JP2018159752A - Method and device for learning map information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To learn positional information of a landmark used for correcting an odometry error caused by the difference in the altitude of travelling paths at appropriate intervals.SOLUTION: The method for learning map information includes the steps of: calculating a distance difference between a first travelling distance and a second travelling distance, the first travelling distance being the length of the travelling path of a vehicle and the second travelling distance being the length of the travelling path on a two-dimensional map (S5); and recording the positional information of a landmark detected around the vehicle into the map data (S7) every time when the vehicle travels a distance specified on the basis of the distance difference (S6).SELECTED DRAWING: Figure 4

Description

  The present invention relates to a map information learning method and a map information learning device.

  As a technique for estimating the position of a traveling vehicle, a vehicle position estimation device described in Patent Document 1 is known. This vehicle position estimation device obtains the amount of movement of the vehicle using vehicle speed information, and estimates the current position of the vehicle therefrom. On the other hand, at the point where detailed map information exists, using the road alignment information such as white lines extracted from the detailed map and landmark information, the point where the scene obtained with the in-vehicle camera and the detailed map information match is identified. Then, a highly accurate vehicle position is obtained. When generating detailed map information, the vehicle position estimation device refers to the elevation information to obtain the lane boundary information extracted from the captured image of the in-vehicle camera as the lane boundary information of the three-dimensional coordinate point sequence. Convert to

JP 2005-265494 A

The self position of the vehicle can be estimated by so-called odometry, in which the moving distance and direction of the vehicle are determined according to the rotation angle and rotation angular velocity of the left and right wheels.
For example, if there is an altitude difference in the travel path of the vehicle, a difference is generated between the travel distance along the road surface and the travel distance on the two-dimensional map. The difference in altitude of the travel path is one of the factors of odometry error that occurs between the self-position estimated by odometry and the self-position on the two-dimensional map.
This odometry error can be corrected by comparing the position information of landmarks around the road previously stored in the map data with the detected positions of landmarks detected around the vehicle.

However, if the landmark position information is stored too densely, the information amount of the landmark position information exceeds the amount of recording media that can be mounted on the vehicle, so that the landmark position information cannot be learned over a wide range.
An object of the present invention is to learn landmark position information at appropriate intervals.

  In the map information learning method according to the aspect of the present invention, the difference between the first travel distance that is the length of the travel locus of the vehicle and the second travel distance that is the length of the travel locus on the two-dimensional map. The distance difference which calculates the distance difference which is and detects the surrounding landmark of a vehicle and records the positional information on a landmark to map data is defined according to a distance difference.

  According to one embodiment of the present invention, landmark position information can be learned at appropriate distance intervals.

It is a figure which shows the example of schematic structure of the driving assistance device which concerns on 1st Embodiment. It is a block diagram which shows an example of a function structure of the controller shown in FIG. It is explanatory drawing of an example of the map information learning method. It is a flowchart of an example of the map information learning method which concerns on 1st Embodiment. It is a flowchart of an example of the modification of a map information learning method. It is explanatory drawing of the detection range of a landmark. It is a block diagram which shows an example of a function structure of the controller which concerns on 2nd Embodiment. It is a flowchart of an example of the map information learning method which concerns on 2nd Embodiment.

  Embodiments of the present invention will be described below with reference to the drawings. The following embodiments of the present invention exemplify apparatuses and methods for embodying the technical idea of the present invention, and the technical idea of the present invention is the structure, arrangement, etc. of components. Is not specified as follows. The technical idea of the present invention can be variously modified within the technical scope defined by the claims described in the claims.

(First embodiment)
(Constitution)
Please refer to FIG. The driving support device 1 includes a self-position estimation device 2 according to an embodiment of the present invention, a vehicle travel controller 50, and a vehicle control actuator group 51.
The self-position estimation device 2 estimates a position on a two-dimensional map of a vehicle (hereinafter simply referred to as “vehicle”) on which the driving support device 1 is mounted, and determines the position of a landmark used for vehicle position estimation. learn.

A landmark is a structure around a road, and may be a three-dimensional object such as a road sign, a traffic light, a pedestrian bridge, a guide board, and a building. A road surface such as a pedestrian crossing, a pedestrian crossing notice, and a traffic classification according to the direction of travel. It may be a display.
The self-position estimation apparatus 2 is an example of a map information learning apparatus according to an aspect of the present invention.

  The self-position estimation device 2 includes an ambient environment sensor group 10, a vehicle sensor group 20, a communication device 30, a map storage unit 31, a GPS (Global Positioning System) receiver 33, a second map 34, and an altitude database. 35 and a controller 40. In addition, in FIG.1, FIG.2, FIG.7 and the following description, an altitude database is described with "elevation DB."

The controller 40 includes an ambient environment sensor group 10, a vehicle sensor group 20, a communication device 30, a map storage unit 31, a GPS receiver 33, a second map 34, and an altitude database 35, and data, signals, or Information can be sent and received.
The map storage unit 31 can transmit / receive data, signals, or information to / from the communication device 30 by wired communication or wireless communication.

The ambient environment sensor group 10 is a sensor group that detects an ambient environment of the vehicle, for example, an object around the vehicle. For example, the ambient environment sensor group 10 may include a photographing device 11 and a distance measuring device 12. The photographing device 11 and the distance measuring device 12 detect the surrounding environment of the vehicle such as an object existing around the vehicle, a relative position between the vehicle and the object, and a distance between the vehicle and the object.
The imaging device 11 may be a stereo camera, for example. The photographing apparatus 11 may be a monocular camera, and may photograph the same object from a plurality of viewpoints with a monocular camera and calculate the distance to the object.

The distance measuring device 12 may be, for example, a laser range finder (LRF) or a radar.
The imaging device 11 and the distance measuring device 12 output the ambient environment information, which is detected ambient environment information, to the controller 40. The photographing device 11 and the distance measuring device 12 also output ambient environment information to the vehicle travel controller 50 that performs vehicle driving support control or automatic driving control.

  The vehicle sensor group 20 includes a sensor that detects a driving operation performed by the driver, a sensor that detects a traveling state of the vehicle, and other sensors 28. Sensors that detect driving operations performed by the driver include an accelerator sensor 21, a steering sensor 22, and a brake sensor 23. Sensors that detect the traveling state of the vehicle include a vehicle speed sensor 24, an acceleration sensor 25, a gyro sensor 26, and a wheel speed sensor 27.

The accelerator sensor 21 detects the accelerator opening of the vehicle and outputs information on the accelerator opening to the controller 40 and the vehicle travel controller 50. For example, the accelerator sensor 21 detects the depression amount of the accelerator pedal of the vehicle as the accelerator opening.
The steering sensor 22 detects the steered wheel turning angle, and outputs the detected turning angle information to the controller 40 and the vehicle travel controller 50.

The brake sensor 23 detects a brake operation amount by the driver, and outputs information on the detected operation amount to the controller 40 and the vehicle travel controller 50. For example, the brake sensor 23 detects the amount of depression of the brake pedal of the vehicle as a brake operation amount.
The vehicle speed sensor 24 calculates the speed of the vehicle, and outputs the calculated vehicle speed information to the controller 40 and the vehicle travel controller 50.

The acceleration sensor 25 detects the acceleration in the longitudinal direction of the vehicle and the acceleration in the vehicle width direction, and outputs information on these accelerations to the controller 40 and the vehicle travel controller 50.
The gyro sensor 26 detects the traveling direction of the vehicle and outputs traveling direction information to the controller 40 and the vehicle traveling controller 50.
The wheel speed sensor 27 detects the wheel speed of the vehicle and outputs wheel speed information to the controller 40 and the vehicle travel controller 50.
The information on the accelerator opening, the turning angle, the brake operation amount, the vehicle speed, the acceleration, the traveling direction, and the wheel speed output from the vehicle sensor group 20 may be referred to as “vehicle state information”.

  The communication device 30 communicates with a communication device outside the vehicle. For example, the communication device 30 performs inter-vehicle communication with other vehicles, performs road-to-vehicle communication with devices installed on the road shoulder, or wirelessly communicates with an information server installed outside the vehicle. By performing the above, various information can be acquired from the external device.

The map storage unit 31 includes a first map 32 that is two-dimensional map data used for driving support of the vehicle by the driving support device 1. The map storage unit 31 may be a car navigation system or a map database, for example.
The first map 32 is a lane level information including road markings such as white lines, stop lines, pedestrian crossings, road markings on roads, and point sequence information indicating the positions of steps such as curbs that indicate road edges and median strips. Map data.

The first map 32 may include point sequence information indicating the positions of landmarks around the road. The point sequence information indicating the position of the landmark has a three-dimensional coordinate value corresponding to the three-dimensional shape of the landmark, and a value indicating the position on the two-dimensional map as a two-dimensional coordinate value indicating the horizontal position. Have.
The map storage unit 31 may acquire the first map 32 from the outside via the communication device 30 or may receive the latest map information periodically and update the first map 32. The map storage unit 31 may accumulate the travel path on which the vehicle actually traveled in the first map 32 as map information.

The GPS receiver 33 detects the position of the vehicle by receiving a satellite positioning signal transmitted from a GPS satellite. The GPS receiver 33 outputs the detected vehicle position information to the controller 40.
The second map 34 is two-dimensional map data used for complementing the first map 32. For example, the second map 34 is learned by the self-position estimation device 2 in an area where the point sequence information is not stored in the first map 32 or an area where the stored point sequence information is small and not sufficient for vehicle position detection. It may be used to store map information.

For example, the second map 34 may store point sequence information indicating the positions of landmarks learned by the self-position estimation device 2. This point sequence information has a three-dimensional coordinate value corresponding to the three-dimensional shape of the landmark, and has a value indicating a position on a two-dimensional map as a two-dimensional coordinate value indicating a horizontal position.
The elevation DB 35 stores elevation data for each point in a predetermined geographical range.

  The controller 40 is an electronic control unit that executes processing for estimating the position of the vehicle on the two-dimensional map and processing for learning the position of the landmark used for vehicle position estimation. The controller 40 includes peripheral components such as a processor 41 and a storage device 42. The processor 41 may be, for example, a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit).

  The controller 40 may be realized by a functional logic circuit set in a general-purpose semiconductor integrated circuit. For example, the controller 40 may include a programmable logic device (PLD) such as a field-programmable gate array (FPGA).

The storage device 42 may include any one of a semiconductor storage device, a magnetic storage device, and an optical storage device. The storage device 42 may include a register, a cache memory, and a memory such as a ROM (Read Only Memory) and a RAM (Random Access Memory) used as a main storage device.
The storage device 42 stores a computer program that is executed on the processor 41 and causes the controller 40 to execute a vehicle position estimation process and a landmark learning process.

  The vehicle travel controller 50 is an electronic control unit that performs vehicle travel control in accordance with the driving operation of the occupant detected by the vehicle sensor group 20. Furthermore, the vehicle travel controller 50 is based on the current position of the vehicle estimated by the self-position estimation device 2, the ambient environment detected by the ambient environment sensor group 10, and the first map 32 stored in the map storage unit 31. Provide driving assistance such as automatic driving, emergency stop, and obstacle avoidance.

The vehicle travel controller 50 includes a processor and peripheral components such as a storage device. The processor may be a CPU or an MPU, for example.
The vehicle travel controller 50 may be realized by a functional logic circuit set in a general-purpose semiconductor integrated circuit. For example, the vehicle travel controller 50 may include a programmable logic device such as a field programmable gate array.
The storage device may include any one of a semiconductor storage device, a magnetic storage device, and an optical storage device. The storage device may include a register, a cache memory, and a memory such as a ROM and a RAM used as a main storage device.

The vehicle travel controller 50 controls the travel of the vehicle by driving the vehicle control actuator group 51. The vehicle control actuator may include, for example, a steering actuator 52, an accelerator opening actuator 53, and a brake control actuator 54.
The steering actuator 52 controls the steering direction and amount of steering of the vehicle. The accelerator opening actuator 53 controls the accelerator opening of the vehicle. The brake control actuator 54 controls the braking operation of the vehicle brake device.

The vehicle travel controller 50 drives the vehicle control actuator group 51 based on the current position of the vehicle estimated by the self-position estimation device 2 when driving the vehicle. Control the amount of vehicle behavior to generate.
For example, the vehicle travel controller 50 determines the timing for changing the accelerator opening, the timing for changing the steering amount, and the timing for operating the brake device based on the current position of the vehicle. For example, the vehicle travel controller 50 changes the accelerator opening amount, the steering amount, and the braking amount by the brake device based on the current position of the vehicle.

An example of the functional configuration of the controller 40 is shown in FIG. The controller 40 includes an odometry calculation unit 60, a trajectory calculation unit 61, a white line detection unit 62, a landmark detection unit 63, a matching unit 64, a landmark correction unit 65, a trajectory correction unit 66, and a landmark recording. The unit 67 is provided.
The functions of the odometry calculation unit 60, trajectory calculation unit 61, white line detection unit 62, landmark detection unit 63, matching unit 64, landmark correction unit 65, trajectory correction unit 66, and landmark recording unit 67 are stored in the storage device 42. This is realized by executing the stored computer program by the processor 41.

The odometry calculation unit 60 calculates the amount of movement and the direction of movement of the vehicle per unit time using the vehicle state information obtained from the vehicle sensor group 20 and the positioning information of the GPS receiver 33.
The trajectory calculation unit 61 calculates the travel locus of the vehicle on the road using the movement amount and the movement direction calculated by the odometry calculation unit 60. Note that, when calculating the travel locus of the vehicle, a travel locus model obtained by learning the travel locus on the road and a vehicle model of the vehicle may be used.

The white line detection unit 62 determines the road markings such as white lines, stop lines, pedestrian crossings, road markings on the road, curbs that indicate road edges, median strips, and the like from changes in the surrounding environment information obtained from the surrounding environment sensor group 10. The position of the step (hereinafter referred to as “white line etc.”) is detected.
The landmark detection unit 63 detects the position of the landmark around the road from the change in the surrounding environment information obtained from the surrounding environment sensor group 10. Examples of the landmark include a power pole, a sign, a post, a building, a thing that exists around the host vehicle and can be detected as a straight line, and a thing that can be detected as a 90-degree edge. The landmark detection unit 63 may detect a landmark within a predetermined range around the vehicle. For example, the landmark detection unit 63 may detect landmarks within a predetermined distance from the vehicle. The landmark detection unit 63 outputs point sequence information indicating the position of the detected landmark to the landmark correction unit 65 and the trajectory correction unit 66.

  The matching unit 64 estimates the position of the vehicle on the two-dimensional map by collating the position of the white line or the like detected by the white line detection unit 62 with the position of the white line or the like stored in the first map 32. When the first map 32 includes landmark point sequence information, the matching unit 64 includes the landmark point sequence information detected by the landmark detection unit 63 and the landmark point sequence stored in the first map 32. You may estimate the position on the two-dimensional map of a vehicle by collating with information. The matching unit 64 outputs the estimated position information on the two-dimensional map of the vehicle to the landmark correction unit 65.

  The landmark correction unit 65 calculates the position of the landmark detected by the landmark detection unit 63 on the two-dimensional map based on the position of the vehicle on the two-dimensional map estimated by the matching unit 64. For example, the landmark correction unit 65 is configured to detect the land detected by the landmark detection unit 63 based on the position of the vehicle on the two-dimensional map and the relative positional relationship between the vehicle and the landmark detected by the landmark detection unit 63. The position of the mark on the two-dimensional map may be calculated.

The landmark correction unit 65 can calculate the positions of landmarks that are not stored in the first map 32 on the two-dimensional map. The landmark correction unit 65 outputs point sequence information indicating the calculated position of the landmark on the two-dimensional map to the landmark recording unit 67.
The trajectory correction unit 66 corrects the travel locus of the vehicle calculated by the trajectory calculation unit 61 based on the position of the vehicle on the two-dimensional map estimated by the matching unit 64, and estimates the position of the vehicle on the two-dimensional map. To do. When correcting, the vehicle trajectory model described above may be used.
Further, the trajectory correction unit 66 re-estimates the position on the two-dimensional map such as the white line detected by the white line detection unit 62 based on the estimated vehicle position, corrects the traveling locus again, and estimates the position of the vehicle. The accuracy may be improved.

In an area where point sequence information is not stored in the first map 32 or an area where the stored point sequence information is not sufficient for vehicle position detection, the trajectory correction unit 66 detects the landmark detected by the landmark detection unit 63. And the position of the landmark stored in the second map 34 are collated to estimate the position of the vehicle on the two-dimensional map.
The trajectory correction unit 66 corrects the travel locus of the vehicle calculated by the trajectory calculation unit 61 based on the estimated position of the vehicle on the two-dimensional map, and estimates the position of the vehicle on the two-dimensional map.

The trajectory correction unit 66 re-estimates the position of the landmark detected by the landmark detection unit 63 on the two-dimensional map based on the estimated vehicle position, corrects the traveling locus again, and corrects the vehicle position estimation accuracy. May be improved.
The trajectory correction unit 66 outputs self-position information indicating the estimated position of the vehicle on the two-dimensional map to the vehicle travel controller 50.

  The landmark recording unit 67 records point sequence information indicating the position of the landmark output from the trajectory correction unit 66 on the two-dimensional map on the second map 34. For example, the landmark recording unit 67 may store landmark point sequence information in an area where point sequence information is not stored in the first map 32, or in an area where the stored point sequence information is small and not sufficient for vehicle position detection. Record in the second map 34. For example, the landmark recording unit 67 records point sequence information on the second map 34 for landmarks that are not stored in the first map 32.

The landmark recording unit 67 is the first travel distance L _odm calculated by the odometry calculation unit 60, that is, the length of the travel track traveled along the road surface, and the length of the travel track on the two-dimensional map. A first distance difference D1 = | L _map −L _odm | which is a difference from the second travel distance L _map is calculated.
The landmark recording unit 67 detects landmarks around the vehicle and determines a distance interval between points at which landmark point sequence information is recorded on the second map 34 according to the first distance difference D1.
For example, the landmark recording unit 67 records the point sequence information of the landmark on the second map 34 every time the vehicle travels at an interval (distance) determined based on the first distance difference D1.

The landmark recording unit 67 includes a second travel distance calculation unit 70, a condition determination unit 71, and a recording unit 72.
The second traveling distance calculating unit 70, a first travel distance L _{odm-calculated} by odometry calculation unit 60, and the positioning information of the GPS receiver 33, based on the altitude data stored in the elevation DB 35, two-dimensional travel locus A second travel distance L _map that is the length on the _map is calculated.
The odometry calculation unit 60 corresponds to a first travel distance calculation unit that calculates the first travel distance L _odm .

Please refer to FIG. Vehicle 80 mounting the driving support device 1 is traveling on the point A at time t _0, during the period from the time t ₀ to time t _i, it is assumed that moving from point A to point B.
The second travel distance calculation unit 70 acquires the coordinates of the points A and B based on the positioning information of the GPS receiver 33 at time t ₀ and time t _i , respectively. The second travel distance calculation unit 70 acquires the elevations of the points A and B from the elevation DB 35 based on the coordinates of the points A and B. The second travel distance calculation unit 70 calculates an elevation difference H between the A point and the B point.

The second travel distance calculation unit 70 calculates the angle θ between the road surface between the point A and the point B and the horizontal plane and the second travel distance L _{map according} to the following expressions (1) and (2). .
θ = Sin ⁻¹ (H / L _odm ) (1)
L _map = L _odm × cos θ (2)

Please refer to FIG. The condition determination unit 71 includes a first travel distance L _odm that is the length of the travel locus of the vehicle calculated by the odometry calculation unit 60, and a second travel distance L _map that is the length of the travel locus on a two-dimensional map. The first distance difference D1 = | L _map −L _odm |
The condition determination unit 71 determines whether or not the vehicle has traveled a distance determined based on the first distance difference D1.

For example, the condition determination unit 71 determines whether or not the vehicle has traveled a distance determined based on the first distance difference D1 and one of the first travel distance L _odm and the second travel distance L _map. Good.
For example, the condition determination unit 71 determines whether or not the vehicle has traveled a distance at which one of the first travel distance L _odm and the second travel distance L _map is maximum while the first distance difference D1 is equal to or less than a predetermined threshold Th1. It may be determined.
Further, for example, the condition determination unit 71 determines the distance at which one of the first travel distance L _odm and the second travel distance L _map reaches a predetermined value while the first distance difference D1 is equal to or less than the predetermined threshold Th1. It may be determined whether or not the vehicle has traveled.

The recording unit 72 records the landmark point sequence information detected by the landmark detection unit 63 on the second map 34 in accordance with the determination result of the condition determination unit 71.
For example, the recording unit 72 may record the landmark point sequence information on the second map 34 each time the condition determination unit 71 determines that the vehicle has traveled a distance determined based on the first distance difference D1. That is, the recording unit 72 records the landmark point sequence information on the second map 34 at intervals determined based on the first distance difference D1.

For example, in the recording unit 72, the condition determination unit 71 determines that the vehicle has traveled a distance determined based on the first distance difference D1 and one of the first travel distance L _odm and the second travel distance L _map. Each time, landmark point sequence information may be recorded on the second map 34. That is, the recording unit 72 _displays the landmark point sequence information in the second map at intervals determined based on the first distance difference D1 and one of the first travel distance L _odm and the second travel distance L _map. 34.

For example, the recording unit 72 determines that the vehicle has traveled a distance at which one of the first travel distance L _odm and the second travel distance L _map is maximum while the first distance difference D1 is equal to or less than a predetermined threshold Th1. Each time 71 determines, landmark point sequence information may be recorded on the second map 34. That is, the recording unit 72 records the landmark point sequence information at intervals at which one of the first travel distance L _odm and the second travel distance L _map is maximum while the first distance difference D1 is equal to or less than the predetermined threshold Th1. May be recorded on the second map 34.
Further, for example, the recording unit 72 indicates that the condition determination unit 71 determines that one of the first travel distance L _odm and the second travel distance L _map has reached a predetermined value while the first distance difference D1 is not more than the predetermined threshold Th1. Each time a determination is made, landmark point sequence information may be recorded on the second map 34.
If landmark point sequence information is recorded at such an interval, an odometry error that occurs between the first travel distance L _odm along the road surface and the second travel distance L _map on the two-dimensional map due to an altitude difference is Th1. The odometry error can be corrected by using the landmark so that it is within the range. In the present embodiment, as an example, an odometry error generated between the first travel distance L _odm due to the altitude difference and the second travel distance L _map on the two-dimensional map is shown, but the present invention is not necessarily limited thereto. In addition, odometry errors also occur due to tire slippage, road surface unevenness, and the like, so that this embodiment can be implemented in response to such cases.

For example, the interval at which landmark point sequence information is recorded may be determined by the following process.
When the landmark point sequence information is recorded on the second map 34 at the time t ₀ , the odometry calculation unit 60 sets the position of the vehicle at the time t _{0 as} the starting point of the first travel distance L _odm . The odometry calculating unit 60 calculates the first travel distance L _odm from the starting point to the first time t _(i−1) . The odometry calculation unit 60 calculates the first travel distance L _odm from the starting point to the second time t _i following the first time t _(i−1) .

The second travel distance calculation unit 70 calculates a second travel distance L _map that is the length on the two-dimensional map of the travel locus from the starting point to the first time t _(i-1) . Further, the second traveling distance calculating unit 70 calculates a second running distance _{L map} from counting point to the second time _{t i.}
Condition determining unit 71 calculates the first distance difference at the first time _{t (i-1) D1 (} i-1), a first distance difference D1 _i at the second time _{t i,} respectively.
The recording unit 72 acquires the landmark point sequence information detected by the landmark detection unit 63 from the landmark correction unit 65 at the first time t _(i−1) and the second time t _i .

The condition determination unit 71 determines that the first distance difference D1 _{(i-1) at} the first time t _(i-1) is less than or equal to the threshold Th1, and the first distance difference D1 _{i at} the second time t _i is greater than the threshold Th1. Judge whether it is large or not. When the first distance difference D1 _(i-1) is equal to or smaller than the threshold Th1 and the first distance difference D1 _i is larger than the threshold Th1, the first distance difference D1 is the threshold Th1 at the first time t _(i-1) . One of the first travel distance L _odm and the second travel distance L _map becomes the maximum as follows.

When the first distance difference D1 _(i-1) is equal to or smaller than the threshold Th1 and the first distance difference D1 _i is larger than the threshold Th1, the recording unit 72 detects the landmark detection unit 63 at the first time t _(i-1) . The point sequence information of the landmarks detected in step 2 is recorded on the second map 34.
After the landmark is recorded on the second map 34, the odometry calculation unit 60 sets the position of the vehicle at the first time t _{(i-1) as} the starting point of the first travel distance _Lodm .
By repeating the above-described processing, the landmark becomes the second interval at which one of the first travel distance _Lodm and the second travel distance L _map is maximized while the first distance difference D1 is equal to or less than the predetermined threshold Th1. Recorded on the map 34.

(Operation)
Next, an example of the map information learning method according to the first embodiment will be described. Please refer to FIG. Steps S1 to S8 shown in FIG. 4 are respectively executed at times t _i (i = 1, 2, 3,...) Visited at a predetermined cycle (for example, every 100 milliseconds).

In step S1, the odometry calculation unit 60 calculates the length of the travel locus of the vehicle from the starting point of the first travel distance L _odm to the time t _i as the first travel distance L _odm .
Second traveling distance calculating unit 70 in step S2 calculates the altitude difference H between the point of counting point and time t _i of the first travel distance L _odm-.
Second traveling distance calculating unit 70 in Step S3 calculates the second traveling distance _{L map} until time _{t i} from counting point of the first travel distance _{L odm-.}

Landmark detection unit 63 in Step S4 detects the position of the landmarks near the road at time t _i. The recording unit 72 acquires the landmark point sequence information detected by the landmark detection unit 63 at time ti from the landmark correction unit 65.
In step S5, the condition determination unit 71 calculates a first distance difference D1 _i at time t _i .

Condition determining unit 71 in step S6, the time _{t (i-1)} a first distance difference in _{D1 (i-1)} is not less threshold Th1 or less, and the first distance difference D1 _i at time _{t i} is larger than the threshold Th1 Determine whether or not.
When the first distance difference D1 _(i-1) is equal to or smaller than the threshold Th1 and the first distance difference D1 _i is larger than the threshold Th1 (step S6: Y), the process proceeds to step S7. If the first distance difference D1 _i is less than the threshold Th1 (step S6: N), the process proceeds to step S8.

In step S <b> 7, the recording unit 72 records the landmark point sequence information detected by the landmark detection unit 63 at time t _(i−1) on the second map 34. The odometry calculation unit 60 sets the position of the vehicle at the time t _{(i−1) as} a starting point of the first travel distance L _odm . Thereafter, the process proceeds to step S8.
In step S <b> 8, the landmark recording unit 67 determines whether or not the learning of landmarks on the second map 34 should be terminated.

For example, the landmark recording unit 67 may determine that the learning of the landmark should end when the ignition switch of the vehicle is turned off. Further, for example, when a vehicle enters an area where point sequence information is stored in the first map 32, it may be determined that the learning of landmarks should be terminated.
If it is determined that the learning of the landmark should be terminated (step S8: Y), the process ends. If it is not determined that the landmark learning should be terminated (step S8: N), the process returns to step S1.

(Effect of 1st Embodiment)
The landmark recording unit 67 is a difference between the first travel distance L _odm that is the length of the travel locus of the vehicle and the second travel distance L _map that is the length of the travel locus on the two-dimensional map. A certain first distance difference D1 is calculated. The landmark recording unit 67 detects landmarks around the vehicle and determines a distance interval between points at which landmark point sequence information is recorded on the second map 34 according to the first distance difference D1.
Thereby, landmark position information can be learned at appropriate distance intervals. In addition, since the landmark position information can be recorded at appropriate distance intervals, the position of the host vehicle can be accurately identified.

The recording unit 72 records the position information of landmarks detected around the vehicle on the second map 34 every time the vehicle travels a distance determined based on the first distance difference D1.
The first distance difference D1 reflects an odometry error that occurs due to an elevation difference in the travel path. For this reason, by recording landmarks at intervals determined based on the first distance difference D1, the interval at which landmark position information is recorded can be determined in accordance with the odometry error caused by the elevation difference in the travel path. become able to.
Therefore, it is possible to learn the landmark position information used for correcting the odometry error caused by the elevation difference of the traveling road at an appropriate distance interval.

For example, if a conventional vehicle is driven at a speed of 36 km / h, 33 frames per second are shot with two stereo cameras, 30 landmarks per frame are recorded, and the data size per point is 10 KB. It was necessary to record a data amount of 1 GB per hit.
According to the present invention, for example, when the landmark recording interval is determined so that the first distance difference D1 is kept within 10 cm on a traveling road having a slope of 5 degrees, it is sufficient to record the landmarks at 40 locations per 1 km on the two-dimensional map. . For this reason, the amount of data to be recorded is 13.2 MB, which can be reduced to 1/75.
The landmark information is not necessarily limited to the point sequence, and may be any information that can identify the landmark. The landmark information is not necessarily recorded on the map, but may be recorded so that the position information can be specified. Further, the landmark information need not be recorded during traveling, and the recording timing is not limited as long as the landmark information can be recorded at distances between points.

Recording unit 72, a first distance difference D1, at one and the distance interval determined based on the first traveling distance L _odm-and second traveling distance L _map, the position of the landmark detected by the surroundings of the vehicle Information may be recorded on the second map 34.
For example, when the first distance difference D1 is equal to or less than a predetermined threshold value Th1, the recording unit 72 determines that the vehicle travels when one of the first travel distance L _odm and the second travel distance L _map reaches a predetermined value. The position information of landmarks detected around can be recorded on the second map 34.
By determining the interval for recording landmarks in this way, the interval can be set so that the interval for recording landmarks is as long as possible while suppressing odometry errors. Thereby, the amount of data required for landmark recording can be reduced.

The second travel distance calculation unit 70 may calculate the second travel distance L _map based on the elevation difference H of the travel path of the vehicle. The condition determination unit 71 may calculate a first distance difference D1 between the first travel distance L _odm and the second travel distance L _map . That is, the second travel distance calculation unit 70 and the condition determination unit 71 may calculate the first distance difference D1 based on the elevation difference H of the travel path of the vehicle.
Based on the altitude difference H, the first distance difference D1 can be calculated relatively easily and accurately.

The second mileage calculation unit 70 may measure the elevation difference H using the satellite positioning signal and the elevation DB 35. The difference in elevation is very small within the error range of the satellite positioning signal. For this reason, the altitude difference H can be accurately acquired by measuring the altitude difference H using the satellite positioning signal and the altitude DB 35.
The map data that records the position information of the landmark is a second map 34 that complements the first map 32 used for driving support of the vehicle. However, the recording unit 72 may record landmark position information on the first map 32.
By supplementing the first map 32 with the second map 34 in which landmarks have been learned with individual vehicles, the area where point sequence information is not stored in the first map 32, or the amount of stored point sequence information is small. Even if there is an area that is not sufficient for position detection, it is possible to correct the odometry error within a range necessary for an individual vehicle.

(Modification)
(1) When the landmark recording interval is determined so that the odometry error caused by the altitude difference is less than or equal to the threshold Th1, a threshold Th2 that is sufficiently smaller than the threshold Th1 is set, and the distance that the first distance difference D1 exceeds the threshold Th2 is set. Each time the vehicle travels, landmark position information detected around the vehicle may be recorded on the second map 34. That is, landmarks may be recorded on the second map 34 at intervals where the first distance difference D1 exceeds the threshold Th2.

Please refer to FIG. Steps S10 to S17 shown in FIG. 5 are respectively executed at times t _i (i = 1, 2, 3,...) Visited at a predetermined cycle (for example, every 100 milliseconds).
The process of steps S10 to S14 is the same as the process of steps S1 to S5 in FIG.
In step S15, the condition determining unit 71 determines whether or not the first distance difference D1 _{i at} time t _i is greater than the threshold value Th2.
When the first distance difference D1 _i is larger than the threshold Th2 (step S15: Y), the process proceeds to step S16. When the first distance difference D1 _i is less than the threshold Th2 (step S15: N), the process proceeds to step S17.

In step S <b> 16, the recording unit 72 records the landmark point sequence information detected by the landmark detection unit 63 at time t _i on the second map 34. The odometry calculating unit 60 sets the position of the vehicle at the time t _{i as} a starting point of the first travel distance L _odm . Thereafter, the process proceeds to step S17.
The process of step S17 is the same as the process of step S8 of FIG.
Also in this modification, similarly to the effect of the first embodiment, the interval for recording landmarks can be determined according to the odometry error caused by the elevation difference of the traveling road.
For this reason, the position information of the landmark used for correcting the odometry error caused by the altitude difference of the traveling road can be learned at an appropriate distance interval.

  (2) The second travel distance calculation unit 70 may measure the elevation difference H using the gyro sensor 26. By using the gyro sensor 26, the interval for recording landmarks can be determined without using the elevation DB 35. The same applies to the second embodiment described later.

(3) A landmark may not be detected due to a cause such as occlusion. As a result, when the vehicle has traveled a distance determined based on the first distance difference D1, the landmark detection unit 63 may not be able to detect a landmark within a predetermined range around the vehicle.
For this reason, when the landmark within the predetermined range around the vehicle cannot be detected when the vehicle travels the distance determined based on the first distance difference D1, the recording unit 72 is detected in the area around the predetermined range. Landmark position information may be recorded on the second map 34. The same applies to the second embodiment described later.

Please refer to FIG. When the landmark in the predetermined range 81 around the vehicle 80 cannot be detected, the landmark detection unit 63 detects the landmark detected in the region 82 having the predetermined width W around the predetermined range 81.
The recording unit 72 records the position information of the landmark detected in the area 82 on the second map 34.
Accordingly, even when the landmark detection unit 63 cannot detect a landmark within the predetermined range 81 when the vehicle 80 travels a distance determined based on the first distance difference D1, the land used for correcting the odometry error. Marks can be recorded. For this reason, accumulation of odometry errors can be prevented.

  (4) The threshold values Th1 and Th2 may be fixed values. It may be a dynamically changing value. The larger the thresholds Th1 and Th2, the wider the interval for recording landmarks, and the easier it is to accumulate odometry errors. The smaller the thresholds Th1 and Th2, the narrower the interval for recording landmarks, and the more difficult it is to accumulate odometry errors. That is, the estimation accuracy of the position of the vehicle is increased.

  For example, the condition determination unit 71 may change the threshold values Th1 and Th2 according to the driving environment. For example, in an environment where the degree of driving difficulty is high, the threshold values Th1 and Th2 may be reduced to increase the estimation accuracy of the position of the vehicle, and for example, deviation from the lane or approach to the end of the lane may be prevented. In an environment where the driving difficulty is low, the thresholds Th1 and Th2 may be increased to reduce the amount of data recorded on the second map 34. The same applies to the second embodiment described later.

For example, the condition determination unit 71 may change the threshold values Th1 and Th2 according to the season. For example, in the winter season when the driving difficulty such as road surface freezing tends to be high, the threshold values Th1 and Th2 may be made smaller than in the summer season.
For example, the condition determination unit 71 may change the threshold values Th1 and Th2 according to the location. For example, it is required to estimate the position of a vehicle with high accuracy in order to ensure safety in a place near a specific store or facility or in a place where congestion occurs, such as an intersection.

For this reason, the condition determination unit 71 determines whether or not the vehicle is traveling in a place where congestion is expected by referring to the first map 32. If the vehicle is traveling in a place where congestion is expected, the threshold value is determined. Th1 and Th2 may be reduced.
For example, the condition determination unit 71 may change the thresholds Th1 and Th2 according to the type and shape of the road on which the vehicle travels. For example, the thresholds Th1 and Th2 may be made smaller during traveling on a curve than when traveling on a straight road. The thresholds Th1 and Th2 may be made smaller during traveling on the highway than when traveling on a general road. The thresholds Th1 and Th2 may be made smaller during traveling on the intersection than when traveling on the road.

(5) When the second travel distance L _map, which is the length on the two-dimensional map, is known, the landmark recording unit 67 travels along the road surface based on the second travel distance L _map and the altitude difference H. _May be calculated, and the first distance difference D1 = | L _map −L _odm | may be calculated.
The landmark recording unit 67 may include a first travel distance calculation unit that calculates the first travel distance L _odm based on the following equations (3) and (4).
θ = Tan ⁻¹ (H / L _map ) (3)
L _odm = L _map / cos θ (4)
The same applies to the second embodiment described later.

(Second Embodiment)
Then, the self-position estimation apparatus 2 of 2nd Embodiment is demonstrated. In the second embodiment, a coefficient G to be multiplied by the first travel distance L _odm is calculated to correct the odometry error generated in the first travel distance L _odm calculated based on the rotation speed of the wheel, and the second map 34 is calculated. Record. Hereinafter, the coefficient G is referred to as “odometry gain G”.
The odometry calculation unit 60 corrects the odometry error by multiplying the odometry gain G by the first travel distance L _odm .

For example, the odometry gain G may be a coefficient for multiplying the first travel distance L _odm to calculate the second travel distance L _map .
The odometry calculation unit 60 calculates the second travel distance L _map by multiplying the first travel distance L _odm by the odometry gain G as shown in the following equation (5), and corrects the odometry error.
L _map = G × L _odm (5)

  The landmark recording unit 67 calculates and records the odometry gain G on the second map 34 every time the condition determination unit 71 determines that the vehicle has traveled a distance determined based on the first distance difference D1.

Please refer to FIG. The functional configuration of the controller 40 according to the second embodiment is similar to the functional configuration shown in FIG. 2, and the same reference numerals are given to the same components.
The controller 40 includes a third travel distance calculation unit 90. The landmark recording unit 67 includes an odometry gain calculation unit 91.
The functions of the third mileage calculation unit 90 and the odometry gain calculation unit 91 are realized by the processor 41 executing a computer program stored in the storage device 42.

The third mileage calculation unit 90 acquires a detection result of the surrounding environment of the vehicle from the surrounding environment sensor group 10. The third travel distance calculation unit 90 calculates the length of the travel locus of the vehicle along the road surface based on the detection result of the surrounding environment. The length of the travel locus calculated by the third travel distance calculation unit 90 is denoted as the third travel distance L _hi .
For example, the third travel distance calculation unit 90 may calculate the third travel distance L _hi based on a white line or the like extracted from the detection result of the ambient environment sensor group 10 or a landmark. The third travel distance L _hi may be calculated by visual odometry based on the detection result.

The third travel distance calculation unit 90 outputs the third travel distance L _hi to the landmark recording unit 67. The third travel distance L _hi is the length of the travel locus of the vehicle along the road surface, like the first travel distance L _odm . However, since the third travel distance L _hi is calculated from the detection result of the ambient environment sensor group 10, the accuracy is generally higher than the first travel distance L _odm calculated based on the rotation speed of the wheel.

The odometry gain calculation unit 91 calculates an odometry gain G.
The odometry gain calculation unit 91 may calculate the odometry gain G based on, for example, the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi . For example, the odometry gain calculation unit 91 may calculate an odometry gain such as G = L _map / L _odm or G = L _hi / L _odm .
The odometry gain calculation unit 91 may further calculate the odometry gain G in consideration of parameters such as road slope, speed, and road shape.

The odometry gain calculation unit 91 calculates the odometry gain G and records it in the second map 34 every time the condition determination unit 71 determines that the vehicle has traveled a distance determined based on the first distance difference D1.
In the second embodiment, the odometry gain calculation unit 91 includes a second distance difference D2 = | L _map −L _hi | that is a difference between the third travel distance L _hi and the second travel distance L _map , Every time the vehicle travels a distance determined based on the distance difference D1, an odometry gain G is calculated and recorded on the second map 34.

For example, the first travel distance L _odm , the second travel distance L _map , or the third travel distance L _hi is maximized while the first distance difference D1 is equal to or less than the predetermined threshold Th1, or the second distance difference D2 Odometry gain G is calculated each time the vehicle travels a distance at which one of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi is the maximum while the vehicle travels below the predetermined threshold Th3 And recorded on the second map 34.
That is, any of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi becomes the maximum while the first distance difference D1 is not more than the predetermined threshold Th1, or the second distance difference D2 The second map 34 calculates the odometry gain G at intervals at which one of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi is the maximum while maintaining the predetermined threshold Th3 or less. To record.
Further, for example, the first travel distance L _odm , the second travel distance L _map , or the third travel distance L _hi reaches a predetermined value while the first distance difference D1 is equal to or less than the predetermined threshold Th1, or the second distance difference D2 is a predetermined threshold value Th3 following while the first travel distance _{L odm-,} when the second travel distance _{L map,} either of the third traveling distance _{L hi} reaches a predetermined value, calculates the odometry gain G Record on the second map 34.

Similarly, the recording unit 72 determines whether any of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi is the maximum while the first distance difference D1 is equal to or less than the predetermined threshold Th1. Alternatively, every time the vehicle travels a distance at which one of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi is the maximum while the second distance difference D2 is equal to or less than the predetermined threshold Th3. The landmark point sequence information is recorded on the second map 34.
That is, any of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi becomes the maximum while the first distance difference D1 is not more than the predetermined threshold Th1, or the second distance difference D2 In the second map 34, the landmark point sequence information is at intervals at which one of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi is the maximum while the value is equal to or less than the predetermined threshold Th3. To record.
Further, for example, a first distance difference D1 is left with a predetermined threshold value Th1 following first travel distance _{L odm-,} second running distance _{L map,} either of the third traveling distance _{L hi} reaches a predetermined value, or the The landmark point sequence information when one of the first travel distance L _odm , the second travel distance L _map , and the third travel distance L _hi reaches a predetermined value while the two-distance difference D2 is not more than the predetermined threshold Th3. Is recorded on the second map 34.

As described above, the odometry gain G is calculated and recorded together with the landmark at an interval at which the second distance difference D2 between the third travel distance L _hi and the second travel distance L _map is equal to or less than the threshold Th3. Thus, the landmark correction can be performed before the error between the third travel distance L _hi calculated based on the detection result of the surrounding environment and the second travel distance L _map on the two-dimensional map becomes excessive. For this reason, it is possible to prevent an increase in the correction error of the odometry gain G due to an excessive correction amount by the odometry gain G.

Further, the odometry gain G is calculated and recorded together with the landmark at an interval at which the first distance difference D1 between the first travel distance L _odm and the second travel distance L _map is equal to or less than the threshold Th1. Accordingly, the correction by the landmark can be performed before the error of the first travel distance L _odm that is the odometry calculated based on the rotation of the wheel becomes excessive. For this reason, it is possible to prevent an increase in the correction error of the odometry gain G due to an excessive correction amount by the odometry gain G.

For example, the interval for recording the odometry gain G and the landmark may be determined by the following process.
When the odometry gain G and the landmark are recorded on the second map 34 at time t ₀ , the odometry calculation unit 60 sets the position of the vehicle at time t _{0 as} the starting point of the first travel distance L _odm . In addition, the third travel distance calculation unit 90 sets the position of the vehicle at time t _{0 as} a starting point for the third travel distance L _hi .

The odometry calculating unit 60 calculates the first travel distance L _odm from the starting point to the first time t _(i−1) . The odometry calculation unit 60 calculates the first travel distance L _odm from the starting point to the second time t _i following the first time t _(i−1) .
The third travel distance calculation unit 90 calculates the third travel distance L _hi from the starting point to the first time t _(i−1) . The third travel distance calculation unit 90 calculates a third travel distance L _hi from the starting point to the second time t _i .

The second travel distance calculation unit 70 calculates the second travel distance L _map from the starting point to the first time t _(i−1) and from the starting point to the second time t _i , respectively.
Condition determining unit 71 calculates the first distance difference at the first time _{t (i-1) D1 (} i-1), a first distance difference D1 _i at the second time _{t i,} respectively.
Condition determining unit 71 calculates a second distance difference at the first time _{t (i-1) D2 (} i-1), the second distance difference D2 _i in the second time _{t i,} respectively.
The recording unit 72 acquires the landmark point sequence information detected by the landmark detection unit 63 from the landmark correction unit 65 at the first time t _(i−1) and the second time t _i .

The odometry gain calculating unit 91 at the first time t _(i-1) , the first travel distance L _odm and the second travel distance L _{map at} the first time t _(i-1) , or the first travel distance L _odm and Based on the third travel distance L _hi , an odometry gain G that corrects an odometry error occurring at the first time t _(i−1) is calculated. Similarly odometry gain calculation unit 91 calculates the odometry gain G for correcting the odometry error caused in the second time t _i.

The condition determination unit 71 determines that the first distance difference D1 _{(i-1) at} the first time t _(i-1) is less than or equal to the threshold Th1, and the first distance difference D1 _{i at} the second time t _i is greater than the threshold Th1. Whether the second distance difference D2 _{(i-1) at} the first time t _(i-1) is less than or equal to the threshold Th3 and the second distance difference D2 _{i at} the second time t _i is greater than the threshold Th3 Judge whether or not.
The odometry gain calculation unit 91 has the first distance difference D1 _{(i-1) at} the first time t _(i-1) equal to or smaller than the threshold value Th1, and the first distance difference D1 _{i at} the second time t _i is equal to the threshold value Th1. When the condition determination unit 71 determines that the value is larger, the odometry gain G calculated at the first time t _(i−1) is recorded in the second map 34. If the second distance difference D2 _{(i-1) at} the first time t _(i-1) is less than or equal to the threshold Th3, and the second distance difference D2 _{i at} the second time t _i is greater than the threshold Th3, the condition determining unit 71 Is determined, the odometry gain G calculated at the first time t _(i−1) is recorded in the second map 34.

In the recording unit 72, the first distance difference D1 _{(i-1) at} the first time t _(i-1) is equal to or less than the threshold Th1, and the first distance difference D1 _{i at} the second time t _i is greater than the threshold Th1. When the condition determination unit 71 determines that the landmark point sequence information detected by the landmark detection unit 63 at the first time t _(i−1) is recorded in the second map 34. If the second distance difference D2 _{(i-1) at} the first time t _(i-1) is less than or equal to the threshold Th3, and the second distance difference D2 _{i at} the second time t _i is greater than the threshold Th3, the condition determining unit 71 Is determined, the landmark point sequence information detected by the landmark detection unit 63 at the first time t _(i−1) is recorded in the second map 34.
After the landmark is recorded on the second map 34, the odometry calculation unit 60 sets the position of the vehicle at the first time t _{(i-1) as} the starting point of the first travel distance _Lodm . The third travel distance calculation unit 90 sets the position of the vehicle at the first time t _{(i−1) as} a starting point for the third travel distance L _hi .

Subsequently, an example of the map information learning method according to the second embodiment will be described. Please refer to FIG. Steps S20 to S30 shown in FIG. 8 are respectively executed at times t _i (i = 1, 2, 3,...) Visited at a predetermined cycle (for example, every 100 milliseconds).
The processes in steps S20 to S22 are the same as the processes in steps S1 to S3 in FIG.
In step S23, the third travel distance calculation unit 90 calculates the length of the travel locus of the vehicle from the starting point of the third travel distance L _hi to the time t _i as the third travel distance L _hi .

The process in step S24 is the same as the process in step S4 in FIG.
Odometry gain calculating unit 91 in step S25 calculates the odometry gain G at time _{t i.}
The process of step S26 is the same as the process of step S5 of FIG.
In step S27, the condition determination unit 71 calculates the second distance difference D2 _i at time t _i .
Condition determining unit 71 in step S28, the time _{t (i-1)} a first distance difference in _{D1 (i-1)} is not less threshold Th1 or less, and the first distance difference D1 _i at time _{t i} is larger than the threshold Th1 Or the second distance difference D2 _{(i-1) at} time t _(i-1) is less than or equal to the threshold Th3, and it is determined whether the second distance difference D2 _{i at} time t _i is greater than the threshold Th3. .

Processing is performed when the first distance difference D1 _{(i-1) at} time t _(i-1) is equal to or smaller than the threshold Th1 and the first distance difference D1 _{i at} time t _i is larger than the threshold Th1 (step S28: Y). Advances to step S29. Even when the second distance difference D2 _{(i-1) at} time t _(i-1) is equal to or smaller than the threshold Th3 and the second distance difference D2 _{i at} time t _i is larger than the threshold Th3 (step S28: Y). The process proceeds to step S29. In other cases (step S28: N), the process proceeds to step S30.

In step S <b> 29, the recording unit 72 records the landmark point sequence information detected by the landmark detection unit 63 at time t _(i−1) on the second map 34. The odometry gain calculation unit 91 records the odometry gain G calculated at time t _(i−1) on the second map 34.
The odometry calculation unit 60 sets the position of the vehicle at the first time t _{(i−1) as} a starting point of the first travel distance L _odm . The third travel distance calculation unit 90 sets the position of the vehicle at the first time t _{(i−1) as} a starting point for the third travel distance L _hi .
The process in step S30 is the same as the process in step S8 in FIG.

(Effect of 2nd Embodiment)
The odometry calculation unit 60 calculates the first travel distance L _odm based on the rotation speed of the wheel. Odometry gain calculation unit 91, an odometry gain G is multiplied to the first travel distance L _odm-to correct an error generated in the first traveling distance L _odm-, the distance determined on the basis of the first distance difference D1 vehicle travels Each time it is calculated, the position information of the landmarks detected around the vehicle and the odometry gain G are recorded on the second map 34.

For example odometry gain G can be a factor to be multiplied to the first travel distance _{L odm-for} calculating a second travel distance _{L map} from the first travel distance _{L odm-.}
This makes it possible to perform position correction using the odometry gain G until position correction is performed using landmarks recorded at intervals determined based on the first distance difference D1, thereby improving vehicle position estimation accuracy. Can do.

The condition determination unit 71 calculates a first distance difference D1 that is a difference between the first travel distance L _odm and the second travel distance L _map . The condition determination unit 71 calculates a second distance difference that is a difference between the third travel distance L _hi that is the length of the travel locus of the vehicle calculated based on the detection result of the surrounding environment of the vehicle and the second travel distance L _map. D2 is calculated.
Each time the vehicle travels a distance determined based on the first distance difference D1 and the second distance difference D2, the odometry gain G is calculated, and the landmark position information and the odometry gain G detected around the vehicle are calculated. Record in the second map 34.

Thus, the landmark correction can be performed before the error between the third travel distance L _hi calculated based on the detection result of the surrounding environment and the second travel distance L _map on the two-dimensional map becomes excessive. For this reason, it is possible to prevent an increase in the correction error of the odometry gain G due to an excessive correction amount by the odometry gain G.
Further, the correction by the landmark is possible before the error of the first travel distance L _odm which is the odometry calculated based on the rotation of the wheel becomes excessive. For this reason, it is possible to prevent an increase in the correction error of the odometry gain G due to an excessive correction amount by the odometry gain G.
Further, the odometry gain G is recorded at an interval determined by the first distance difference D1, which is the difference between the first travel distance L _odm and the second travel distance L _map . As a result, the odometry gain G can be recorded at an appropriate interval, and the appropriate odometry gain G can be set while suppressing the amount of data, so that the travel distance of the host vehicle can be accurately calculated. become able to. The odometry gain G does not need to be recorded on the map, and the recording method and medium are not limited as long as the position information of the position using the odometry gain G can be specified.

(Modification)
(1) As in the first embodiment, the vehicle determines a distance at which one of the first travel distance L _odm and the second travel distance L _map is maximum while the first distance difference D1 is equal to or less than the predetermined threshold Th1. Each time the vehicle travels, the odometry gain G may be calculated and the landmark position information and the odometry gain G may be recorded on the second map 34.
Further, a threshold value Th2 that is sufficiently smaller than the threshold value Th1 is set, and each time the vehicle travels a distance where the first distance difference D1 exceeds the threshold value Th2, the odometry gain G is calculated and the landmark position information and the odometry gain G are calculated. You may record on the 2nd map 34.
Further, a threshold value Th4 that is sufficiently smaller than the threshold value Th3 is set, and each time the vehicle travels a distance where the second distance difference D2 exceeds the threshold value Th4, the odometry gain G is calculated and the landmark position information and the odometry gain G are calculated. You may record on the 2nd map 34.
(2) As in the modification of the first embodiment, the condition determination unit 71 may change the threshold Th3 according to the driving environment.

  It goes without saying that the present invention includes various embodiments not described herein. Therefore, the technical scope of the present invention is defined only by the invention specifying matters according to the scope of claims reasonable from the above description.

DESCRIPTION OF SYMBOLS 1 ... Driving assistance device, 2 ... Self-position estimation device, 10 ... Ambient environment sensor group, 11 ... Image pick-up device, 12 ... Distance measuring device, 20 ... Vehicle sensor group, 21 ... Accelerator sensor, 22 ... Steering sensor, 23 ... Brake Sensor, 24 ... Vehicle speed sensor, 25 ... Acceleration sensor, 26 ... Gyro sensor, 27 ... Wheel speed sensor, 28 ... Other sensor, 30 ... Communication device, 31 ... Map storage unit, 32 ... First map, 33 ... GPS receiver 34 ... Second map 35 ... Elevation DB 40 ... Controller
DESCRIPTION OF SYMBOLS 41 ... Processor, 42 ... Memory | storage device, 50 ... Vehicle travel controller, 51 ... Vehicle control actuator group, 52 ... Steering actuator, 53 ... Accelerator opening actuator, 54 ... Brake control actuator, 60 ... Odometry calculation part, 61 ... Trajectory calculation , 62 ... white line detection unit, 63 ... landmark detection unit, 64 ... matching unit, 65 ... landmark correction unit, 66 ... trajectory correction unit, 67 ... landmark recording unit, 70 ... second travel distance calculation unit, 71 ... Condition determining unit, 72 ... Recording unit, 80 ... Vehicle, 90 ... Third travel distance calculating unit, 91 ... Odometry gain calculating unit

Claims (14)

Calculating a distance difference that is a difference between a first travel distance that is a length of a travel locus of the vehicle and a second travel distance that is a length of the travel locus on a two-dimensional map;
Detecting landmarks around the vehicle,
A distance interval for recording the detected position information of the landmark is determined according to the distance difference.
A map information learning method characterized by that. Records position information of landmarks detected around the vehicle at a distance interval determined based on the distance difference.
The map information learning method according to claim 1, wherein:   The map information learning method according to claim 2, wherein, as the distance interval, the landmark position information is recorded when the distance difference exceeds a predetermined threshold.   The position information of the landmark is recorded at a distance interval determined based on the distance difference and one of the first travel distance and the second travel distance. Map information learning method.   The distance is detected around the vehicle when the distance between the first travel distance and the second travel distance reaches a predetermined value while the distance difference is not more than a predetermined threshold. 5. The map information learning method according to claim 4, wherein position information of a landmark is recorded. Calculating the first mileage based on the rotational speed of the wheel;
The map information learning method according to any one of claims 1 to 5, wherein a coefficient used for correcting an error occurring in the first travel distance is recorded.   The map information learning method according to claim 6, wherein the coefficient is a coefficient by which the first travel distance is multiplied in order to calculate the second travel distance from the first travel distance.   The map information learning method according to claim 6 or 7, wherein the coefficient is recorded at a distance interval determined based on the distance difference. Calculating a first distance difference that is a difference between the first travel distance and the second travel distance;
Calculating a second distance difference that is a difference between a third travel distance that is a length of a travel locus of the vehicle calculated based on a detection result of the surrounding environment of the vehicle and the second travel distance;
The map information learning method according to claim 6 or 7, wherein the coefficient is recorded at a distance interval determined based on the first distance difference and the second distance difference.   The map information learning method according to any one of claims 1 to 9, wherein the distance difference is calculated based on an elevation difference of the traveling path of the vehicle.   The map information learning method according to claim 10, wherein the altitude difference is measured using a satellite positioning signal and an altitude database, or a gyro sensor.   12. The map information learning method according to claim 1, wherein the recording of the landmark position information is used to complement map data used for driving support of the vehicle.   The landmark position information detected in a predetermined width region around the predetermined range is recorded when a landmark within the predetermined range around the vehicle is not detectable. The map information learning method as described in any one of Claims. A sensor that detects landmarks around the vehicle;
A distance difference which is a difference between a first travel distance which is a length of the travel locus of the vehicle and a second travel distance which is a length of the travel locus on a two-dimensional map is calculated, and the sensor A controller that determines a distance interval for recording the position information of the detected landmark in a predetermined storage device according to the distance difference;
A map information learning device comprising:

Images