map area including the current position of the vehicle A acquired from a GPS receiver (not shown) is selected. Subsequently, the controller 4 (landmark selection unit 4a) acquires the position of the landmark B existing in the selected map area (position in the world coordinate system) from the map information read in step S101.

Subsequently, the process proceeds to step S104, where the controller 4 (landmark selection unit 4a) estimates the current position of the vehicle A for each map area divided in step S102 based on the position of the landmark B acquired in step S103. Priority level setting processing for setting a priority level for the combination of landmarks B used in the above is executed. Details of the priority setting process will be described later.
Subsequently, the process proceeds to step S105, and the controller 4 (landmark selection unit 4a) selects the combination of landmarks B having the highest priority ("1") in the detected map area according to the priority set in step S104. A pair of landmarks B (estimation landmarks B1 and B2) are selected.

  Subsequently, the process proceeds to step S106, where the controller 4 (distance direction detection unit 4b) estimates based on the detection result (distance and direction from the vehicle A to the target) output by the distance detection unit 3 in step S105. The distance (landmark distance) between each of the landmarks B1 and B2 and the vehicle A is detected. Subsequently, the controller 4 (distance direction detection unit 4b) detects the directions (landmark directions) of the estimation landmarks B1 and B2 with respect to the vehicle A. When the combination of landmarks B having the highest priority (“1”) cannot be detected by the distance detection unit 3 due to an obstacle or the like, the landmark B having the priority “2” set in S104 is used as a pair of landmarks. The mark B (estimation landmarks B1 and B2) is selected again, and the landmark distance and the landmark direction are detected.

For example, if the controller 4 (distance direction detection unit 4b) detects the current position of the vehicle A in step S107 described later during the execution of the current position calculation process, the vehicle A detected in the previous step S107. The distance and direction of the estimation landmarks B1 and B2 are estimated in consideration of the current position and direction of the vehicle A that are predicted based on the current position of the vehicle A and the traveling state such as the vehicle speed, yaw rate, and lateral acceleration of the vehicle A. . Then, the controller 4 (distance direction detection unit 4b) uses the landmarks around the estimated distance and direction as the estimation landmarks B1 and B2 based on the detection result output by the distance detection unit 3, and determines the landmark distance and the land. Detect the mark direction.
Subsequently, the process proceeds to step S107, and the controller 4 (landmark position detection unit 4c) determines the position (movement) of the estimation landmarks B1 and B2 with respect to the vehicle A based on the landmark distance and landmark direction detected in step S106. The positions (x1, y1), (x2, y2)) of the estimation landmarks B1, B2 in the coordinate system (relative coordinate system) are detected.

Subsequently, the process proceeds to step S108, where the controller 4 (landmark collation unit 4d) stores the position (x1, y1), (x2, y2) of the estimation landmarks B1, B2 calculated in step S106, and the map information storage. The positions (X _G 1, Y _G 1) and (X _G 2, Y _G 2) of the landmark B (estimation landmarks B1, B2) on the map stored in the part 2 are collated. Subsequently, the controller 4 (current position estimation unit 4e) estimates the current position and direction (X, Y, φ) of the vehicle A based on the collation result. Specifically, the controller 4 (current position estimation unit 4e), as shown in FIG. 4, positions (x1, y1), (x2, y2) of the estimation landmarks B1, B2 in the movement coordinate system (relative coordinate system). y2) and an estimation landmark B1 in the world coordinate system (absolute coordinate system) based on the current position and orientation candidates (X ′, Y ′, φ ′) of the vehicle A in the world coordinate system (absolute coordinate system) , B2 position candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′) are set.

  Here, as the current position and orientation candidates (X ′, Y ′, φ ′) of the vehicle A in the world coordinate system (absolute coordinate system), for example, during the execution of the current position calculation process, the vehicle A in step S107. When the current position of the vehicle A is detected, the current position of the vehicle A predicted based on the current position of the vehicle A detected in the previous step S107 and the traveling state of the vehicle A such as the vehicle speed, yaw rate, and lateral acceleration. The position and orientation, and the positions and orientations around these are adopted. In addition, as candidates (X ′, Y ′, φ ′), for example, when the current position of the vehicle A is not detected in step S107 during the execution of the current position calculation process, a GPS receiver (not shown) is illustrated. The current position and orientation of the vehicle A acquired from the above, and the positions and orientations around these are adopted.

Subsequently, the controller 4 (landmark matching unit 4d) selects one candidate (X1 ′, Y1 ′), (X2 ′) from the set candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′). , Y2 ′). Subsequently, the controller 4 (landmark matching unit 4d) selects the selected candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′) and the world coordinate system (absolute coordinates) represented by the map information acquired in step S101. An estimation evaluation value is calculated based on the positions (X _G 1, Y _G 1) and (X _G 2, Y _G 2) of the estimation landmarks B1, B2. The estimated evaluation value, for example, the selected candidate (X1 ', Y1') and the position and the _{(X G 1, Y G 1} ) a distance between d1, the selected candidate (X2 ', Y2') and the position (X _There is a sum (d1 + d2) of the distance d2 with _G2 , YG2).

Then, the controller 4 (landmark position detection unit 4c) sequentially selects the other candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′), repeatedly executes the above flow, and all candidates (X1 The estimation evaluation value (d1 + d2) is calculated for ', Y1') and (X2 ', Y2'). As a result, the controller 4 (landmark collation unit 4d) makes the estimation landmarks B1 and B2 positions (x1, y1) and (x2, y2) and the land on the map stored in the map information storage unit 2. The positions (X _G 1, Y _G 1) and (X _G 2, Y _G 2) of the mark B (estimation landmarks B1, B2) are collated. Subsequently, the controller 4 (current position estimation unit 4e) has the candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′) for which the estimation evaluation value (d1 + d2), which is the collation result of the landmark collation unit 4d, becomes the minimum. ) Is calculated. The candidates (X ′, Y ′, φ ′) of the current position and orientation (X, Y, φ) of the vehicle A used for setting the calculated candidates (X1 ′, Y1 ′), (X2 ′, Y2 ′) The current position and direction (X, Y, φ) of the vehicle A are set (estimated).

Then, it transfers to step S109 and the controller 4 determines whether an ignition switch is an OFF state. If the controller 4 determines that the ignition switch is in the off state (Yes), the controller 4 ends this calculation process. On the other hand, if the controller 4 determines that the ignition switch is not in an off state, that is, the ignition switch is in an on state (No), the controller 4 proceeds to step S110.
In step S110, the controller 4 determines in which map area of the plurality of map areas divided in step S102 the current position of the vehicle A calculated in step S108 is, and then proceeds to step S103. Transition.

(Priority setting process)
Next, priority order setting processing executed by the controller 4 will be described.
As shown in FIG. 5, in step S201, the controller 4 (landmark selecting unit 4a, combination excluding unit 4aa) is based on the position of the landmark B acquired in step S103 for each map area divided in step S102. A combination group of a pair of landmarks B that can be selected from a plurality of landmarks B present in the map area is acquired from the map information stored in the map information storage unit 2. At that time, the controller 4 (landmark selecting unit 4a, combination excluding unit 4aa) sets a distance (relative distance L) along the vehicle traveling direction between the two landmarks B from the combination group. (For example, a distance shorter than a general arrangement interval 30 [m] of utility poles. 20 [m]). Thereby, the controller 4 (combination exclusion unit 4aa) excludes the estimation landmarks B1 and B2 from the selection targets.

  As shown in FIG. 6A, the estimation accuracy of the current position of the vehicle A becomes higher as the angle (relative angle θ) formed by the straight lines extending from the vehicle A to the estimation landmarks B1 and B2 increases. However, as shown in FIG. 6B, when the estimation landmark B1 directly on the left of the vehicle A and the estimation landmark B2 on the right of the vehicle A are used for estimating the current position of the vehicle A. As the vehicle A travels, the relative angle θ between the estimation landmarks B1 and B2 rapidly decreases, and the estimation accuracy of the current position of the vehicle A rapidly decreases. Therefore, when a combination of landmarks B having a short distance along the traveling direction of the vehicle A is used, there is a possibility that fluctuations in the detection accuracy of the current position of the vehicle A increase with time. Therefore, the controller 4 excludes a combination in which the distance along the traveling direction of the vehicle A (relative distance L) is equal to or less than a set distance (for example, 20 [m]) from the selection of the estimation landmarks B1 and B2. .

  Subsequently, the process proceeds to step S202, and the controller 4 (landmark selection unit 4a) assumes that the vehicle A has entered the center of the lane at the entrance of the map area for each map area divided in step S102. The relative angle θ is calculated for each of the two landmarks B represented by the combination group (combination group after exclusion) acquired in step S2. For example, the relative angle θ is calculated based on the map information stored in the map information storage unit 2.

  Subsequently, the process proceeds to step S203, in which the controller 4 (first selection unit 4ab) has a large relative angle θ calculated in step S202 from the combination group (combination group after exclusion) acquired in step S201. The combination of B and the combination of small landmarks B are distinguished and a combination of landmarks B having a large relative angle θ is selected. Specifically, the controller 4 (first selection unit 4ab) excludes the combination in which the relative angle θ calculated in step S202 is less than a set angle (for example, 30 °) from the selection of the estimation landmarks B1 and B2. To do. And the controller 4 (1st selection part 4ab) sets (selects) the combination after exclusion as a 1st group (1st selection process).

  Thereby, the controller 4 (first selection unit 4ab) exists in the region (map region) based on the storage contents of the map information storage unit 2 before the vehicle A enters the predetermined region (map region). Among combinations, the combination of landmarks B with a large relative angle θ and the combination of landmarks B with a small relative angle are distinguished, and the combination of landmarks B with a large relative angle θ (set angle (for example, 30 °) or more). Select a combination).

  Note that if the controller 4 (first selection unit 4ab) has executed step S205, which will be described later, during the execution of the priority order setting process, the current step S203 is selected from the group set in step S205. The highest group (first group) is selected from the groups not yet selected. The order of groups is higher as the number is smaller, such as first group> second group> third group> fourth group. Subsequently, the controller 4 (the first selection unit 4ab) determines a division threshold (for example, an intermediate value between the maximum value and the minimum value of the relative angle θ) in which the relative angle θ is predetermined among the combinations belonging to the selected group. The above combination is the highest group (first group) among the groups not yet selected in step S203. Subsequently, the controller 4 (the first selection unit 4ab), among the combinations belonging to the selected group, the combination whose relative angle θ is less than the division threshold is not selected in the current step S203. The highest group (second group). Then, the controller 4 (first selection unit 4ab) sequentially selects the groups set in step S205 from the upper group, repeatedly executes the above flow, and sets new groups for all the groups set in step S205. Reset it.

  Subsequently, the process proceeds to step S204, where the controller 4 (landmark selection unit 4a) assumes that the vehicle A has entered the center of the lane at the entrance of the map area for each map area segmented in step S102, and step S203. The relative distance L is calculated for each of the two landmarks B represented by the combination group grouped in step (b). The calculation of the relative distance L is performed based on the map information stored in the map information storage unit 2, for example.

  Subsequently, the process proceeds to step S205, where the controller 4 (second selection unit 4ac) selects a land having a long relative distance L calculated in step S204 from the combination group (combination group of group 1) grouped in step S203. The combination of the mark B and the combination of the short landmark B are distinguished and the combination of the landmark B having a long relative distance L is selected. Specifically, the controller 4 (second selection unit 4ac) selects a combination of the estimation landmarks B1 and B2 for combinations in which the relative distance L calculated in step S204 is less than a set distance (for example, 30 [m]). Exclude from And the controller 4 (2nd selection part 4ac) sets (selects) the combination after exclusion as a 1st group (2nd selection process).

  As a result, the controller 4 (second selection unit 4ac) allows the vehicle A to enter the region (map region) based on the stored contents of the map information storage unit 2 before the vehicle A enters the predetermined region (map region). In the case where it is assumed that the vehicle has entered, the combination of the landmarks B having a long relative distance L and the combination of the short landmarks B are distinguished from the combination group existing in the region (map region). A combination of landmarks B with a long (a combination that is a set distance (for example, 30 [m]) or more) is selected.

Note that if the controller 4 (second selection unit 4ac) has executed step S205 during the execution of the priority order setting process, the controller 4 (second selection unit 4ac) is not yet selected in the current step S205 from the group set in step S203. Among the unselected groups, the highest group (first group) is selected. Subsequently, the controller 4 (second selection unit 4ac) determines a division threshold (for example, the maximum value and the minimum value of the relative distance L) in which the relative distance L is predetermined among the combinations belonging to the selected group (first group). The above combination is the highest group (first group) among the groups not yet set in step S205. Subsequently, the controller 4 (second selection unit 4ac) has not yet set a combination whose relative distance L is less than the division threshold among the combinations belonging to the selected group (first group) in the current step S205. It is assumed that the highest group (second group) among the groups. Then, the controller 4 (second selection unit 4ac) selects the groups set in step S203 in order from the upper group, repeatedly executes the above flow, and sets new groups for all the groups set in step S203. Reset it.
Thereby, in 1st Embodiment, the combination whose relative distance L is less than the predetermined setting distance (for example, 20 [m]) is excluded from a combination group (step S201). Therefore, a combination in which the relative angle θ changes relatively greatly as the vehicle A progresses can be excluded in advance, so that the calculation load required for selecting the landmark B can be reduced.

  Subsequently, the process proceeds to step S206, and the controller 4 (first selection unit 4ab) determines whether the combinations grouped in step S205 belong to different groups. If the controller 4 (first selection unit 4ab) determines that the combinations belong to different groups (Yes), the process proceeds to step S207. On the other hand, if the controller 4 (first selection unit 4ab) determines that any of the combinations belong to the same group (No), the process returns to step S203.

Subsequently, the process proceeds to step S207, and the controller 4 (selection execution unit 4ad) sets the priority order for each of the combinations grouped in step S205, and then ends this calculation process. The priority is higher for the higher group.
Thereby, in the first embodiment, a combination of landmarks B having a large relative angle θ is distinguished from a combination of landmarks B having a large relative angle θ and a combination of landmarks B having a large relative angle θ. (Step S203). Also, from the combination group, the combination of the landmark B having the long relative distance L is distinguished from the combination of the landmark B having the long relative distance L, and the combination of the landmark B having the long relative distance L is selected (step S205). Then, a pair of landmarks B (estimation landmarks B1 and B2) are selected from the combination of landmarks B selected in both of them (steps S205 and S105). Therefore, it is possible to select a pair of landmarks B (estimation landmarks B1 and B2) in which both the relative angle θ and the relative distance L are relatively large. Therefore, a change in the relative angle θ between the two landmarks B (estimation landmarks B1 and B2) accompanying the traveling of the vehicle A can be suppressed. Therefore, in 1st Embodiment, it can be used for a long time, without changing a landmark, and the fluctuation | variation of the estimation precision of the present position of the vehicle A which generate | occur | produces with switching of a landmark can be suppressed.

Here, when the positional relationship between the pair of landmarks B (estimation landmarks B1 and B2) and the vehicle A changes with the progress of the vehicle A, and the detection accuracy of the current position of the vehicle A decreases, The landmark B (estimation landmarks B1, B2) needs to be switched. However, when the pair of landmarks B (estimation landmarks B1, B2) are switched, the positional relationship between the pair of landmarks B (estimation landmarks B1, B2) and the vehicle A immediately before and immediately after the switching. Therefore, the estimation result of the current position of the vehicle A may vary.
Therefore, for example, when the turning control of the vehicle A is performed using the estimation result of the current position of the vehicle A, the turning state may vary. On the other hand, in the first embodiment, since it is possible to suppress a decrease in the estimation accuracy of the current position of the vehicle A, it is possible to suppress the switching of the estimation landmarks B1 and B2, and the estimation result of the current position of the vehicle A accompanying the switching. Can be suppressed for a long time.

  Further, in the first embodiment, before the vehicle A enters a predetermined area (for example, the map area that travels ahead of the vehicle A), the area (map area) is determined based on the stored contents of the map information storage unit 2. When it is assumed that the vehicle A has entered, a combination of the landmark B having a large relative angle θ and a combination of a small landmark B are distinguished from a combination group existing in the region (map region). A combination of landmarks B with a large angle θ is selected. Subsequently, the region (map region) when it is assumed that the vehicle A has entered the region (map region) based on the storage contents of the map information storage unit 2 before the vehicle A enters the region (map region). The combination of landmarks B having a long relative distance L is selected by distinguishing the combination of landmarks B having a long relative distance L from the combination of landmarks B having a short relative distance L. Subsequently, when it is determined that the vehicle A has entered the area (map area), a pair of landmarks B (estimation landmarks B1 and B2) are selected from the combinations of the landmarks B selected in both of them. Select. Accordingly, a pair of landmarks B (estimation landmarks B1 and B2) used for estimation of the current position of the vehicle A can be selected based on the stored contents of the map information storage unit 2. Therefore, the distance and direction need only be detected for the selected pair of landmarks B, and the load for detecting the distance and direction can be reduced.

(Operation other)
Next, the operation of the vehicle A equipped with the vehicle position estimation device 1 of the first embodiment will be described.
It is assumed that the driver performs an operation for starting the current position calculation process. Then, the controller 4 reads the map information of the area where the vehicle A is predicted to travel from the map information storage unit 2 (step S101 in FIG. 2). Subsequently, as shown in FIG. 3, the controller 4 divides the area (area) into a plurality of map areas arranged in the vehicle traveling direction as shown in FIG. 3 (step S102 in FIG. 2). Subsequently, the controller 4 acquires the position of the landmark B existing in the map area for each divided map area from the read map information (step S103 in FIG. 2).

  Subsequently, the controller 4 can select from the plurality of landmarks B present in the map area based on the position of the landmark B acquired from the map information in the map information storage unit 2 for each divided map area. Then, a combination group of a pair of landmarks B is acquired (step S201 in FIG. 5). At that time, the controller 4 excludes combinations whose distance along the traveling direction of the vehicle A is equal to or less than a predetermined set distance (for example, 20 [m]) from the combination group (step S201 in FIG. 5). . For example, when the vehicle A is traveling on the road shown in FIG. 7, the combinations BaBd, BbBe, and BcBf are excluded. Thus, the controller 4 obtains the combination groups BaBb, BaBc, BaBe, BaBf, BbBc, BbBd, BbBf, BcBd, BcBe, BdBe, BdBf, and BeBf.

  Subsequently, the controller 4 detects the relative angle θ for each acquired combination group (step S202 in FIG. 5). Subsequently, as shown in FIG. 8A, the controller 4 distinguishes the combination of the landmark B having a large relative angle θ and the combination of a small landmark B from the acquired combination group, A combination of landmarks B with a large relative angle θ is selected. Specifically, the controller 4 excludes combinations BbBc, BbBf, BcBe, and BeBf in which the calculated relative angle θ is less than a set angle (for example, 30 °) (step S203 in FIG. 5). Subsequently, the controller 4 sets (selects) the combination group BaBb, BaBc, BaBe, BaBf, BbBd, BcBd, BdBe, and BdBf after the exclusion as the first group (step S203 in FIG. 5).

  Subsequently, the controller 4 calculates the relative distance L for each combination group belonging to the first group in FIG. 8A (step S204 in FIG. 5). Subsequently, as shown in FIG. 8B, the controller 4 combines the combination of the landmark B having a long relative distance L and the short landmark B out of the combinations belonging to the first group in FIG. The combination of landmarks B having a long relative distance L is selected. Specifically, the controller 4 excludes combinations BaBb, BaBe, BbBd, and BdBe whose calculated relative distance L is less than a predetermined set value (for example, 30 [m]) (step S205 in FIG. 5). Subsequently, the controller 4 sets (selects) the combination group BaBc, BaBf, BcBd, and BdBBf as the first group (step S205 in FIG. 5).

  Thus, in the first embodiment, the distance (relative distance L) along the vehicle traveling direction between two landmarks is less than a predetermined set distance (for example, 20 [m]) from the combination group. Exclude combinations. Subsequently, a combination of landmarks B with a large relative angle θ and a combination of landmarks B with a small relative angle θ are distinguished from a combination group from which combinations less than a set distance (for example, 20 [m]) are excluded. A combination of landmarks B having a large angle θ is selected (steps S201 and S203). Subsequently, the controller 4 distinguishes the combination of the landmark B having a large relative angle θ from the combination of the small landmark B from the selected combinations, and selects a combination of the landmark B having a large relative angle θ. (Step S205). Therefore, a combination in which the relative angle θ changes relatively greatly with the progress of the vehicle A can be excluded from the selection targets of the estimation landmarks B1 and B2 in advance, so that the calculation load required for selecting the landmark B can be reduced. .

  Subsequently, as shown in FIG. 8C, the controller 4 selects the first group (old first group) set in step S205, and calculates the calculated relative among the combinations belonging to the selected old first group. A combination of landmarks B having a large relative angle θ is selected by distinguishing a combination of landmarks B having a large angle θ from a combination of landmarks B having a small angle θ. Specifically, the controller 4 sets the combinations BcBd and BdBf whose calculated relative angle θ is equal to or greater than the division threshold (for example, 63 °) as the first group (step S203 in FIG. 5). Subsequently, among the combinations belonging to the selected first group (old first group), the controller 4 sets the combinations BaBc and BaBf whose relative angle θ is less than the division threshold (for example, 63 °) as the second group ( Selection) (step S203 in FIG. 5).

  Subsequently, as shown in FIG. 8D, the controller 4 selects the first group (old first group) of FIG. 8C set in step S203, and the combinations belonging to the selected old first group. Among these, the combination of the landmark B having the long relative distance L is distinguished from the combination of the landmark B having the long relative distance L, and the combination of the landmark B having the long relative distance L is selected. Specifically, the controller 4 sets (selects) a combination in which the detected relative distance L is equal to or greater than the division threshold (for example, 65 [m]) as the first group (step S205 in FIG. 5). Subsequently, the controller 4 sets (selects) a combination whose relative distance L is less than the division threshold (for example, 65 [m]) as the second group among the combinations belonging to the selected old first group (FIG. 5). Step S205). Subsequently, the controller 4 selects the second group (old second group) of FIG. 8C set in step S203, and among the combinations belonging to the selected old second group, the relative distance L is the dividing threshold value. A combination of (for example, 62.5 [m]) or more is set as the third group (step S205 in FIG. 5). Subsequently, among the combinations belonging to the selected old second group, the controller 4 sets a combination whose relative distance L is less than the division threshold (62.5 [m]) as the fourth group (step S205 in FIG. 5). ).

  Subsequently, the controller 4 determines that each grouped combination belongs to a different group (step S206 “Yes” in FIG. 5). Subsequently, as shown in FIG. 8F, the controller 4 sets the priority “1” for the combination BcBd, sets the priority “2” for the combination BdBF, and sets the priority “3” for the combination BaBc. Then, the priority “4” is set to the combination BaBf (step S207 in FIG. 5).

  Subsequently, the controller 4 selects, as a pair of landmarks B (estimation landmarks B1 and B2), the combination landmark B having the highest priority ("1") in the detected map area according to the set priority. (Step S105 in FIG. 2). Subsequently, the controller 4 determines the distance (landmark distance) between each of the selected estimation landmarks Bc and Bd and the vehicle A based on the detection result (distance and direction to the target) output by the distance detector 3. ) Is detected (step S106 in FIG. 2). Subsequently, the controller 4 detects the directions (landmark directions) of the estimation landmarks Bc and Bd with respect to the vehicle A (step S106 in FIG. 2).

  Subsequently, the controller 4 detects the positions of the estimation landmarks Bc and Bd with respect to the vehicle A based on the detected landmark distance and landmark direction (step S107 in FIG. 2). Subsequently, as shown in FIG. 4, the controller 4 obtains a comparison result between at least the positions of the estimation landmarks Bc and Bd and the positions of the estimation landmarks Bc and Bd stored in the map information storage unit 2. Based on this, the current position and direction of the vehicle A are estimated (step S108 in FIG. 2).

  As described above, in the first embodiment, based on the relative angle θ between the two landmarks B and the relative distance L between the two landmarks B, the pair of landmarks B (estimation landmarks Bc, Bd ) Is selected. Therefore, in consideration of both the relative angle between the two landmarks B and the relative distance in the vehicle traveling direction, a pair of landmarks B (estimation landmarks) in which both the relative angle θ and the relative distance L are relatively large. Bc, Bd) can be selected. Therefore, it is possible to suppress a change in the relative angle θ between the pair of landmarks B (estimation landmarks Bc and Bd) as the vehicle A travels. As a result, in the first embodiment, the same estimation landmarks Bc and Bd can be used for a long time, and the current state of the vehicle A that occurs when the pair of landmarks B (estimation landmarks Bc and Bd) are switched. Variations in position estimation accuracy can be suppressed.

  More specifically, a combination of landmarks B having a large relative angle θ and a combination of landmarks B having a large relative angle θ are distinguished from a combination group of landmarks B having a large relative angle θ. , 30 °) or more). Further, the combination of landmarks B having a long relative distance L and the combination of landmarks B having a long relative distance L are distinguished from the combination of landmarks B having a long relative distance L (set distance (30 [m]). Select the combination). Then, a pair of landmarks B (estimation landmarks Bc and Bd) are selected from the combination of the landmarks B selected by both of them.

Here, when the positional relationship between the pair of landmarks B (estimation landmarks Bc and Bd) and the vehicle A changes as the vehicle A progresses, the detection accuracy of the current position of the vehicle A decreases. Switching of the landmark B (estimation landmarks Bc, Bd) is required. However, when the pair of landmarks B (estimation landmarks Bc, Bd) are switched, the positional relationship between the pair of landmarks B (estimation landmarks Bc, Bd) and the vehicle A immediately before and immediately after the switching. Therefore, the estimation result of the current position of the vehicle A may vary.
Therefore, for example, when the turning control of the vehicle A is performed using the estimation result of the current position of the vehicle A, the turning state may vary. On the other hand, in the first embodiment, since it is possible to suppress a decrease in the estimation accuracy of the current position of the vehicle A, the switching of the estimation landmarks Bc and Bd can be suppressed, and the estimation result of the current position of the vehicle A accompanying the switching. Can be suppressed for a long time.

  In the first embodiment, before the vehicle A enters a predetermined area (map area), two landmarks B existing in the area (map area) based on the storage contents of the map information storage unit 2. Priorities for selecting a pair of landmarks B are set for each of the combination groups. Subsequently, a pair of landmarks B (estimation landmarks Bc and Bd) are selected according to the set priority order. Thereby, the priority used for estimation of the current position of the vehicle A can be set based on the stored contents of the map information storage unit 2. Therefore, the distance and direction need only be detected for the landmark B having a high priority, and the load for detecting the distance and direction can be reduced.

  In the first embodiment, the landmark selection unit 4a in FIG. 1 constitutes a landmark selection unit. Similarly, the distance direction detection unit 4b in FIG. 1 constitutes the distance direction detection unit. The landmark position detection unit 4c in FIG. 1 constitutes a landmark position detection unit. Further, the map information storage unit 2 in FIG. 1 constitutes a landmark position storage unit. Further, the landmark matching unit 4d in FIG. 1 constitutes a landmark matching unit. Further, the current position estimation unit 4e in FIG. 1 constitutes a current position estimation unit. Further, the first selection unit 4ab in FIG. 1 constitutes a first selection unit. Further, the second selection unit 4ac in FIG. 1 constitutes a second selection unit. Further, the selection execution unit 4ad in FIG. 1 constitutes a selection execution unit. Further, the combination exclusion unit 4aa in FIG. 1 constitutes a combination exclusion unit.

(Effect of 1st Embodiment)
The vehicle position estimation apparatus 1 according to the first embodiment has the following effects.
(1) According to the vehicle position estimation apparatus 1 according to the first embodiment, the controller 4 detects the positions of the plurality of landmarks B with respect to the vehicle A based on the distances and directions of the plurality of landmarks B with respect to the vehicle A. . Subsequently, the controller 4 collates the detected positions of the plurality of landmarks B with the positions of the landmarks B on the map. Subsequently, the controller 4 determines the vehicle A based on the collation result between the position of at least one pair of landmarks B (estimation landmarks B1 and B2) and the position of the landmark on the map among the plurality of landmarks B. Estimate the current position. At that time, the controller 4 is based on the angle (relative angle θ) formed by the straight lines extending from the vehicle to each of the two landmarks and the distance (relative distance L) along the vehicle traveling direction between the two landmarks. Then, a pair of landmarks B (estimation landmarks B1 and B2) are selected.

  According to such a configuration, in consideration of both the relative angle between the two landmarks and the relative distance in the vehicle traveling direction, the pair of landmarks B ( The estimation landmarks B1, B2) can be selected. Therefore, it is possible to suppress a change in the relative angle θ between the selected landmarks B (estimation landmarks B1 and B2) as the vehicle A travels. As a result, the same landmark B (estimation landmarks B1, B2) can be used for a long time, and the current position of the vehicle A is generated when the selected landmark B (estimation landmarks B1, B2) is switched. The fluctuation of the estimation accuracy can be suppressed.

(2) According to the vehicle position estimation apparatus 1 according to the first embodiment, the controller 4 determines that the relative distance L is less than the predetermined set distance (20 [m]) from the combination group of the pair of landmarks B. Are excluded from selection targets (a combination group of a pair of landmarks B).
According to such a configuration, since the combination in which the relative angle θ changes relatively greatly as the vehicle A travels can be excluded in advance, the calculation load required for selecting the landmark B can be reduced.

(3) According to the vehicle position estimation apparatus 1 according to the first embodiment, the controller 4 is based on the storage contents of the map information storage unit 2 before the vehicle A enters the predetermined area (map area). Priorities for selecting a pair of landmarks B (estimation landmarks B1 and B2) are set for each combination group of two landmarks B present in the region (map region). Subsequently, the controller 4 selects a pair of landmarks B (estimation landmarks B1 and B2) according to the set priority order.
According to such a configuration, the priority order used for estimating the current position of the vehicle A can be set based on the stored contents of the map information storage unit 2. Therefore, the distance and direction need only be detected for the landmark B having a high priority, and the load for detecting the distance and direction can be reduced.

(4) According to the vehicle position estimation apparatus 1 according to the first embodiment, the controller 4 distinguishes the combination of the landmark B having a large relative angle θ from the combination of the small landmark B so that the relative angle θ is large. A combination of landmarks B (set angle (for example, 30 ° or more)) is selected. Further, the controller 4 distinguishes the combination of the landmark B having a long relative distance L from the combination of the short landmark B, and the combination of the landmark B having a long relative distance L (set distance (for example, 30 [m]). Above). Then, the controller 4 selects a pair of landmarks B from the combination of the landmarks B selected by both of them.
According to such a configuration, a pair of landmarks B (estimation landmarks B1 and B2) in which both the relative angle θ and the relative distance L are relatively large can be selected more reliably. Therefore, a change in the relative angle θ between the selected landmarks B (estimation landmarks B1 and B2) accompanying the progress of the vehicle A can be more reliably suppressed.

(5) According to the vehicle position estimation method according to the first embodiment, the pair of landmarks for the vehicle A based on the distance and direction of the selected pair of landmarks B (estimation landmarks B1 and B2) with respect to the vehicle A. The position of B (estimation landmarks B1, B2) is detected. Subsequently, the position of the detected pair of landmarks B (estimation landmarks B1, B2) and the position of the landmark B (estimation landmarks B1, B2) on the map are collated. Subsequently, the current position of the vehicle A is estimated based on the collation result. At that time, based on the angle (relative angle θ) formed by the straight lines extending from the vehicle to each of the two landmarks and the distance (relative distance L) along the vehicle traveling direction between the two landmarks, The landmark B (estimation landmarks B1 and B2) is selected.
According to such a configuration, in consideration of both the relative angle between the two landmarks and the relative distance in the vehicle traveling direction, the pair of landmarks B ( The estimation landmarks B1, B2) can be selected. Therefore, it is possible to suppress a change in the relative angle θ between the selected landmarks B (estimation landmarks B1 and B2) as the vehicle A travels. As a result, the same landmark B (estimation landmarks B1, B2) can be used for a long time, and the current position of the vehicle A is generated when the selected landmark B (estimation landmarks B1, B2) is switched. The fluctuation of the estimation accuracy can be suppressed.

(Modification)
In the first embodiment, map information of a region where the vehicle A is predicted to travel is read from the map information storage unit 2 (step S101 in FIG. 2), and the read map information includes a plurality of regions (regions) arranged in the vehicle traveling direction. 2 (step S102 in FIG. 2), and from the read map information, the position of the landmark B existing in the map area is obtained for each divided map area (step S103 in FIG. 2). Based on the position of the mark B, a priority order is set for the combination of the landmark B used for estimating the current position of the vehicle A for each divided map area (step S104 in FIG. 2), and detection is performed according to the set priority order. The combination of landmarks B having the highest priority (“1”) in the map area is selected as a pair of landmarks B (estimation landmarks B1 and B2). 2 (step S105 in FIG. 2), based on the detection results (distance to target, direction) output from the distance detector 3, the distance between the selected estimation landmarks Bc and Bd and the vehicle A (landmark distance). ) And the respective directions (landmark directions) of the estimation landmarks Bc and Bd with respect to the vehicle A (step S106 in FIG. 2), and the estimation land for the vehicle A based on the detected landmark distance and landmark direction. The positions of the marks Bc and Bd are detected (step S107 in FIG. 2), and at least the positions of the estimation landmarks Bc and Bd and the positions of the estimation landmarks Bc and Bd stored in the map information storage unit 2 are detected. Based on the collation result, the current position and direction of the vehicle A are estimated (step S108 in FIG. 2), but the relative angle between the two landmarks B is Based on θ and the relative distance L between the two landmarks B, a pair of landmarks B (estimation landmarks Bc, Bd) are selected and at least a pair of landmarks B (estimation landmarks Bc, Any configuration may be used as long as the current position and direction of the vehicle A are estimated on the basis of the collation result between the position Bd) and the positions of the estimation landmarks Bc and Bd stored in the map information storage unit 2.

  For example, as a modification, the processing order shown in FIG. 2 can be changed to the processing order shown in FIG. Based on the detection result (distance to target, direction) output by the distance detector 3, the distance between the landmark B and the vehicle A (landmark distance) and the direction of the landmark B relative to the vehicle A ( (The landmark direction) is detected (step S106 in FIG. 9), and then the position of the landmark B with respect to the vehicle A is detected based on the map information and the detected landmark distance and landmark direction (step S107 in FIG. 9). Thereafter, based on the position of the detected landmark B, a priority is set for the combination in the detected landmark B (step S104 in FIG. 9), and then detected according to the set priority. The combination of landmarks B having the highest priority (“1”) in the selected map area is designated as a pair of landmarks B (estimation landmarks). 1 and B2) (step S105 in FIG. 9), then, at least the positions of the selected estimation landmarks Bc and Bd and the estimation landmarks Bc and Bd stored in the map information storage unit 2 A configuration of estimating the current position and direction of the vehicle A based on the collation result with the position (step S108 in FIG. 9) may be employed. In this modified example, in consideration of both the relative angle between the two landmarks B and the relative distance in the vehicle traveling direction, the pair of landmarks B (for estimation) both of the relative angle θ and the relative distance L are relatively large. Landmarks Bc, Bd) can be selected. Therefore, it is possible to suppress a change in the relative angle θ between the pair of landmarks B (estimation landmarks Bc and Bd) as the vehicle A travels. As a result, as in the first embodiment, the same estimation landmarks Bc and Bd can be used for a long time, and the vehicle A is generated when the pair of landmarks B (estimation landmarks Bc and Bd) are switched. Variations in the accuracy of estimation of the current position can be suppressed.

(Second Embodiment)
Next, a second embodiment according to the present invention will be described with reference to the drawings. In addition, about the structure similar to the said embodiment, the same code | symbol is used and the detail is abbreviate | omitted.
In the second embodiment, combinations of the landmark B existing on the left side of the vehicle A and the landmark B existing on the right side of the vehicle A are excluded from the combination groups that are candidates for the estimation landmarks B1 and B2. This is different from the first embodiment.

Specifically, the processing content of step S201 in FIG. 5 is different in the second embodiment.
In step S201, the controller 4 (landmark selection unit 4a) determines, for each map area divided in step S102, a plurality of landmarks B existing in the map area based on the position of the landmark B acquired in step S103. A combination group of a pair of landmarks B that can be selected from among them is acquired. At that time, the controller 4 (landmark selection unit 4a) excludes a combination of the landmark B present on the left side of the vehicle A and the landmark B present on the right side of the vehicle A. Note that the controller 4 (landmark selection unit 4a) has a predetermined distance (20 [m]) in which the distance (relative distance L) along the vehicle traveling direction between the two landmarks B is selected from the combination group. It is good also as a structure which excludes the following combinations.

  Thereby, the controller 4 (landmark selection unit 4a), for example, has the landmarks Bg, Bh, Bi on the right side of the vehicle A and the landmarks Bj, Bk, Bl on the left side as shown in FIG. In this case, six combinations BgBh, BgBi, BhBi, BjBk, BjBl, and BkBl are acquired as a combination group. Then, processing similar to that in the first embodiment is performed on these six combinations, and estimation landmarks B1 and B2 are selected.

  The estimation accuracy of the current position of the vehicle A increases as the angle (relative angle θ) formed by the straight lines extending from the vehicle A to the estimation landmarks B1 and B2 increases. However, when the estimation landmark B1 on the left side of the vehicle A and the estimation landmark B2 on the right side of the vehicle A are used for estimation of the current position of the vehicle A, the estimation landmark as the vehicle A advances. The relative angle θ between B1 and B2 decreases rapidly, and the estimation accuracy of the current position of the vehicle A rapidly decreases. For example, as shown in FIG. 11A, when the landmarks Bm and Bn exist on the right side of the vehicle A and the landmarks Bo and Bp exist on the left side, the vehicle A moves from the entrance to the exit of the map area. Think about the case. In this case, the relative angle θ of the combination Bm, Bp of the estimation landmark Bp existing on the left side of the vehicle A and the estimation landmark Bm existing on the right side of the vehicle A is shown by a broken line in FIG. In addition, it decreases rapidly after reaching 180 °.

  Therefore, if the combination of the estimation landmark B1 existing on the left side of the vehicle A and the estimation landmark B2 existing on the right side of the vehicle A is used, the detection accuracy of the current position of the vehicle A can be improved over time. Variations can increase. Therefore, the controller 4 (landmark selection unit 4a) excludes the combination of the estimation landmark B1 existing on the left side of the vehicle A and the estimation landmark B2 existing on the right side of the vehicle A. For example, the relative angle θ of the combination Bo and Bp of the estimation landmarks Bo and Bp existing on the left side of the vehicle A does not reach 180 ° as shown by the solid line in FIG. Hold the angle for a long time.

  Thus, in 2nd Embodiment, the combination of the landmark B which exists in the left side of the vehicle A, and the landmark B which exists in the right side of the vehicle A is excluded from a combination group (step S201). Therefore, combinations in which the relative angle θ changes relatively greatly as the vehicle A progresses can be excluded. Therefore, the change in the relative angle θ between the two landmarks B can be more appropriately suppressed. Therefore, it is possible to more appropriately suppress a decrease in the estimation accuracy of the current position of the vehicle A.

(Effect of 2nd Embodiment)
The vehicle position estimation device 1 according to the second embodiment has the following effects.
(1) According to the vehicle position estimation device 1 according to the second embodiment, the controller 4 selects a combination of the landmark B present on the left side of the vehicle A and the landmark B present on the right side of the vehicle A. Excluded from (combination group of a pair of landmarks B).
According to such a configuration, combinations in which the relative angle θ changes relatively greatly as the vehicle A progresses can be excluded. Therefore, the change in the relative angle θ between the two landmarks B can be more appropriately suppressed. Therefore, it is possible to more appropriately suppress a decrease in the estimation accuracy of the current position of the vehicle A.

2 Map information storage unit (landmark position storage unit)
4a Landmark selection unit 4aa Combination exclusion unit 4ab First selection unit 4ac Second selection unit 4ad Selection execution unit 4b Distance direction detection unit 4c Landmark position detection unit 4d Landmark collation unit 4e Current position estimation unit

Claims (7)

A distance direction detection unit for detecting distances and directions of a plurality of landmarks existing around the vehicle with respect to the vehicle;
A landmark position detection unit that detects positions of the plurality of landmarks with respect to the vehicle based on the distance and direction detected by the distance direction detection unit;
A landmark position storage unit storing the positions of landmarks on the map;
A landmark collating unit for collating the positions of the plurality of landmarks detected by the landmark position detecting unit with the positions of the landmarks on the map;
A landmark selection unit for selecting a pair of landmarks from the plurality of landmarks;
A current position estimation unit that estimates a current position of the vehicle based on a comparison result between at least the position of the pair of landmarks and the position of the landmark on the map;
The landmark selection unit is based on a relative angle that is an angle formed by straight lines extending from the vehicle to each of the two landmarks, and a relative distance that is a distance along the vehicle traveling direction between the two landmarks. A vehicle position estimation device that selects the pair of landmarks.   2. The vehicle position according to claim 1, wherein the landmark selection unit excludes a combination of a landmark present on the left side of the vehicle and a landmark present on the right side of the vehicle from the selection target. Estimating device.   The vehicle position estimation apparatus according to claim 1, wherein the landmark selection unit excludes a combination of landmarks whose relative distance is less than a predetermined set distance from the selection target. When selecting the pair of landmarks for each combination group of two landmarks existing in the area based on the stored contents of the landmark position storage unit before the vehicle enters the predetermined area. A priority setting unit for setting the priority of
4. The vehicle position estimation according to claim 1, wherein the landmark selection unit selects the pair of landmarks according to the priority set by the priority setting unit. 5. apparatus.   The distance direction detection unit detects a distance and a direction of the pair of landmarks of the combination having the highest priority among the detected landmarks according to the priority set by the priority setting unit. The vehicle position estimation apparatus according to claim 4. The landmark selection unit
A first selection unit that distinguishes between a combination of landmarks with a large relative angle and a combination of landmarks with a small relative angle and selects a combination of landmarks with a large relative angle;
A second selection unit that distinguishes a combination of landmarks having a long relative distance from a combination of short landmarks and selects a combination of landmarks having a long relative distance;
2. A selection execution unit that selects the pair of landmarks from a combination of landmarks selected by both the first selection unit and the second selection unit. The vehicle position estimation device according to any one of claims 5 to 6. The distance and direction of the plurality of landmarks existing around the vehicle are detected, the positions of the plurality of landmarks with respect to the vehicle are detected based on the detected distance and direction, and the detected plurality of landmarks are detected. The position is compared with the position of the landmark on the map, a pair of landmarks is selected from the plurality of landmarks, and at least the position of the pair of landmarks and the position of the landmark on the map Based on the matching result, the current position of the vehicle is estimated,
Based on a relative angle that is an angle formed by straight lines extending from the vehicle to each of the two landmarks and a relative distance that is a distance along the vehicle traveling direction between the two landmarks, the pair of landmarks is A vehicle position estimation method comprising: selecting a vehicle position.

Images