VEHICLE POSITION ESTIMATE DEVICE

Abstract

Vehicle position estimation device comprising the following: - an image information reference unit (4) configured to obtain image information obtained by taking an image of a road in front of a vehicle; - a recognition unit (5) for applied lines, configured to recognize relative positions of left and right applied lines in relation to the vehicle based on the image information obtained from the image information reference unit (4), wherein the left and right applied lines are applied to the road; - a processing unit (6, 15) for detection results of the applied line, configured to calculate whether the vehicle has made a lane change, and a direction of the lane change if the vehicle has made the lane change, wherein the relative positions of the left and right sides detected by the detection unit (5) for applied lines are positions of left and right dividing lines defining a lane of the road; - a map information reference unit (11) configured to obtain map information which includes at least one lane connection relation which is a lane connection relation in one direction of travel of the road; - a vehicle lateral position calculation unit (7) configured to calculate a lane where the vehicle is in motion and a vehicle lateral position, which is a lateral position of the vehicle in the lane, based on whether the vehicle has made a lane change, as well as the direction of the lane change, which has been calculated by the processing unit (6, 15) for recognition results of the applied line, and the lane connection relation obtained from the map information reference unit (11); - a calculation unit (10) for vehicle advance distances, configured to calculate an advance distance, which is a distance by which the vehicle has moved forward; and - a calculation unit (8) for vehicle positions on the map, configured to calculate a position of the vehicle on the map on the road, based on the map information obtained from the map information reference unit (11), the lane where the vehicle is traveling, the vehicle lateral position calculated by the vehicle lateral position calculation unit (7), and the advance distance calculated by the vehicle advance distance calculation unit (10), wherein the processing unit (15) for recognition results of the plotted line comprises: a continuity correction unit (16) for applied lines, configured to perform a correction so that the positions of the left and right applied lines are temporarily made continuous, based on the relative positions of the left and right sides, detected by the detection unit (5) for applied lines, a determination unit (17) for a lane change state, configured to determine a lane change state of the vehicle, which includes the direction of the lane change, based on the relative positions of the left and right applied lines, corrected by the continuity correction unit (16) for applied lines, and past position of the left and right dividing lines, a calculation unit (18) for division line positions, configured to calculate the positions of the left and right division lines based on the relative positions of the left and right applied lines, corrected by the continuity correction unit (16) for applied lines, and the lane change state of the vehicle, determined by the determination unit (17) for a lane change state, and an accumulation unit (19) for time series data, configured to accumulate as time series data the position of the left and right division lines calculated by the calculation unit (18) for division line positions, wherein the map information reference unit (11) includes a lane width which is the width of the lane, and wherein the continuity correction unit (16) for applied lines determines that the right-hand drawn line corresponds to the left-hand division line at a present time if a difference between the position of the left-hand division line, which is a division line adjacent to the left side of the vehicle, referenced to a predetermined time prior to the present time, accumulated by the time-series data accumulation unit (19), and the position of the right-hand drawn line, which is a drawn line adjacent to the right side of the vehicle at the present time, detected by the drawn line detection unit (5), lies within a predetermined distance, and calculates the position of the right-hand division line, which is a division line adjacent to the right side of the vehicle at the present time, based on the left-hand division line at the present time and the lane width contained in the map information, and determines that the left applied line corresponds to the right applied line at a present time if a difference between the position of the right division line, referenced to a predetermined time prior to the present time, accumulated by means of the time series data accumulation unit (19), and the position of the left applied line, which is an applied line adjacent to the left side of the vehicle at the present time, detected by means of the applied line detection unit (5), lies within a predetermined distance, and calculates the position of the left division line at the present time based on the right division line at the present time and the lane width.

Description

Technical area

[0001] The present invention relates to a vehicle position estimating device that estimates a current position of a vehicle. State of the art

[0002] Generally, navigation systems keep vehicle position estimation coordinates obtained by dead reckoning with global navigation satellite (GNSS) coordinate information received from satellites, as well as vehicle speed signal information calculated from a wheel speed of a vehicle, acceleration information obtained from a gyro sensor, and the like taken into account.

[0003] Additionally, when using a navigation map that displays roads only with links, nodes, and attribute information associated with the links and nodes, a navigation system calculates a link of a road where the vehicle is traveling and a vehicle position on the link through map matching, thus determining a display position of the vehicle position on the map. Note that the vehicle position on the map is also referred to as a vehicle marker.

[0004] The navigation map described above does not include detailed lane-level shape information, and it has an entire road including a plurality of lanes represented by line segments, curves, or the like. Therefore, such a navigation map does not provide map information used to estimate a lane, where the vehicle is traveling, or a relative lateral position of the vehicle in the lane through map matching. The lateral direction of the vehicle is the direction perpendicular to the vehicle's traveling direction.

[0005] In general, the estimation accuracy of the vehicle position obtained by dead reckoning is susceptible to satellite reception conditions such as multipath propagation, and an error of approximately 10 m occurs depending on the reception condition. Therefore, even if the detailed shape of each lane is already known, it is less likely that an actual traveling lane and a relative position of the vehicle in the lateral direction in the lane can be accurately estimated based on a comparison of the dead reckoning result with the detailed shape of each lane.

[0006] As described above, it is difficult to obtain the lateral position of the vehicle, which is necessary for estimating the traveling lane, only from the vehicle position calculated by dead reckoning, which uses GNSS coordinate information, vehicle speed signal information, and angular velocity information together. In response, a technology for identifying the vehicle position by simultaneously using the result of analyzing image information acquired by a camera on the vehicle has been developed.

[0007] Analysis of the image information captured by the on-vehicle camera makes it possible to detect a plotted line, which represents a line plotted on a road corresponding to each of the right and left dividing lines of the traveling lane. Hereinafter, among the dividing lines that define the traveling lanes of the road, a dividing line adjacent to the left side of the vehicle—that is, a line forming the boundary between the traveling lane and a lane located immediately to the left of the traveling lane—is referred to as a left dividing line.

[0008] In addition, a dividing line adjacent to the right side of the vehicle—that is, a line forming the boundary between the traveling lane and a lane immediately to the right of the traveling lane—is called a right dividing line. The color and shape applied to the dividing line vary from country to country or region to region. For example, in Japan, white and yellow are often used as the colors for the dividing line.

[0009] The vehicle position estimation system enables highly accurate estimation of the vehicle's position in the lane based on the relative position of the plotted line with respect to the vehicle and the absolute position of the dividing line contained in the map information. The accuracy of the position of a plotted line detected by a camera is generally on the order of a few tens of centimeters, and therefore, the accuracy of a vehicle's lateral position can also be expected to be on the order of a few tens of centimeters.

[0010] The above-described plotted line recognition technology using a camera has been increasingly applied to various driver assistance functions, and a universal camera with plotted line recognition capability designed for use in an advanced driver assistance system (ADAS) has been developed. The plotted line recognition capability of such a camera is generally based on a specification in which two plotted lines immediately adjacent to the vehicle are detected, that is, plotted lines immediately adjacent to the left and right sides of the vehicle, and then the two plotted lines thus detected are represented with a polynomial, that is, a curved line, in an orthogonal coordinate system based on the position and traveling direction of the vehicle, and then output.

[0011] It is believed that the specification in which two plotted lines are output is intended to apply the camera's plotted line detection capability to a lane keeping capability.

[0012] The accuracy of detecting the plotted line using the camera varies depending on the driving conditions around the vehicle. Initially, it is expected that both the left and right dividing lines would be detected, but often only a single dividing line is detected, for example, due to a contrast difference on the road caused by the road structure, insufficient lighting in a tunnel, a plotted line obscured by a parallel vehicle, smearing or blurring of the plotted line, or the like.

[0013] To deal with such uncertainty in detecting a plotted line using a camera, a technology for estimating a traveling direction of a vehicle is known in which correction is performed based on a transition pattern of a pre-predicted line type for a line type of a detected plotted line and time series data of the thus-detected line type (see, for example, Patent Document 1). Examples of the line type of the plotted line include a branch line, a solid line, and a dashed line. The branch line has a larger width and a shorter dashed line length than a typical dividing line located between lanes. Examples of the traveling direction of the vehicle include a branch direction and a main line direction.

[0014] In Patent Document 1, when a branch line is detected in front of the vehicle based on map data, a camera image obtained by capturing an area in front of the vehicle is divided into three areas in the lateral direction, and plotted line detection is performed for each of the areas. Then, the line type of the plotted line thus detected is corrected based on a comparison of the plotted line detection result for each of the areas with a line type transition pattern predicted in advance in each of the areas.

[0015] Patent Document 2 discloses a method for determining transverse position information of a motor vehicle on a roadway. The method comprises the following steps: recording radar data describing at least a portion of the roadway with at least one radar sensor of the motor vehicle; detecting and locating environmental features describing the position of a roadway boundary by evaluating the radar data; determining a course of the roadway boundaries and lateral distances of the motor vehicle from the roadway boundaries from the environmental features; and determining the transverse position information as or as a function of the lateral distances of the motor vehicle from the roadway boundaries.

[0016] Patent Document 3 discloses a navigation device that includes a device for detecting the position of the vehicle and a direction of travel of the vehicle. Furthermore, the navigation device includes a device for detecting a lane marking line in an image of a road located in front of or behind the vehicle, captured by a camera.

[0017] From Patent Document 4, a lane departure prevention control device for a vehicle is known, which comprises: an assist control device that determines whether the vehicle deviates from a lane in which the vehicle is traveling, and, when it is determined that the vehicle deviates from the lane, executes an assist control.The support control device includes: a first support control device that calculates a predicted lane departure time until the vehicle deviates from the lane in the future based on a vehicle lateral position in a width direction of the lane in which the vehicle is traveling, a yaw angle of the vehicle with respect to the lane, and the vehicle speed, and determines the lane departure from the predicted lane departure time; and a second support control device that determines whether the vehicle deviates from the lane when the yaw angle of the vehicle with respect to the lane points in a direction deviating from the lane.The assist control means selects the first assist control means when the vehicle lateral position is closer to a lane center than a predetermined position set according to a traveling environment, and selects the second assist control means when the vehicle lateral position is closer to a lane edge than the predetermined position. State of the artPatent documents Patent Document 1: Japanese Patent Application Laid-Open No. JP 2008-102684 A Patent document 2: DE 10 2015 001 386 A1 Patent document 3: DE 11 2005 001 307 T5 Patent document 4: DE 10 2015 104 472 A1 SummaryProblem to be solved with the invention

[0018] Patent Document 1 assumes that the plotted line detection result is used in each of the regions resulting from dividing the camera image into three regions in the lateral direction. That is, in Patent Document 1, up to three plotted lines must be detected. Therefore, the use of only two plotted lines immediately adjacent to the vehicle is not possible based on the standard output specification of the plotted line detection result applied to the driver assistance function described above. Furthermore, in Patent Document 1, it is necessary to use a dedicated plotted line detection module capable of detecting three plotted lines, and thus, it is difficult to apply such a module to a wider range of products.

[0019] Furthermore, in Patent Document 1, the correction target is only the line type of the plotted line detected by the camera, and no description is given regarding the detection position of the plotted line or the presence or absence of the plotted line. Therefore, applying Patent Document 1 directly to the vehicle position estimation technology does not contribute to an increase in vehicle position accuracy.

[0020] The present invention was conceived to solve such problems. It is therefore an object of the present invention to provide a vehicle position estimation device capable of estimating a vehicle position with high accuracy. Ways to solve the problem

[0021] The problem is solved by a vehicle position estimation device having the features of claim 1. Advantageous further developments emerge from subclaims 2 to 6. Effects of the invention

[0022] According to the present invention, the vehicle position estimation device comprises: an image information acquisition unit configured to acquire image information obtained by capturing an image of a road ahead of a vehicle; the plotted line recognition unit configured to recognize relative positions of left and right plotted lines with respect to the vehicle based on the image information acquired by the image information acquisition unit, the left and right plotted lines being plotted on the road; the plotted line recognition result processing unit configured to calculate whether the vehicle has made a lane change and a direction of the lane change when the vehicle has made the lane change, the relative positions of the left and right sides,which are recognized by the plotted line recognition unit are positions of left and right dividing lines defining a lane of the road; the map information acquisition unit configured to acquire map information including at least a lane connection relation, which is a connection relation of the lane in a traveling direction of the road; the vehicle lateral position calculation unit configured to calculate a lane where the vehicle is traveling and a vehicle lateral position, which is a lateral position of the vehicle in the lane, based on whether the vehicle has completed the lane change, the direction of the lane change calculated by the plotted line recognition result processing unit, and the lane connection relation acquired by the map information acquisition unit;the vehicle ahead distance calculation unit configured to calculate a ahead distance, which is a distance by which the vehicle has moved forward, and the vehicle map position calculation unit configured to calculate a position of the vehicle on the map based on the map information acquired by the map information acquisition unit, the lane where the vehicle is traveling, and the vehicle lateral position calculated by the vehicle lateral position calculation unit, as well as the ahead distance calculated by the vehicle ahead distance calculation unit. Therefore, it is possible to estimate the vehicle position with high accuracy.

[0023] The objects, features, aspects and advantages of the present invention will become more apparent from the following detailed description and the accompanying drawings. Short description of the drawings Fig. 1 is a block diagram showing an example of a structure of a vehicle position estimation system according to a first embodiment of the present invention. Fig. 2 is a diagram for describing the operation of a drawn line detection unit according to the first embodiment of the present invention. Fig. 3 is a diagram showing time series data of plotted line positions detected by the plotted line detection unit according to the first embodiment of the present invention. Fig. 4 is a diagram showing the time series data of the plotted line positions detected by the plotted line detection unit according to the first embodiment of the present invention. Fig. 5 is a flowchart showing an example of the operation of a vehicle lateral position calculation unit according to the first embodiment of the present invention. Fig. 6 is a flowchart showing an example of the operation of a vehicle position calculation unit on the map according to the first embodiment of the present invention. Fig. 7 is a diagram showing an example of a lane connection relation according to the first embodiment of the present invention. Fig. 8 is a diagram showing an example of the lane connection relation according to the first embodiment of the present invention. Fig. 9 is a diagram for explaining the operation of a vehicle position estimating apparatus according to the first embodiment of the present invention. Fig. 10 is a block diagram showing an example of a structure of a vehicle position estimation system according to a second embodiment of the present invention. Fig. 11 is a flowchart showing an example of the operation of a plotted line continuity correction unit according to the second embodiment of the present invention. Fig. 12 is a flowchart showing an example of an operation of a lane change state determining unit according to the second embodiment of the present invention. Fig. 13 is a diagram for describing a calculation order of local variables in a plotted line recognition result processing unit according to the second embodiment of the present invention. Fig. 14 is a block diagram showing an example of a structure of a vehicle position estimation system according to a third embodiment of the present invention. Fig. 15 is a flowchart showing an example of the operation of a position integration unit on the map according to the third embodiment of the present invention. Fig. 16 is a diagram for explaining the operation of a vehicle position estimating apparatus according to the third embodiment of the present invention. Description of embodiments

[0024] Embodiments of the present invention will be described below with reference to the drawings. First embodiment structure

[0025] Fig. 1 is a block diagram showing an example of the structure of a vehicle position estimation system 1 according to a first embodiment of the present invention. The vehicle position estimation system 1 estimates a vehicle position on a road where a mobile object, such as a moving vehicle, is located. The following description assumes that the mobile object is a vehicle.

[0026] The vehicle position estimation system 1 includes a vehicle position estimator 2, a road imaging device 3, and a satellite positioning device 12. The vehicle position estimator 2 includes an image information acquisition unit 4, a plotted line recognition unit 5, a plotted line recognition result processing unit 6, a vehicle lateral position calculation unit 7, a vehicle position on the map calculation unit 8, a vehicle position on the map initialization unit 9, a vehicle ahead distance calculation unit 10, and a map information acquisition unit 11.

[0027] The road imaging device 3 is, for example, a monocular camera or a stereo camera, and is configured to capture an image of an area in front of the vehicle. The road imaging device 3 captures an image of an area in the traveling direction of the vehicle, ie, a plotted line or a road sign located in front of the vehicle. The image information acquisition unit 4 acquires image information captured by the road imaging device 3.

[0028] The plotted line recognition unit 5 analyzes the image information acquired by the image information acquisition unit 4, thus detecting a plotted line presumed to be a dividing line marking on a road. Furthermore, the plotted line recognition unit 5 identifies, among the plotted lines thus detected, plotted lines adjacent to the vehicle, preferably plotted lines adjacent to the left and right sides of the vehicle and approximately parallel to the vehicle, and calculates relative positions of the plotted line with respect to the vehicle.That is, the plotted line recognition unit 5 recognizes the relative positions of the left and right plotted lines with respect to the vehicle on the basis of the image information acquired from the image information acquisition unit 4, the left and right plotted lines being plotted on the road.

[0029] The plotted line detection result processing unit 6 calculates whether the vehicle has made a lane change based on time-series changes in the position of the two plotted lines detected by the plotted line detection unit 5. If the vehicle has made a lane change, it calculates the direction of the lane change. The plotted line detection result processing unit 6 also calculates the positions of the left and right dividing lines of a lane where the vehicle is traveling.

[0030] The vehicle ahead distance calculation unit 10 calculates a ahead distance of the vehicle based on a detection result such as the tire rotation speed or the tire rotation angle when the vehicle moves forward.

[0031] The map information acquisition unit 11 is a hard disk drive (HDD), a memory card, or the like, and stores high-precision map information including a detailed shape of each lane. In the high-precision map information, the shapes of a lane centerline and dividing lines of each lane are represented with a point sequence. Furthermore, the high-precision map information includes a lane width at each point on the lane centerline, as well as a lane connection relation, which is a connection relation of the lane in a traveling direction of the road. Note that the lane width can be calculated from the point sequence representing the shape of the dividing line. The map information acquisition unit 11 can be configured to acquire the high-precision map information externally as needed.

[0032] The vehicle lateral position calculation unit 7 calculates, based on a recognition result from the plotted line recognition result processing unit 6 and the lane connection relation stored in the map information acquisition unit 11, a vehicle lateral position, which is a lateral position of the vehicle in a lane where the vehicle is traveling.

[0033] The vehicle map position calculation unit 8 updates a vehicle lateral position on the map upon notification of the vehicle lateral position from the vehicle lateral position calculation unit 7, and updates a vehicle position on the map in the forward-backward direction upon notification of the vehicle's forward travel distance from the vehicle forward travel distance calculation unit 10. The vehicle position in the forward-backward direction is a position along the vehicle's travel direction and is also referred to as a vehicle longitudinal position.

[0034] The satellite positioning device 12 is connected to a GNSS antenna for receiving radio waves from satellites, and measures the current position of the vehicle, including the latitude and longitude, based on the radio waves from the satellites. The vehicle map position initialization unit 9 notifies the vehicle map position calculation unit 8 of the current position measured by the satellite positioning device 12 to locate the vehicle on the map for the first time.

[0035] It should be noted that the description is based on Fig. 1, wherein the satellite positioning device 12 is arranged outside the vehicle position estimation device 2; however, the present invention is not limited to such a structure. For example, the satellite positioning device 12 may also be provided inside the vehicle position estimation device 2. OperationOperation of the detection unit 5 for applied lines

[0036] The plotted line recognition unit 5 detects, from an image representing an area ahead of the vehicle acquired by the image information acquisition unit 4, a plurality of plotted lines assumed to be dividing line markers on the road. Then, the plotted line recognition unit 5 identifies, among the plurality of plotted lines thus detected, two plotted lines assumed to be dividing lines immediately adjacent to the left and right sides of the vehicle, and calculates a relative positional relationship between each of the two plotted lines and the vehicle.

[0037] More precisely: As in Fig. As shown in Figure 2, the plotted line recognition unit 5 represents the shapes of two plotted lines with a one- or multi-dimensional polynomial in a two-dimensional orthogonal coordinate system, where the vehicle position is the origin, the forward direction of the vehicle is the X-axis direction, and the right direction of the vehicle is the y-axis direction. That is, the two plotted lines are represented by the following equations (1) and (2), respectively. y=L0+L1⋅x+fL(x) y=R0+R1⋅x+fR(x)

[0038] In equations (1), (2) f L (x) and f R(x) are each a two- or multi-dimensional polynomial and are used as error terms in the present embodiment. Furthermore, the lateral positions of the two plotted lines relative to the vehicle are denoted by L0 and R0, respectively, and the gradients of the two plotted lines relative to the vehicle's direction of travel are denoted by L1 and R1, respectively.

[0039] Hereinafter, L0 denotes a lateral position of a left plotted line, R0 denotes a lateral position of a right plotted line, L1 denotes a slope of the left plotted line, and R1 denotes a slope of the right plotted line. Although the origin is the vehicle position, the origin is more specifically a position where the road imaging device 3 is installed. Operation of the processing unit 6 for recognition results of the applied line

[0040] The plotted line detection result processing unit 6 calculates whether the vehicle has completed a lane change based on the lateral positions of the two plotted lines calculated by the plotted line detection unit 5. If the vehicle has completed the lane change, it calculates the direction of the lane change. Whether a lane change has been completed is indicated by "No," "Yes," and "Undetermined."

[0041] "No" means that no lane change has occurred. "Yes" means that a lane change has occurred. "Undetermined" means that an error occurred in determining whether a lane change has occurred based on the lateral positions of the two plotted lines calculated by the plotted line detection unit 5.

[0042] Next, the calculation of whether a lane change has been completed and the direction of the lane change are described.

[0043] The positions of the left plotted line and the right plotted line at a current time (t) calculated by the plotted line recognition unit 5 are represented by the following equations (3), (4), respectively, and the positions of the left plotted line and the right plotted line at a previous time (t - 1) are represented by the following equations (5), (6), respectively. y=L0,t+L1,t⋅x+fL,t(x) y=R0,t+R1,t⋅fR,t(x) y=L0,t−1+L1,t−1⋅x+fL,t−1(x) y=R0,1−t+R1,t−1⋅x+fR,t−1(x)

[0044] In equations (3) to (6), L 0,tthe lateral position of the left plotted line at a time t, which is calculated by the plotted line recognition unit 5, and R 0,t denotes the lateral position of the right plotted line at time t, calculated by the plotted line recognition unit 5. L 1,t denotes the slope of the left plotted line at time t, calculated by the plotted line recognition unit 5, and R 1,t denotes the slope of the right plotted line at time t, calculated by the plotted line detection unit 5.

[0045] In addition, f L,t (x) and f R,t(x) a two- or more-dimensional polynomial of the left plotted line at time t calculated by the plotted line recognition unit 5, and a two- or more-dimensional polynomial of the right plotted line at time t calculated by the plotted line recognition unit 5, respectively, and they are used as error terms in the present embodiment.

[0046] For example, if both of the following inequalities (7), (8) are satisfied, it can be recognized that both the left plotted line and the right plotted line have moved to the right by a distance equivalent to one lane, as in Fig. 3. This causes the plotted line recognition result processing unit 6 to determine that a lane change has been made to the right lane, causing "whether a lane change has been made" to be set to "Yes" and the lane change direction to be set to "right." (L0,t−L0,t−1)>θLchange∗Widt (R0,t−R0,t−1)>θLchange∗Widt

[0047] In inequalities (7), (8), Widt denotes the lane width, which may be the width of a lane near the current point acquired by the map information acquisition unit 11, or it may be calculated from the following equation (9), which takes into account the difference between the lane width captured by the camera and the actual lane width. In addition, θ denotes Lchangea parameter for allowing an error in the lane width, and it is preferably set to less than 1 and close to 1, for example to 0.7. Widt=(R(D)0,t−1−L(D)0,t−1)∗(WMap,t / WMap,t−1)

[0048] In equation (9) W Map,t the lane width of a lane where the vehicle is traveling, obtained from the map information obtaining unit 11 at the current time (t), and W Map,t-1 denotes the lane width of a lane where the vehicle is traveling, obtained from the map information acquisition unit 11 at the previous time (t - 1).

[0049] Similarly, if both of the following inequalities (10), (11) are satisfied, it can be seen that both the left plotted line and the right plotted line have moved to the left by a distance equivalent to one lane, as in Fig. 4. This causes the plotted line recognition result processing unit 6 to determine that a lane change has been made to the left lane, so that “whether a lane change has been made” is set to “Yes” and the direction of the lane change is set to “left”. (L0,t−L0,t−1)<θLchange∗Widt (R0,t−L0,t−1)<θLchange∗Widt

[0050] If it cannot be determined whether a lane change has occurred, "whether a lane change has occurred" is marked "Undetermined." Specific examples of a case where it cannot be determined whether a lane change has occurred include the following: (A) the case where the plotted line recognition unit 5 could not recognize one or both of the two plotted lines; and (B) the case in which a change in the left plotted line and a change in the right plotted line are not synchronized. A determining criterion for the case of (B) can, for example, be a case in which both of the following inequalities (12) and (13) are satisfied. ((L0,t−L0,t−1)−θLchange∗Widt)∗((R0,t−R0,t−1)−θLchange∗Widt)<0 ((L0,t−L0,t−1)+θLchange∗Widt)∗((R0,t−R0,t−1)+θLchange∗Widt)<0

[0051] The lateral position of a left dividing line corresponds to the lateral position L 0,t the left plotted line detected by the plotted line detection unit 5 at the current time (t), and the lateral position of a right dividing line corresponds to the lateral position R 0,t the right plotted line detected by the plotted line detection unit 5 at the current time (t). Operation of the vehicle lateral position calculation unit 7

[0052] Fig. 5 is a flowchart showing an example of the operation of the vehicle lateral position calculation unit 7.

[0053] In step S101, the vehicle lateral position calculation unit 7 waits until it is notified of a position calculation result of the dividing line from the plotted line recognition result processing unit 6 or the lane connection relation from the vehicle position calculation unit 8 on the map, and proceeds to step S102 upon notification of either the position calculation result of the dividing line or the lane connection relation.

[0054] In step S102, the vehicle lateral position calculation unit 7 determines which of the position calculation results of the dividing line and the lane connection relation was notified. If the position calculation result of the dividing line has been notified, the vehicle lateral position calculation unit 7 proceeds to step S103. If the lane connection relation has been notified, the vehicle lateral position calculation unit 7 proceeds to step S106.

[0055] In step S103, if the position calculation result of the dividing line notified by the plotted line recognition result processing unit 6 indicates that “whether a lane change has been made” is “Yes,” the vehicle lateral position calculation unit 7 increases or decreases a lane number held therein by 1. If the position calculation result of the dividing line indicates that “whether a lane change has been made” is “No” or “Undetermined,” the vehicle lateral position calculation unit 7 makes the lane number identical to the lane number at the previous time point.

[0056] For example, if the occupied lane number immediately before the dividing line position calculation result is notified is 2, and the thus-notified dividing line position calculation result indicates that "whether a lane change has been completed" is "yes" and the lane change direction is "left," the occupied lane number is updated to 1. The occupied lane number of 1 indicates that the vehicle is traveling in a first lane of the road. The occupied lane number of 2 indicates that the vehicle is traveling in a second lane of the road.

[0057] In step S104, the vehicle lateral position calculation unit 7 calculates the vehicle lateral position in a lane from the position of the left dividing line and the position of the right dividing line, as transmitted from the plotted line recognition result processing unit 6. In step S105, the vehicle lateral position calculation unit 7 updates the vehicle lateral position held therein to the vehicle lateral position calculated in step S104. More specifically, for example, the vehicle lateral position Y car in the lane is calculated from the following equation (14). Ycar=(−L(D)0,t) / (R(D)0,t−1−L(D)0,t−1)

[0058] In equation (14) L (D) 0,t the lateral position of the left dividing line at time (t) as determined by the plotted line recognition result processing unit 6, and R (D) 0,tdenotes the lateral position of the right dividing line at time (t), as determined by the plotted line recognition result processing unit 6. In addition, Y car = 0, that the vehicle moves on the left plotted line, Y car = 1 means that the vehicle is moving on the right plotted line, and Y car = 0.5 means that the vehicle is moving in the middle of the lane.

[0059] In step 106, the vehicle lateral position calculation unit 7 updates the lane number based on the lane connection relation notified by the vehicle position calculation unit 8 on the map. The lane connection relation thus transmitted includes a connection relation between a lane configuration at the previous travel point and a lane configuration at the current travel point.

[0060] The Fig. 7 and Fig. 8 show an example of the lane connection relation. In the example according to Fig. 7, the first lane in use at the previous travel point is connected to the second lane at the current travel point. The second lane in use at the previous travel point is connected to the third lane at the current travel point. Fig. Figure 8 shows the correspondence relations of the above connections. For example, if the lane connection relation shown in Fig. 7 and Fig. 8, in a case where the previous traveled lane number is 1, the traveled lane number is updated to 2. Operation of the calculation unit 8 for vehicle positions on the map

[0061] Fig. 6 is a flowchart showing an example of the operation of the vehicle position calculation unit 8 on the map.

[0062] In step S201, the map vehicle position calculation unit 8 waits until it is notified of the ahead distance from the vehicle ahead distance calculation unit 10 or the vehicle lateral position from the vehicle lateral position calculation unit 7, and proceeds to step S202 upon notification of either the ahead distance or the vehicle lateral position.

[0063] In step S202, the vehicle position calculation unit 8 determines the vehicle position on the map that has been notified from the ahead distance and the vehicle lateral position. If the ahead distance has been notified, the vehicle position calculation unit 8 proceeds to step S203. If the vehicle lateral position has been notified, the vehicle position calculation unit 8 proceeds to step S206.

[0064] In step S203, the vehicle position calculation unit 8 advances the vehicle position with the vehicle lateral position unchanged, and updates a vehicle position on the map held therein to the vehicle position thus advanced.

[0065] In step S204, the map vehicle position calculation unit 8 obtains the lane connection relations at the vehicle positions before and after the forward movement from the map information acquisition unit 11. Then, in step S205, the map vehicle position calculation unit 8 notifies the vehicle lateral position calculation unit 7 of the lane connection relations obtained from the map information acquisition unit 11.

[0066] In step S206, the vehicle position on the map calculation unit 8 moves the vehicle position in the lateral direction within the road so as to make the vehicle position identical to the vehicle lateral position thus notified with the vehicle longitudinal position remaining unchanged, and updates a vehicle position on the map held therein to the vehicle position thus moved in the lateral direction.

[0067] From the above, the map vehicle position calculation unit 8 adds the ahead distances obtained from the map vehicle ahead distance calculation unit 10 to the initial position given by the map vehicle position initialization unit 9, thus calculating the current vehicle longitudinal position on the map. Furthermore, the map vehicle position calculation unit 8 identifies the current vehicle lateral position on the map based on the vehicle lateral position obtained from the vehicle lateral position calculation unit 7. Note that when the vehicle reaches a branch point, the vehicle's traveling direction is identified based on a lane where the vehicle is traveling and the position in the lane. Operation of the initialization unit 9 for vehicle positions on the map

[0068] The vehicle map position initialization unit 9 notifies the vehicle map position calculation unit 8 of the absolute coordinates serving as a basis when the vehicle starts traveling. For example, if satellite positioning becomes available immediately after the vehicle starts traveling, the vehicle map position initialization unit 9 may notify the vehicle map position calculation unit 8 of a satellite positioning result obtained from the satellite positioning device 12 only once.

[0069] Note that when the division line position calculation result notified by the plotted line recognition result processing unit 6 indicates that "whether a lane change has been performed" is "undetermined," the type of lane change is unknown while "whether a lane change has been performed" remains "undetermined." This prevents the subsequent lane being identified. Such a case also requires that the vehicle position be identified on the map by the vehicle position initialization unit 9, that is, that the vehicle position be reinitialized.

[0070] Therefore, not only immediately after the vehicle starts traveling, but whenever “whether a lane change has been completed” is “undetermined” or the vehicle position on the map calculated by the vehicle position on the map calculation unit 8 is significantly different from the satellite positioning result obtained from the satellite positioning device 12, the vehicle position on the map initialization unit 9 preferably notifies the vehicle position on the map calculation unit 8 of the satellite positioning result obtained from the satellite positioning device 12.

[0071] In addition, at a specific point where the vehicle position can be identified at a lane level, such as a point where the vehicle enters a highway or at an installation point of an optical beacon, information concerning the position thus identified may be transmitted to the vehicle position calculation unit 8 on the map. Overall operation of vehicle position estimation device 2

[0072] A description will be given of the overall operation of the vehicle position estimating device 2 with reference to Fig. 9.As a prerequisite, it is assumed that the map vehicle position initialization unit 9 has already notified the map vehicle position calculation unit 8 of the information concerning a position 1000.

[0073] First, the map vehicle position calculation unit 8 acquires the ahead distance from the vehicle ahead distance calculation unit 10. A vehicle position resulting from reflecting the thus acquired ahead distance becomes a position 1001. Accordingly, the map vehicle position calculation unit 8 acquires the lane connection relation at position 1001 from the map information acquisition unit 11, and notifies the vehicle lateral position calculation unit 7 of the thus acquired lane connection relation.

[0074] In the example according to Fig. 9, the lane connection relation includes information representing that another lane entering from the left side of the road causes the first lane before the intersection to become the second lane at position 1001, and causes the second lane before the intersection to become the third lane at position 1001. The vehicle lateral position calculation unit 7 switches the traveled lane from the first lane to the second lane while maintaining the position in the lane.

[0075] Next, the plotted line recognition unit 5 performs plotted line recognition, and the plotted line recognition result processing unit 6 determines whether a lane change has been performed is "No." The vehicle lateral position calculation unit 7 notifies the vehicle map position calculation unit 8 of the vehicle lateral position calculated based on the division line position calculation result notified by the plotted line recognition result processing unit 6. A vehicle lateral position on the map, which is a result of the recalculation performed by the vehicle map position calculation unit 8, becomes a position 1002.

[0076] Next, the map vehicle position calculation unit 8 acquires the ahead distance from the vehicle ahead distance calculation unit 10. A vehicle position resulting from reflecting the thus acquired ahead distance becomes a position 1003. Furthermore, the map vehicle position calculation unit 8 acquires the lane connection relation at the position 1003 from the map information acquisition unit 11, and notifies the vehicle lateral position calculation unit 7 of the thus acquired lane connection relation.

[0077] Next, the plotted line recognition unit 5 performs plotted line recognition, and the plotted line recognition result processing unit 6 determines whether a lane change has been completed is "Yes" based on the confirmation that the vehicle has crossed the left plotted line. The vehicle lateral position calculation unit 7 changes the vehicle lateral position to the first lane. The vehicle lateral position calculation unit 7 notifies the vehicle lateral position calculation unit 8 of the vehicle lateral position. A vehicle lateral position on the map, which is a result of the recalculation performed by the vehicle lateral position calculation unit 8, becomes a position 1004. Effects

[0078] In the first embodiment, the vehicle lateral position calculation unit 7 calculates the vehicle lateral position in a lane based on the lane change information and the dividing line detection position calculated by the plotted line detection result processing unit 6. The vehicle map position calculation unit 8, in turn, calculates the vehicle longitudinal position based on the ahead distance obtained from the vehicle ahead distance calculation unit 10. This makes it possible to estimate the vehicle lateral position and the vehicle longitudinal position in a lane without using satellite positioning while the vehicle is moving.

[0079] Furthermore, the vehicle position on the map calculation unit 8 is capable of calculating the vehicle position on the map by considering the lane connection relation information near the branch point. Since the position of the plotted line captured by the camera is used in the first embodiment, it is possible to prevent the accumulation of direction errors that occur in dead reckoning and prevent a position deviating from the road from being displayed. Consequently, it is possible to accurately calculate the vehicle position. Second embodiment structure

[0080] In the first embodiment, the use of the image captured by the camera and the preceding distance calculated by the preceding vehicle distance calculation unit 10 enables an estimation of the vehicle position, which allows for a reduction in the cumulative error. However, a lateral position error due to noise or a recognition error included in the plotted line recognition result from the plotted line recognition unit 5 is still a problem. The second embodiment of the present invention solves such a problem.

[0081] Fig. 10 is a block diagram showing an example of a structure of a vehicle position estimation system 13 according to a second embodiment of the present invention. As shown in Fig. As shown in Figure 10, a vehicle position estimation device 14 according to the second embodiment includes a plotted line recognition result processing unit 15. The remaining components and operations are identical or equivalent to those of the first embodiment. Therefore, detailed descriptions of the remaining components and operations are omitted herein.

[0082] The plotted line recognition result processing unit 15 includes a plotted line continuity correction unit 16, a lane change state determination unit 17, a dividing line position calculation unit 18, and a time series data accumulation unit 19.

[0083] The plotted line continuity correction unit 16 corrects, based on the fact that the plotted lines on a road surface are graphically continuous, the noise or the recognition error included in the plotted line recognition result from the plotted line recognition unit 5 using information accumulated in the time series data accumulation unit 19, and outputs a continuous plotted line recognition result.

[0084] That is, the plotted line continuity correction unit 16 performs correction based on relative positions of the left and right sides detected by the plotted line detection unit 5 to make the positions of the left and right plotted lines time-continuous.

[0085] The lane change state determination unit 17 determines a lane change state based on the correction result from the plotted line continuity correction unit 16 and the information accumulated in the time series data accumulation unit 19. That is, the lane change state determination unit 17 determines the lane change state of the vehicle, including the lane change direction, based on the relative positions of the left and right plotted lines corrected by the plotted line continuity correction unit 16, as well as the past positions of the left and right dividing lines. The lane change state includes "lane change started," "lane change in progress," "lane change complete," and "lane change interrupted."

[0086] The dividing line position calculation unit 18 determines the position of the plotted line based on the correction result from the plotted line continuity correction unit 16 and the lane change state determination result determined by the lane change state determination unit 17, and calculates the left dividing line position and the right dividing line position of the lane where the vehicle is traveling.

[0087] That is, the dividing line position calculation unit 18 calculates the positions of the left and right dividing lines based on the relative position of the left and right plotted lines corrected by the plotted line continuity correction unit 16 and the lane change state of the vehicle determined by the lane change state determination unit 17.

[0088] The time series data accumulation unit 19 stores and accumulates the positions of the left and right division lines calculated by the division line position calculation unit 18 as time series data. When the accumulated data exceeds a buffer size, the data in the buffer is deleted in chronological order. Operation of the continuity correction unit 16 for applied lines

[0089] Fig. 11 is a flowchart showing an example of the operation of the plotted line continuity correction unit 16.

[0090] In step S301, the plotted line continuity correction unit 16 waits until it is notified of the plotted line detection result from the plotted line detection unit 5. Upon notification of the plotted line detection result, the plotted line continuity correction unit 16 proceeds to step S302 and then corrects the plotted line detection result to maintain continuity based on the time series. At this time, a past plotted line detection result is based on time series data accumulated in the time series data accumulation unit 19.

[0091] Since the plotted line detection result from the plotted line detection unit 5 contains noise due to the driving environment or the condition of the road markings, the lateral position and slope of the plotted line may contain errors. Specifically, it is conceivable that, even if the road marking of the dividing line is continuous, the lateral position of the plotted line is detected as continuous, that a linear object different from the road marking of the dividing line is erroneously detected, or that a road marking of a dividing line of a lane different from the lane where the vehicle is traveling is detected. Therefore, considering the temporal changes in the detection positions of the left plotted line and the right plotted line, the following is performed: (C) detecting and correcting a position detection error of the plotted line under the influence of disturbances, ie detecting and correcting the position of the plotted line, and (D) Detection and correction of the movement of the detected plotted line to the adjacent plotted line, ie detection and correction of a change in the plotted line.

[0092] First, (C) Detection and correction of the plotted line position under the influence of disturbances is described. A condition for correcting the plotted line position is that the value of the temporal change in the lateral position of the plotted line exceeds a threshold. This corresponds to proceeding to step S304 if a determination is made in step S303 that the plotted line position can be corrected.

[0093] However, if both the temporal changes in the lateral position of the right plotted line and the lateral position of the left plotted line exceed the threshold, it is determined that the position of the plotted line cannot be corrected. This corresponds to proceeding to step S305 if a determination is made in step S303 that the position of the plotted line cannot be corrected.

[0094] In a process for correcting the position of the plotted line performed in step S306, it is determined that one of the left plotted line and the right plotted line whose temporal lateral movement value is smaller has higher reliability, and based on the position of the one plotted line having higher reliability and using the lane width included in the map information, the position of the other plotted line is calculated. This process corresponds to step S304.

[0095] That is, the plotted line continuity correction unit 16 determines that the right plotted line corresponds to the left division line at the current time when the difference between the position of the left division line, which is a division line adjacent to a left side of the vehicle, obtained at a predetermined time before a current time, accumulated in the time series data accumulation unit 19, and the position of the right plotted line, which is a plotted line adjacent to a right side of the vehicle, at the current time and detected by the plotted line detection unit 5, is within a predetermined distance, and calculates the position of the right division line, which is a division line adjacent to the right side of the vehicle,at the current time based on the left dividing line at the current time and the lane width included in the map information, and determines that the left plotted line corresponds to the right dividing line at the current time when the difference between the position of the right dividing line obtained at a predetermined time before a current time, accumulated in the time series data accumulation unit 19, and the position of the left plotted line, which is a plotted line adjacent to the left side of the vehicle, at the current time and detected by the plotted line detection unit 5, is within the predetermined distance, and calculates the position of the left dividing line based on the right dividing line at the current time and the lane width.

[0096] For example, it is assumed that the position of the left plotted line and the position of the right plotted line at the current time (t) by the plotted line recognition unit 5 are represented by the following equations (15), (16), respectively. y=L0,t+L1,t⋅x+fL,t(x) y=R0,t+R1,t⋅x+fR,t(x)

[0097] In equations (15), (16) the following applies: L 0,t denotes the lateral position of the left plotted line at time t, calculated by the plotted line recognition unit 5, and R 0,t denotes the lateral position of the right plotted line at time t, calculated by the plotted line detection unit 5. L 1,t denotes the slope of the left plotted line at time t, calculated by the plotted line recognition unit 5, and R 1,tdenotes the slope of the right plotted line at time t, calculated by the plotted line detection unit 5.

[0098] In addition, f L,t (x) and f R,t (x) a two- or more-dimensional polynomial of the left plotted line at time t calculated by the plotted line recognition unit 5, and a two- or more-dimensional polynomial of the right plotted line at time t calculated by the plotted line recognition unit 5, respectively, and they are used as error terms in the present embodiment.

[0099] In addition, it is assumed that the position of the left dividing line and the position of the right dividing line at the previous time point (t-1) accumulated in the time series data accumulation unit 19 are represented by the following equations (17), (18), respectively. y=L(D)0,t−1+L(D)1,t−1⋅x+f(D)L,t−1(x) y=R(D)0,t−1+R(D)1,t−1⋅x+f(D)R,t−1(x)

[0100] In equations (17), (18) the following applies: L (D) 0,t denotes the lateral position of the left dividing line at time t, determined by the dividing line position calculation unit 18 and accumulated in the time series data accumulation unit 19, R (D) 0,t denotes the lateral position of the right dividing line at time t, determined by the dividing line position calculation unit 18 and accumulated in the time series data accumulation unit 19, L (D) 1,t denotes the slope of the left dividing line at time t, determined by the dividing line position calculation unit 18 and accumulated in the time series data accumulation unit 19, and R (D) 1,tdenotes the slope of the right dividing line at time t, determined by the dividing line position calculation unit 18 and accumulated in the time series data accumulation unit 19.

[0101] In addition, f (D) L,t (x) and f (D) R,t (x) a two- or more-dimensional polynomial of the left plotted line at time t determined by the dividing line position calculation unit 18, and a two- or more-dimensional polynomial of the right plotted line at time t determined by the dividing line position calculation unit 18, respectively, and are used as error terms in the present embodiment.

[0102] With the variables defined as described above, it is assumed that a relation between the parameters satisfies, for example, the following inequalities (19), (20). |L0,t−L(D)0,t−1|≤θ0,diff |R0,t−R(D)0,t−1|>θ0,diff

[0103] In inequalities (19), (20) the following applies: θ 0,diff denotes a threshold parameter which represents an upper limit of the continuity of a zero-order coefficient of the plotted line, and which can be changed according to the output frequency of the detection result of the plotted line, so that, for example, when the detection result of the plotted line is output at 10 Hz, θ 0,diff is set equal to 50 cm, and when the detection result of the plotted line is output at 2 Hz, θ 0,diff is set equal to 100 cm In this configuration, it can be assumed that the left plotted line is continuously changed from inequality (19), but it can be assumed that the right plotted line is not continuously changed from inequality (20).

[0104] Therefore, the position of the right plotted line is changed based on the position of the left plotted line. Specifically, the coefficients of the position of the left plotted line and the position of the right plotted line after correction to satisfy continuity, as shown in Equations (21) and (22), are calculated from the following Equations (23) to (28). y=L(C)0,t+L(C)1,t⋅x+f(C)L,t(x) y=R(C)0,t+R(C)1,t⋅x+f(C)R,t(x) L(C)0,t=L0,t L(C)1,t=L1,t f(C)L,t(x)=fL,t(x) R(C)0,t=L(C)0,t+Widt R(C)1,t=L(C)1,t+Dirt f(C)R,t(x)=f(C)L,t(x)

[0105] In equations (21) to (28) the following applies: L (c) 0,t denotes the lateral position of the left plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16, R(C) 0,t denotes the lateral position of the right plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16, L (C) 1,t denotes the slope of the left plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16, and R (C) 1,t denotes the slope of the right plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16.

[0106] In addition, f (C) L,t (x) and f (C) R,t(x) a two- or more-dimensional polynomial of the left plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16, and a two- or more-dimensional polynomial of the right plotted line at time t after the continuity correction performed by the plotted line continuity correction unit 16, respectively, and they are used as error terms in the present embodiment.

[0107] In addition, Widt denotes the lane width, which may be the width of a lane near the current point acquired from the map information acquisition unit 11, or which may be calculated from the following equation (29) which takes into account a difference between the lane width captured by the camera and the actual lane width. Widt=(R(D)0,t−1−L(D)0,t−1)∗(WMap,t / WMap,t−1)

[0108] In equation (29) W Map,t the lane width of a lane where the vehicle is traveling, obtained from the map information obtaining unit 11 at the current time (t), and W Map,t-1 denotes the lane width of a lane where the vehicle is traveling, obtained from the map information obtaining unit 11 at the previous time (t-1).

[0109] Furthermore, dirt refers to the angular difference between the right and left dividing lines. At most points, except near a point where the number of lanes increases or decreases, the right and left dividing lines are parallel to each other, so dirt can be specified as 0 or calculated from the following equation (30), taking into account the angular difference between the left and right dividing lines. Dirt=DirMap,R,t−DirMap,L,t

[0110] In equation (30) Dir Map,R,t the direction of the right dividing line of the lane where the vehicle is traveling, at a traveling point at the current time t, obtained from the map information obtaining unit 11, and Dir Map,L,t denotes the direction of the left dividing line of the lane where the vehicle is traveling, at the traveling point at the current time t, acquired from the map information acquisition unit 11.

[0111] The above description of (C) detection and correction of the plotted line position under the influence of disturbances was made with reference to a case in which inequalities (19) and (20) are satisfied, i.e., the correction is made based on the position of the left plotted line. The same applies to the case in which the correction is made based on the position of the right plotted line, i.e., the following inequalities (31) and (32) are satisfied. |L0,t−L(D)0,t−1|>θ0,diff |R0,t−R(D)0,t−1|≤θ0,diff

[0112] Next, a description will be given of (D) the detection and correction of the movement of the detected plotted line to the adjacent plotted line, ie, detection and correction of a change in the plotted line. When the difference between the lateral position of the left plotted line at the previous time (t-1) and the lateral position of the right plotted line at the current time (t) is equal to or smaller than the threshold value, ie, when a determination in step S305 of Fig. 11 results in “Yes”, continue with S306.

[0113] In step S306, the plotted line continuity correction unit 16 determines that two plotted lines are the same, and sets the plotted line detected as the right plotted line at the current time (t) as the left dividing line. At this time, the position of the right dividing line is calculated based on the position of the left dividing line using the lane width included in the map information.

[0114] For example, it is assumed that the position of the left plotted line and the position of the right plotted line at the current time (t), detected by the plotted line detection unit 5, are represented by equations (15), (16), and that the position of the left division line and the position of the right division line at the previous time (t-1), accumulated in the time series data accumulation unit 19, are represented by equations (17), (18). Furthermore, it is assumed that a relationship between the parameters satisfies the following inequality (33). |R0,t−L(D)0,t−1|≤θ0,diff

[0115] In this case, the change value between the left plotted line at the previous time and the right plotted line at the current time is equal to or less than the threshold, and therefore both plotted lines are assumed to be identical to each other. Since it is considered that the plotted line captured by the camera has changed to the left plotted line, the information about the plotted line is replaced.

[0116] More specifically, the coefficients of the position of the left plotted line and the position of the right plotted line after the continuity correction, as shown in the following equations (34), (35), are calculated from the following equations (36) to (41). y=L(C)0,t+L(C)1,t*X+f(C)L,t(X) y=R(C)0,t+R(C)1,t*X+f(C)R,t(X) L(C)0,t=R0,t L(C)1,t=R1,t f(C)L,t(X)=fR,t(X) R(C)0,t=L(C)0,t+Widt R(C)1,t=L(C)1,t+Dirt f(C)R,t(X)=f(C)L,t(X)

[0117] In equations (34) to (41), Widt denotes the lane width, which may be the width of a lane near the current point acquired from the map information acquisition unit 11 or calculated from equation (29).

[0118] In addition, Dirt, which denotes the angle difference between the right division line and the left division line, can be set to 0 or calculated from equation (30).

[0119] Above, the description of (D) of the detection and correction of the movement of the detected plotted line to the adjacent plotted line, that is, detection and correction of a change in the plotted line, is made with reference to a case where the inequality (33) is satisfied, that is, the plotted line captured by the camera has changed to the left plotted line, and a case where the plotted line captured by the camera has changed to the right plotted line, that is, the correction is performed based on the position of the left plotted line. The same applies to the case where the correction is performed based on the position of the right plotted line, that is, the following inequality (42) is satisfied. |L0,t−R(D)0,t−1|≤θ0,diff

[0120] If both determinations in step S303 and step S305 result in "No," that is, if noises are included that cannot be corrected by the plotted line continuity correction unit 16, the plotted line continuity correction unit 16 determines in step S307 that the plotted line cannot be corrected. Specifically, if all of the following inequalities (43) to (46) are satisfied, the plotted line continuity correction unit 16 determines that the plotted line cannot be corrected. |L0,t−L(D)0,t−1|>θ0,diff |R0,t−R(D)0,t−1|>θ0,diff |L0,t−R(D)0,t−1|>θ0,diff |R0,t−L(D)0,t−1|>θ0,diff

[0121] In step S308, the plotted line continuity correction unit 16 notifies the lane change state determination unit 17 and the dividing line position calculation unit 18 of the parameters L (C)0,t , L (C) 1,t , R (C) 0,t , R (C) 1,t and the polynomials f (C) L,t (x), f (C) R,t (x), which have been corrected in this way. Operation of the lane change state determination unit 17

[0122] Fig. 12 is a flowchart showing an example of the operation of the lane change state determination unit 17. The lane change state determination unit 17 determines a lane change state of the vehicle based on the plotted line position information after the continuity correction performed by the plotted line continuity correction unit 16. The lane change state includes the following: "no lane change," "lane change started," "lane change in progress," "lane change complete," "lane change canceled," and "lane change undetermined."

[0123] In step S401, the lane change state determination unit 17 waits until it is notified of the plotted line continuity correction result. Upon notification of the plotted line continuity correction result, the lane change state determination unit 17 proceeds to step S402.

[0124] In step S402, the lane change state determination unit 17 determines whether the plotted line continuity correction unit 16 has made a correction to make the plotted line continuous. To identify the lane change state, the lateral positions and gradients of the continuously changing left plotted line and right plotted line are obtained from the plotted line continuity correction unit 16. If the plotted line continuity correction unit 16 has not made a correction to make the plotted line continuous, the lane change state determination unit 17 proceeds to step S403.

[0125] In turn, if the plotted line continuity correction unit 16 has performed a correction to make the plotted line continuous, the lane change state determination unit 17 proceeds to step S404. In step S403, the lane change state determination unit 17 determines that the lane change state is "lane change indeterminate."

[0126] In step S404, the lane change state determination unit 17 determines whether the previous lane change state is "no lane change," "lane change complete," "lane change started," or "lane change in progress." If "no lane change" or "lane change complete," the lane change state determination unit 17 proceeds to step S405. If "lane change started" or "lane change in progress," the lane change state determination unit 17 proceeds to step S409.

[0127] In step S405, if the previous lane change state is "no lane change," the lane change state determination unit 17 determines whether there is an indication of a lane change start. Specifically, the precedence distance for the vehicle to cross the dividing line is calculated by dividing the lateral position of the left dividing line by the slope of the left dividing line. This is referred to as a precedence distance for crossing, or a crossing precedence distance.

[0128] In addition, the time required for the vehicle to cross the dividing line is calculated by dividing the lateral position of the left dividing line by the value of the temporal change to the left dividing line and multiplying the result of the division by a time coefficient. This is referred to as the time required to cross, or the necessary crossing time.

[0129] For example, under the condition L (C) 0,t < 0 and L (C) 1,t > 0 the distance ahead D L to cross the left plotted line and the time T L , which is necessary to cross the left plotted line, is calculated from the following equations (47), (48). DL=(−L(C)0,t) / L(C)1,t TL=(−L(C)0,t) / L(C,DiffAVE)0,t*TDiff

[0130] In equation (48) T Diff the elapsed time per frame. L(C,DiffAVE) 0,t denotes an average rate of change per frame of the lateral position of the left plotted line around the current time (t) and is calculated from the following equation (49) using an averaging parameter N. L(C,DiffAVE)0,t=(L(C)0,t−L(D)0,t−N) / N

[0131] It is assumed that the lateral position of the left plotted line at time (tN) is a value L (D) 0,t-N after the determination by the correction process performed by the division line position calculation unit 18, is obtained from the time series data accumulation unit 19 and used. Since the lateral position of the left plotted line at the current time (t) is still to be determined, a value L (C) 0,t after the continuity correction performed by the applied line continuity correction unit 16.

[0132] In step S406, the lane change state determination unit 17 determines whether both the leading crossing distance and the crossing time are equal to or less than a threshold. If both the leading crossing distance and the crossing time are equal to or less than the threshold, the lane change state determination unit 17 proceeds to step S407.

[0133] In turn, when one or both of the preceding distance to cross and the time required to cross are larger than the threshold value, the lane change state determination unit 17 proceeds to step S408.

[0134] In step S407, the lane change state determination unit 17 determines that a case where the leading distance D L to cross and the time required to cross T Lbelow predetermined thresholds θ D or θ T fall, ie that the following inequalities (50), (51) are both satisfied, is an indication of a lane change, and it determines that the lane change state is "lane change started". DL<θD TL<θT

[0135] For example, the threshold θ D the distance ahead for crossing is 30 m, and the threshold θ T The time required to cross can be 3 seconds. After the lane change state is determined as "lane change started," the lane change state is automatically determined as "lane change in progress" until it is determined to be either "lane change complete" or "lane change aborted."

[0136] That is, the lane change state determining unit 17 calculates the crossing lead distance and the necessary crossing time based on time series data obtained during a predetermined time, including a current time accumulated in the time series data accumulating unit 19 and the positions of the left dividing line and the right dividing line corrected by the plotted line continuity correcting unit 16, and determines whether the vehicle is performing a lane change based on the crossing lead distance and the necessary crossing time, and determines, when the vehicle starts to change lanes, the direction of the lane change.

[0137] In step S408, the lane change state determination unit 17 determines that the lane change state is “no lane change”.

[0138] In step S409, the lane change state determination unit 17 determines whether the lane change is complete. If the lane change is complete, the lane change state determination unit 17 proceeds to step S410. If the lane change is yet to be completed, the lane change state determination unit 17 proceeds to step S411.

[0139] In step S410, the lane change state determination unit 17 determines that the lane change state is "lane change complete." For example, in the case where the lane change start direction is left, the lane change state is determined to be "lane change complete" when the lateral position of the left dividing line exceeds a threshold value θ. 0,complete which is a predetermined first threshold value greater than 0. The threshold value θ 0,complete is, for example, 0.5 m. As with θ 0,diffthe threshold value θ 0,complete according to the output frequency of the detection result of the plotted line.

[0140] In step S411, the lane change state determination unit 17 determines whether the lane change is canceled. If the lane change is canceled, the lane change state determination unit 17 proceeds to step S412. If the lane change is not canceled, the lane change state determination unit 17 proceeds to step S413.

[0141] In step S412, the lane change state determination unit 17 determines that the lane change state is "lane change canceled." For example, in the case where the lane change starting direction is left, the lane change state is determined as "lane change canceled" when the lateral position of the left dividing line falls below a threshold value -θ. 0,cancelwhich is a predetermined second threshold value less than 0.

[0142] The threshold θ 0,cancel is, for example, -0.8 m. The same applies to the case where the lane change starting direction is right. As with θ 0,diff the threshold value θ 0,cancel according to the output frequency of the detection result of the plotted line.

[0143] In step S413, the lane change state determination unit 17 determines that the lane change state is “lane change in progress”.

[0144] In step S414, the lane change state determining unit 17 notifies the dividing line position calculating unit 18 of the lane change state.

[0145] The above describes the case where the lane change starting direction is left. The same applies to the case where the lane change starting direction is right. Operation of the division line position calculation unit 18

[0146] The dividing line position calculation unit 18 calculates the position of the dividing line based on the position of the plotted line corrected to be continuous by the plotted line continuity correction unit 16 and the lane change state determined by the lane change state determination unit 17. Unless the lane change state determination unit 17 determines that the lane change state is "lane change complete," the position of the dividing line is the same as the position of the plotted line corrected by the plotted line continuity correction unit 16. The following is a description of a process for the case where the lane change state determination unit 17 determines that the lane change state is "lane change complete."

[0147] For example, when the lane change state determination unit 17 determines that the lane change to the right lane is complete, the dividing line position calculation unit 18 calculates both the lateral position of the left dividing line and the right dividing line of the right lane. The lateral positions of the left dividing line and the right dividing line of the right lane are calculated using the following method.

[0148] The left dividing line of the right lane is the right dividing line of the current lane. The right dividing line of the right lane is calculated in the same way as in equations (34) to (41) used to correct the change in the plotted line in the plotted line continuity correction unit 16, using the lane width included in the map information with reference to the left dividing line of the right lane.

[0149] That is, when the lane change state determining unit 17 determines that the lane change is complete, the dividing line position calculating unit 18 calculates the positions of the left dividing line and the right dividing line adjacent to both sides of the vehicle after the lane change, based on the left dividing line after the correction calculated by the plotted line continuity correcting unit 16 and the lane width included in the map information.

[0150] More specifically, the coefficients of the left dividing line position and the right dividing line position, with the lane change considered, as shown in the following equations (52), (53), are calculated from the following equations (54) to (59). y=L(D)0,t+L(D)1,t⋅x+f(D)L,t(x) y=R(D)0,t+R(D)1,t⋅x+f(D)R,t(x) L(D)0,t=R(C)0,t L(D)1,t=R(C)1,t f(D)L,t(x)=f(C)R,t(x) R(D)0,t=L(D)0,t+Widt R(D)1,t=L(D)1,t+Dirt f(D)R,t(x)=f(D)L,t(x)

[0151] In equations (52) to (59) the following applies: L (D) 0,t denotes the lateral position of the left dividing line at time t, determined by the dividing line position calculation unit 18, R (D) 0,t denotes the lateral position of the right dividing line at time t, determined by the dividing line position calculation unit 18, L (D) 1,t denotes the gradient of the left dividing line at time t, determined by the dividing line position calculation unit 18, and R (D) 1,tdenotes the slope of the right dividing line at time t, determined by the dividing line position calculation unit 18. Furthermore, Widt denotes the lane width, which may be the width of a lane near the current point obtained from the map information acquisition unit 11 or calculated from equation (29).

[0152] Next, the determination of whether a lane change has been completed and the direction of the lane change performed by the dividing line position calculation unit 18 will be described. When the lane change state determined by the lane change state determination unit 17 is any of "no lane change," "lane change started," "lane change in progress," and "lane change canceled," the dividing line position calculation unit 18 determines that "whether a lane change has been completed" is "No."

[0153] If the lane change status determined by the lane change status determination unit 17 is "lane change complete," the dividing line position calculation unit 18 determines that "whether a lane change has been performed" is "yes," and the direction of the lane change is a direction notified by the lane change status determination unit 17. If the lane change status determined by the lane change status determination unit 17 is "lane change impossible," the dividing line position calculation unit 18 determines that "whether a lane change has been performed" is "undetermined." Operation of the time series data accumulation unit 19

[0154] The time series data accumulation unit 19 stores and accumulates past positions of the dividing line calculated by the dividing line position calculation unit 18 as time series data. The time series data accumulated in the time series data accumulation unit 19 is read out and used when the plotted line continuity correction unit 16 and the lane change state determination unit 17 perform a comparison using the past positions of the dividing line.

[0155] The buffer of the time series data accumulation unit 19 has a fixed size, and when the buffer is full, the data in the buffer is deleted in chronological order. Therefore, the time series data accumulation unit 19 appropriately has a ring buffer type data structure. Calculation order of the local variables in the processing unit 15 for detection results of the plotted line

[0156] Fig. 13 is a diagram for describing the calculation order of the local variables in the plotted line recognition result processing unit 15.

[0157] In step S501, the plotted line continuity correction unit 16 is notified by the plotted line recognition unit 5 of the following: The lateral position L 0,t the left plotted line at time t, the lateral position R 0,t the right-hand line at time t, the slope L 1,t the left plotted line at time t, and the slope R 1,t the right-hand line at time t.

[0158] In step S502, the plotted line continuity correction unit 16 calculates the following, taking into account the lane width Widt and the direction difference Dirt: the lateral position L (c) 0,t the left plotted line at time t after the continuity correction, the lateral position R (C) 0,t the right plotted line at time t after the continuity correction, the slope L (C) 1,t the left plotted line at time t after the continuity correction, and the slope R (C) 1,t the right-hand line at time t after the continuity correction.

[0159] In step S503, the dividing line position calculation unit 18 calculates the following, taking into account the lane width Widt and the direction difference Dirt: the lateral position L (D) 0,t the left plotted line at time t, the lateral position R(D) 0,t the right-hand line at time t, the slope L (D) 1,t the left plotted line at time t, and the slope R (D) 1,t the right-hand line at time t, namely for L (c) 0,t , R (C) 0,t , L (C) 1,t , and R (C) 1,t which are calculated by means of the continuity correction unit 16 for plotted lines.

[0160] L (D) 0,t , R (D) 0,t , L (D) 1,t , and R (D) 1,t, calculated by the division line position calculation unit 18, are notified to the time series data accumulation unit 19 and the vehicle lateral position calculation unit 7. The operation after the vehicle lateral position calculation unit 7 acquires these pieces of information is the same as in the first embodiment. Effects

[0161] In the second embodiment, the plotted line continuity correction unit 16 performs correction to make the position of the plotted line detected by the plotted line detection unit 5 continuous. The lane change state determination unit 17 determines the lane change state, taking past positions of the dividing line into account. The dividing line position calculation unit 18 calculates the position of the dividing line using all of these pieces of information. This improves robustness against noise or detection error contained in the output from the road imager 3 or the plotted line detection unit 5. Third embodiment structure

[0162] In the first and second embodiments, the vehicle position estimation system updates the lateral position using the camera's imaging result, and updates the longitudinal position based on the leading distance calculated by the vehicle leading distance calculation unit 10. At this time, the initial position corresponds to a positioning position input from the satellite positioning device 12 to the vehicle position initialization unit 9 on the map as the basis for updating.

[0163] However, if the positioning result of the satellite positioning device 12 has an error in the lateral direction, the initial position is set to the wrong lane, causing a problem in that the vehicle position calculation unit 8 on the map may continue to determine the lane in error. The third embodiment of the present invention solves this problem.

[0164] In the first and second embodiments, the vehicle position estimation system updates the vehicle position based on the position initialized on the map by the vehicle position initialization unit 9 with the vehicle lateral position calculated by the vehicle lateral position calculation unit 7 using the camera imaging result as input. In the third embodiment, it is contemplated that this problem is solved by having the vehicle position estimation system perform reliability evaluation based on both the initialization position and the updated position, and only when the reliability meets a certain standard does it determine that the vehicle position has been determined with high accuracy and select the position updated based on the camera imaging result.In turn, if the reliability does not meet the standard, then the vehicle position estimation system selects the initialization position with low accuracy and makes the vehicle position with high accuracy uncertain.

[0165] Fig. 14 is a block diagram showing an example of a structure of a vehicle position estimation system 20 according to the third embodiment. As shown in Fig. As shown in Figure 14, a vehicle position estimation device 21 according to the third embodiment includes a plotted line recognition unit 22, a vehicle position on the map calculation unit 24, and a vehicle lateral position calculation unit 28. The remaining components and operations are identical or equivalent to those of the second embodiment. Therefore, detailed descriptions of the remaining components and operations are omitted here.

[0166] The applied line detection unit 22 has a line type detection unit 23.

[0167] When the plotted line recognition unit 22 detects a plurality of plotted lines on the road from an image representing an area in front of the vehicle acquired from the image information acquisition unit 4, the line type recognition unit 23 analyzes the plotted lines in more detail to estimate the line type of each of the plotted lines. The line type represents the type of line pattern (shape) and is classified into types such as a solid line, a long-dashed line, a short-dashed line, and a double line based on the line types existing in the actual road environment.

[0168] The map vehicle position calculation unit 24 includes a lane reliability estimation unit 25, an image-based vehicle position calculation unit 26, and a position integration unit 27.

[0169] The lane reliability estimation unit 25 makes a comprehensive determination as to whether the plotted line recognition unit 22 has successfully recognized the plotted line, whether the line type recognized by the line type recognition unit 23 is a stable result on a time series basis, and whether a vehicle position on the map calculated by the position integration unit 27 continuously and stably indicates a position near the center of the lane on a time series basis, so that it estimates the reliability of the traveled lane calculated by the vehicle lateral position calculation unit 28.

[0170] The image-based vehicle position calculation unit 26 operates in the same manner as the entire vehicle position calculation unit 8 on the map in the first embodiment. When the vehicle lateral position is notified by the vehicle lateral position calculation unit 28, the vehicle lateral position on the map is updated, and when the vehicle ahead distance is notified by the vehicle ahead distance calculation unit 10, the vehicle position on the map is updated in the forward-backward direction.

[0171] Since the vehicle position calculated in this way is a vehicle position calculated using the lateral position calculated based on the image information input from the image information acquisition unit 4, such a vehicle position is referred to as an image-based vehicle position.

[0172] If the reliability of the traveled lane estimated by the lane reliability estimation unit 25 is less than a threshold, the position integration unit 27 determines that the image-based vehicle position calculated by the image-based vehicle position calculation unit 26 has low reliability, and selects the initialization position output from the map vehicle position initialization unit 9 as the calculation result of the map vehicle position calculation unit 24.

[0173] In turn, when the reliability of the traveling lane estimated by the lane reliability estimation unit 25 is equal to or greater than the threshold value, the position integration unit 27 determines that the image-based vehicle position calculated by the image-based vehicle position calculation unit 26 has high reliability based on a result of evaluating the image-based vehicle position using a plurality of evaluation axes, and uses the image-based vehicle position as the calculation result of the map vehicle position calculation unit 24.

[0174] Once the reliability of the traveling lane estimated by the lane reliability estimation unit 25 becomes equal to or greater than the threshold, as long as the plotted line recognition unit 22 continues to recognize at least one of the left and right plotted lines and the plotted line continuity correction unit 16 continues to correct without errors, the reliability of the traveling lane is determined to be equal to or greater than the threshold, causing the vehicle position estimation system 20 to continue using the image-based vehicle position.

[0175] In addition to the role of the vehicle lateral position calculation unit 7 according to the first embodiment, the vehicle lateral position calculation unit 28 has the role of notifying the lane reliability estimation unit 25 of a plotted line detection state of the plotted line detection unit 22, a correction possibility state of the plotted line continuity correction unit 16, and a result of the vehicle lateral position consistency determination necessary for the lane reliability estimation unit 25 to perform the reliability estimation.

[0176] Here, the result of the consistency determination of the vehicle lateral position means the result of checking whether the vehicle lateral position is outside the lane (a travel-permissible range) due to the lane change determined by the lane change state determination unit 17. Operation of Lane Reliability Estimation Unit 25

[0177] The lane reliability estimation unit 25 makes a comprehensive determination regarding the following: whether the plotted line recognition unit 22 has successfully recognized the plotted line; whether the line type recognized by the line type recognition unit 23 is a stable result on a time-series basis; whether the vehicle position on the map calculated by the position integration unit 27 continuously and stably indicates a position near the center of the lane on a time-series basis; and whether the vehicle lateral position notified by the vehicle lateral position calculation unit 28 is within the travel-permissible range, so that it estimates the reliability of the traveled lane calculated by the vehicle lateral position calculation unit 28.As an example, the following is a description of a specific use case regarding the calculation of the reliability estimate.

[0178] On an expressway, in a scene where the vehicle travels on a ramp having one lane and then travels on a main road having a plurality of lanes, as long as the plotted line detection unit 22 continues to detect at least one of the left and right plotted lines and the plotted line continuity correction unit 16 continues to correct without errors, the traveling lane at the time point of traveling on the main road is uniquely determined, which causes the lane reliability estimation unit 25 to determine that the reliability of the traveling lane at the time point of traveling on the main road is high.

[0179] However, there is no plotted line near a toll booth or the like, so in a scene where the traveling lane cannot be identified by monitoring the plotted line, the lane reliability estimation unit 25 determines that the reliability of the traveling lane is low. During subsequent travel, the following applies: If there is determination material for identifying the traveling lane, the reliability increases according to the degree of the determination material.

[0180] In addition, for example, if the plotted line detection unit 22 has lost the view of the plotted lines on both sides while the vehicle is traveling, the lane reliability estimation unit 25 is expected to determine that the reliability of the traveling lane has decreased according to the length of time for which the plotted line detection unit 22 has lost the view of the plotted lines.

[0181] Examples of the reliability calculation method with the above features include the following calculation equations of a reliability point. (Current reliability point in case of two or more lanes) = min((previous reliability point) + (difference point), (initialization point)) (Current reliability point in the case of a single lane) = 100 (Difference point) = (Line type point Pt1) + (Position point Pt2 in the lane) + (Detection probability for plotted lines point Pt3) + (Continuity correction possibility point Pt4) (Initialization point)=(lane change to point Pt5 outside the road) (Initial value of the reliability point = 0 Pt1=−10 (when the type of the left and right dividing lines detected by the line type detection unit 23 does not match the type of the left and right dividing lines on the map at the position of the position integration unit 27) Pt1=10(in case of agreement) Pt2=10∗(1−(distance from the center of the lane[m])) Pt3=−30(if at least one of the left and right plotted lines was not detected) Pt3=0 (if both plotted lines are successfully detected) Pt4=−30 (if it is determined that the continuity correction cannot be performed) Pt4=0 (if the continuity correction has been successfully performed) Pt5=0 (when the lane change destination determined by the lane change state determination unit 17 is outside the road) Pt5=0×FFFF(maximum permissible value:If not outside the road)

[0182] Here, the reliability point is an index representing the strength of reliability in a point representation, and indicates that the higher the reliability point is, the higher the reliability of the traveled lane estimated by the position integration unit 27 becomes.

[0183] Generally, the lane reliability estimation unit 25 adds the difference point as a reference to the previous reliability point to calculate the current reliability point. However, if a large discrepancy is found in the position estimated up to this point, such as when the lane change target determined by the lane change state determination unit 17 is off-road, the lane reliability estimation unit 25 sets the reliability point accumulated up to this point to 0, thus resetting the cumulatively calculated vehicle position based on the camera in the past.

[0184] The position integration unit 27 is notified of the reliability point of the traveling lane calculated by the lane reliability estimation unit 25 according to equations (60) to (73). Operation of the calculation unit 26 for image-based vehicle positions

[0185] Since the image-based vehicle position calculation unit 26 operates in the same manner as the entire map vehicle position calculation unit 8 in the first embodiment, the flowchart of the operation is as shown in Fig. 6 shown. Operation of the Position Integration Unit 27

[0186] Fig. 15 is a flowchart showing an example of the operation of the position integration unit 27.

[0187] In step S601, the position integration unit 27 waits until a predetermined time (e.g., 100 ms) has elapsed after the previous position integration process. Therefore, the following process is performed at fixed intervals.

[0188] In step S602, the position integration unit 27 obtains the reliability of the traveling lane last estimated by the lane reliability estimation unit 25.

[0189] In step S603, the position integration unit 27 determines whether the reliability of the traveled lane thus obtained is equal to or greater than a threshold value. In the third embodiment, the reliability of the traveled lane is evaluated in the reliability point map, and a determination is made, for example, whether the reliability point is equal to or greater than 100, which is a value predetermined as a threshold value.

[0190] If the reliability point is equal to or greater than the threshold, the position integration unit 27 proceeds to step S604. In this case, it is determined that the lateral position calculated by the vehicle lateral position calculation unit 28 is reliable, so the vehicle position on the map is updated to the image-based vehicle position (a position calculated using the image-based lateral position).

[0191] If the reliability point is again lower than the threshold, the position integration unit 27 proceeds to step S605 and initializes the vehicle position on the map to the initialization position notified by the vehicle position initialization unit 9. This is because a determination has been made that the image-based vehicle position, which is cumulatively calculated based on the image, is unreliable. Thereafter, the position integration unit 27 proceeds to step S606.

[0192] In step S606, the position integration unit 27 notifies the image-based vehicle position calculation unit 26 of the thus initialized vehicle position on the map. This is because, after the notification, the longitudinal position and the lateral position are updated based on the initialized vehicle position on the map.

[0193] Finally, in step S607, the position integration unit 27 notifies the lane reliability estimation unit 25 of the image-based vehicle position or the vehicle position on the map updated to the initial position. The lane reliability estimation unit 25 evaluates the correctness of the vehicle position on the map with respect to the lane in the next lane reliability estimation. Operation of the vehicle lateral position calculation unit 28

[0194] In addition to the operation of the vehicle lateral position calculation unit 7 according to the first embodiment, as shown in Fig. 5, the vehicle lateral position calculation unit 28 determines whether the plotted line recognition unit 22 has successfully recognized the plotted line, whether the plotted line continuity correction unit 16 has successfully performed the continuity correction, and whether the vehicle position is outside the travel allowable range due to a lane change determined by the lane change state determination unit 17, and notifies the lane reliability estimation unit 25 of the results of the determinations. Overall operation of the vehicle position estimation system 20

[0195] A description will be given of the overall operation of the vehicle position estimating device 21 with reference to Fig.16.Assume, as a precondition, that the position integration unit 27 calculated the position 2001 at the previous time, and that the reliability point estimated by the lane reliability estimation unit 25 is 150. In this case, since the reliability point is equal to or greater than 100, which is the threshold value, the image-based vehicle position calculated by the image-based vehicle position calculation unit 26 is selected. Therefore, it is possible to calculate the lateral position by tracing the plotted line detected by the plotted line detection unit 22. However, it is assumed here that the traveling lane detected by the position integration unit 27 is incorrect.

[0196] At position 2001, the line type point Pt1 is set to -10 because the type of the map's division line and the type detected by the line type detection unit 23 are different. Pt2 is set to -8 because a position relatively close to the lane center and 20 cm away from the lane center is calculated for the lane position point Pt2. Both the plotted line detection possibility point Pt3 and the continuity correction point Pt4 are set to 0 because the plotted line detection and continuity correction were successfully performed.

[0197] The lane change outside the road point Pt5 is set to 0 because no lane change has been determined by the lane change state determination unit 17. From the above results, the difference point is calculated as -2, and as a result, the most recent reliability point is calculated as 148 points. Accordingly, the position integration unit 27 selects the output result of the image-based vehicle position calculation unit 26 immediately afterward. Subsequently, the reliability point calculation is performed sequentially in the same manner.

[0198] At position 2002, a lane change to a right lane is determined by the lane change state determination unit 17, and it is considered the off-road lane change, so the off-road lane change is set to 0 at the road point Pt5, and the most recent reliability point is calculated as 0 points. Therefore, the initialization position is selected by the vehicle position initialization unit 9 on the map. At this time, the image-based lateral position accumulated by the vehicle lateral position calculation unit 28 is reset.

[0199] When the position 2003 is reached, the calculated reliability point becomes equal to or greater than 100, which is the threshold value, so that it is determined that the vehicle position inputted from the vehicle position initialization unit 9 on the map becomes stable, and the consistency with the line type information recognized by the line type recognition unit 23 has been confirmed, so that a transition is made to a mode where the output result of the image-based vehicle position calculation unit 26 is selected.

[0200] When the position 2004 is reached, a lane change to a right lane is subsequently determined by the lane change state determination unit 17, so that the image-based vehicle position calculation unit 26 quickly calculates an own vehicle position after the lane change, which enables the position integration unit 27 to calculate the own vehicle position in response to a lane change. Effects

[0201] In the third embodiment, the reliability of the current traveling lane is determined based on the following: the time series of the line type of the plotted line recognized by the line type recognition unit 23; the time series of the positional relationship between the left and right dividing lines of the traveling lane included in the map acquired by the map information acquisition unit 11 and the vehicle position on the map output from the position integration unit 27; and the length of time that the plotted line recognition unit 22 has failed to recognize the plotted line, or that the plotted line continuity correction unit 16 has determined that the plotted line cannot be corrected.

[0202] When the reliability is equal to or greater than the threshold, the lateral position (the traveling lane and the lateral position in the lane) is determined based on the imaging result of the camera, and when the reliability is lower than the threshold, the output from the vehicle position initialization unit 9 on the map is selected and waited until the reliability becomes equal to or greater than the threshold, so that it becomes possible to prevent an error in detecting the traveling lane due to a position error of the vehicle position initialization unit 9 on the map and an error in determining the traveling lane based on the subsequent imaging result of the camera under the influence of the error in detection.

[0203] It should be noted that the present invention can be implemented with any combination of the embodiments within the scope of the present invention, and that each of the embodiments can be appropriately modified or features can be omitted.

[0204] Although the present invention has been described in detail, the above descriptions are merely illustrative in all aspects, and the present invention is not limited by the descriptions. It should be understood that countless non-illustrated modifications can be devised without departing from the scope of the present invention. List of reference symbols 1 Vehicle position estimation system 2 Vehicle position estimation device 3 Street imaging device 4 Image information reference unit 5 Detection unit for applied lines 6 Processing unit for detection results of the applied line 7 Calculation unit for vehicle lateral positions 8 Calculation unit for vehicle positions on the map 9 Initialization unit for vehicle positions on the map 10 Calculation unit for vehicle-ahead distances 11 Map information reference unit 12 Satellite positioning device 13 Vehicle position estimation system 14 Vehicle position estimation device 15 Processing unit for detection results of the applied line 16 Continuity correction unit for applied lines 17 Lane change state determination unit 18 Calculation unit for division line positions 19 Accumulation unit for time series data 20 Vehicle Position Estimation System 21 Vehicle position estimation device 22 Detection unit for applied lines 23 Line type recognition unit 24 Calculation unit for vehicle positions on the map 25 Lane reliability estimation unit 26 Calculation unit for image-based vehicle positions 27 Position integration unit 28 Calculation unit for vehicle lateral positions

Claims

[1] Vehicle position estimation device comprising: - an image information acquisition unit (4) configured to acquire image information obtained by capturing an image of a road ahead of a vehicle; - a plotted line recognition unit (5) configured to recognize relative positions of left and right plotted lines with respect to the vehicle based on the image information acquired by the image information acquisition unit (4), the left and right plotted lines being plotted on the road; - a plotted line detection result processing unit (6, 15) configured to calculate whether the vehicle has completed a lane change and a direction of the lane change when the vehicle has completed the lane change, wherein the relative positions of the left and right sides detected by the plotted line detection unit (5) are positions of left and right dividing lines defining a lane of the road; - a map information obtaining unit (11) configured to obtain map information including at least one lane connection relation which is a connection relation of the lane in a traveling direction of the road; - a vehicle lateral position calculation unit (7) configured to calculate a lane where the vehicle is traveling and a vehicle lateral position, which is a lateral position of the vehicle in the lane, based on whether the vehicle has completed the lane change, the direction of the lane change calculated by the plotted line recognition result processing unit (6, 15), and the lane connection relation obtained from the map information obtaining unit (11); - a vehicle advance distance calculation unit (10) configured to calculate a advance distance, which is a distance by which the vehicle has moved forward; and - a vehicle position on the map calculation unit (8) configured to calculate a position of the vehicle on the map on the road based on the map information obtained by the map information obtaining unit (11), the lane where the vehicle is traveling, the vehicle lateral position calculated by the vehicle lateral position calculation unit (7), and the vehicle ahead distance calculated by the vehicle ahead distance calculation unit (10), wherein the plotted line recognition result processing unit (15) comprises: a plotted line continuity correction unit (16) configured to perform correction so that the positions of the left and right plotted lines are temporarily made continuous, based on the relative positions of the left and right sides detected by the plotted line detection unit (5), a lane change state determination unit (17) configured to determine a lane change state of the vehicle, including the direction of the lane change, based on the relative positions of the left and right plotted lines corrected by the plotted line continuity correction unit (16) and the past position of the left and right dividing lines, a dividing line position calculation unit (18) configured to calculate the positions of the left and right dividing lines based on the relative positions of the left and right plotted lines corrected by the plotted line continuity correction unit (16) and the lane change state of the vehicle determined by the lane change state determination unit (17), and a time series data accumulation unit (19) configured to accumulate the position of the left and right dividing lines calculated by the dividing line position calculation unit (18) as time series data, wherein the map information obtained by the map information obtaining unit (11) includes a lane width which is the width of the lane, and wherein the plotted line continuity correction unit (16) determines that the right plotted line corresponds to the left dividing line at a current time when a difference between the position of the left dividing line, which is a dividing line adjacent to a left side of the vehicle, at a predetermined time before the current time, accumulated by the time series data accumulation unit (19), and the position of the right plotted line, which is a plotted line adjacent to a right side of the vehicle at the current time, recognized by the plotted line recognition unit (5), is within a predetermined distance, and calculates the position of the right dividing line, which is a dividing line adjacent to the right side of the vehicle at the current time, based on the left dividing line at the current time and the lane width included in the map information, and determines that the left plotted line corresponds to the right plotted line at a current time when a difference between the position of the right dividing line, relative to a predetermined time before the current time, accumulated by the time series data accumulation unit (19), and the position of the left plotted line, which is a plotted line adjacent to the left side of the vehicle at the current time, recognized by the plotted line recognition unit (5), is within a predetermined distance, and calculates the position of the left dividing line at the current time based on the right dividing line at the current time and the lane width. [2] Vehicle position estimation device according to claim 1, wherein the lane change state determination unit (17) calculates a crossing lead distance, which is a lead distance for the vehicle to cross the left dividing line or the right dividing line, and calculates a necessary crossing time, which is a time required for the vehicle to cross the left dividing line or the right dividing line, based on the time series data acquired during a predetermined time, including a current time accumulated by the time series data accumulation unit (19), and the position of the left dividing line and the right dividing line after the correction performed by the plotted line continuity correction unit (16), and determines whether the vehicle starts the lane change based on the crossing lead distance and the necessary crossing time, and, when the vehicle starts the lane change, a direction of the lane change, determines that if the position of the plotted line after the correction in the direction of the lane change is equal to or greater than a first threshold value after the vehicle starts the lane change, the vehicle has completed the lane change, and determines that if the position of the plotted line after correction in the direction of the lane change is equal to or less than a second threshold after the vehicle starts the lane change, the vehicle has canceled the lane change. [3] The vehicle position estimation device according to claim 2, wherein the dividing line position calculation unit (18), when the lane change state determination unit (17) determines that the lane change whose direction is left has been completed, calculates the positions of the left dividing line and the right dividing line adjacent to both sides of the vehicle after the lane change on the basis of the left dividing line after the correction calculated by the plotted line continuity correction unit (16) and the lane width included in the map information, and when the lane change state determination unit (17) determines that the lane change whose direction is right has been completed, calculates the positions of the left dividing line and the right dividing line adjacent to the vehicle,after the lane change on the basis of the right dividing line according to the correction calculated by the continuity correction unit (16) for plotted lines and the lane width contained in the map information. [4] Vehicle position estimation device according to claim 1, further comprising a map vehicle position initialization unit (9) configured to output an initialization position for initializing a map position of the vehicle held by the map vehicle position calculation unit (24) on the basis of a positioning result obtained by a positioning device different from the map information acquisition unit (11), - wherein the map vehicle position calculation unit (24) comprises: a lane reliability estimation unit (25) configured to estimate a lane reliability based on the position of the vehicle on the map calculated by the vehicle lateral position calculation unit (28), an image-based vehicle position calculation unit (26) configured to calculate an image-based vehicle position based on the vehicle-ahead distance calculated by the vehicle-ahead distance calculation unit (10) and the vehicle lateral position calculated by the vehicle lateral position calculation unit (28), and a position integration unit (27) configured to determine which of the initialization position output by the vehicle position initialization unit (9) on the map and the image-based vehicle position calculated by the image-based vehicle position calculation unit (26) is selected as a vehicle position, according to the reliability of the traveled lane estimated by the lane reliability estimation unit (25). [5] Vehicle position estimation device according to claim 4, wherein the lane reliability estimation unit (25) calculates the reliability of the traveled lane based on at least one of a time series of a positional relationship between the left and right dividing lines included in the map information acquired by the map information acquisition unit (11) and the position of the vehicle on the map, a duration of a state where the plotted line recognition unit (22) determines that the plotted lines are not successfully recognizable or a duration of a state where the plotted line continuity correction unit (16) determines that the plotted lines are not correctable, and whether the position of the vehicle on the map is off-road due to the lane change determined by the lane change state determination unit (17). [6] Vehicle position estimation device according to claim 5, wherein the applied line recognition unit (22) comprises a line type recognition unit (23) configured to recognize a line type of each of the left and right applied lines on the basis of the image information acquired from the image information acquisition unit (4), and the lane reliability estimation unit (25) calculates the reliability of the traveled lane at a current time based on a time series of the line type of each of the plotted lines recognized by the line type recognition unit (23).

Images