Position estimation device, position estimation method and position estimation program

Abstract

To provide a position estimation device that improves the accuracy of estimating a road on which a position estimation target vehicle is traveling.SOLUTION: A self-position estimation ECU 20 as a position estimation device estimates a position of an own vehicle EV as a position estimation target vehicle by using a PD map generated by collecting probe data from multiple vehicles. The self-position estimation ECU 20 includes: a reliability calculation unit 21 which calculates road reliability indicating whether the own vehicle EV exists for each road that has been already maintained in the PD map; and a travel road estimation unit 23 which estimates a road on which the own vehicle EV is traveling by including the road that has not been maintained in the assumption in the PD map on the basis of the calculation result of the road reliability.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The disclosure of this specification relates to a technology for estimating the position of a vehicle using a PD map generated by collecting probe data from multiple vehicles. [Background technology]

[0002] Patent Document 1 discloses a technique for estimating which road a vehicle is traveling on among an elevated road and a road under an elevated road that are parallel to each other, by using a digital road map database of a navigation device. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2007-71539 A Summary of the Invention [Problem to be solved by the invention]

[0004] Recently, there has been a study on generating PD maps by collecting probe data from multiple vehicles and using them for vehicle travel. On the one hand, the PD map can reflect the latest data collected from vehicles that have actually traveled on the roads, but on the other hand, in order to reduce the management costs of the PD map, it is possible to consider not including some of the roads that actually exist on the PD map.

[0005] In this way, there is a possibility that developed and undeveloped roads coexist on the PD map. Under such circumstances, it is necessary to improve the accuracy of estimating the road on which the target vehicle is traveling.

[0006] One of the objectives of the disclosure of this specification is to provide a position estimation device, a position estimation method, and a position estimation program that improve the accuracy of estimating a road on which a vehicle whose position is to be estimated is traveling. [Means for solving the problem]

[0007] One aspect disclosed herein is a location estimation device that estimates a location of a target vehicle (EV) by using a PD map generated by collecting probe data from a plurality of vehicles, the device comprising: a reliability calculation unit (21, 121) that calculates a road reliability indicating whether or not a vehicle to be estimated is present on one or more roads that have been prepared on a PD map; The system includes a road travel estimation unit (23, 123) that estimates the road on which the vehicle is traveling, taking into consideration roads that are not yet developed in the PD map, based on the calculation results of the road reliability.

[0008] Another disclosed aspect is a location estimation method executed by at least one processor to estimate a location of a location estimation target vehicle (EV) using a PD map generated by collecting probe data from a plurality of vehicles, the method comprising: Calculating road reliability indicating whether or not a vehicle to be estimated is present on one or more roads that have been developed on a PD map; The method includes estimating the road on which the vehicle is traveling, taking into consideration roads that are not yet developed on the PD map, based on the calculation results of the road reliability.

[0009] Another disclosed aspect is a location estimation program that estimates a location of a location estimation target vehicle (EV) using a PD map generated by collecting probe data from a plurality of vehicles, the program comprising: At least one processor has Calculating road reliability indicating whether or not a vehicle to be estimated is present on one or more roads that have been developed on a PD map; The system is configured to estimate the road on which the vehicle to be located is traveling, taking into consideration unimproved roads in the PD map, based on the calculation results of the road reliability.

[0010] According to these aspects, road reliability is calculated for roads that have been developed on a PD map. Then, the estimation of the road on which the location estimation target vehicle is traveling, which is performed based on this calculation result, is performed not only for roads that have been developed on a PD map, but also for roads that have not been developed on a PD map. Therefore, even if the vehicle is traveling on a road that has not been developed on a PD map, it is possible to suppress the possibility of erroneously estimating that the vehicle is traveling on a road that has been developed on a PD map. Therefore, it is possible to improve the estimation accuracy of the road on which the location estimation target vehicle is traveling.

[0011] In addition, the symbols in parentheses included in the claims, etc., are intended to exemplify the corresponding relationship with parts of the embodiments described below, and are not intended to limit the technical scope. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a diagram showing a vehicle equipped with a self-position estimation ECU. [Diagram 2] FIG. 2 is a diagram showing a schematic configuration of a vehicle system and a self-position estimation ECU. [Diagram 3] 6 is a flowchart showing an example of processing performed by a self-position estimation ECU. [Figure 4] 6 is a flowchart showing an example of processing performed by a self-position estimation ECU. [Diagram 5] 6 is a flowchart showing an example of processing performed by a self-position estimation ECU. [Figure 6] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 7] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 8] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 9] FIG. 1 is a diagram showing an example of parallel roads. [Figure 10] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 11] FIG. 1 is a diagram showing an example of parallel roads. [Figure 12] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 13]FIG. 1 is a diagram showing an example of parallel roads. [Figure 14] FIG. 4 is a diagram for explaining an example of calculation of road reliability. [Figure 15] 6 is a flowchart showing an example of processing performed by a self-position estimation ECU. [Figure 16] FIG. 2 is a diagram showing a schematic configuration of a vehicle system and a self-position estimation ECU. [Figure 17] 6 is a flowchart showing an example of processing performed by a self-position estimation ECU. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, a number of embodiments will be described with reference to the drawings. In addition, by assigning the same reference numerals to corresponding components in each embodiment, duplicated descriptions may be omitted. When only a part of the configuration is described in each embodiment, the configuration of the other embodiment described above can be applied to the other parts of the configuration. In addition to the combination of configurations explicitly stated in the description of each embodiment, configurations of multiple embodiments can be partially combined together even if not explicitly stated, as long as there is no particular problem with the combination.

[0014] (First embodiment) The vehicle system 10 of this embodiment is mounted on a vehicle as shown in Figs. 1 and 2. The vehicle system 10 can be used in a vehicle (hereinafter, the host vehicle EV) that is capable of autonomously controlling driving (hereinafter, autonomous driving control), for example, with automation levels 2 to 5. In this embodiment, an example will be described in which the vehicle system 10 is mounted on a host vehicle EV with automation levels 3 to 5. In this embodiment, the host vehicle EV corresponds to a vehicle whose position is to be estimated.

[0015] Automation levels are classified into levels 0 to 5, as defined by the SAE, for example. Level 0 is a level where the driver performs all driving tasks without system intervention. Level 0 corresponds to so-called fully manual driving. Level 1 is a level where the system assists with either steering or acceleration / deceleration. Level 1 corresponds to so-called driving assistance. Level 2 is a level where the system assists with both steering and acceleration / deceleration. Level 2 corresponds to partial driving automation.

[0016] Level 3 is a level where the system can perform all driving tasks under certain conditions, and the driver performs driving operations in an emergency. Level 3 corresponds to so-called conditional driving automation. Level 4 is a level where the system can perform all driving tasks except in certain situations such as incompatible roads and extreme environments. Level 4 corresponds to so-called high driving automation. Level 5 corresponds to so-called full driving automation. Levels 3 to 5 may be classified as automated driving.

[0017] As shown in FIG. 2, the vehicle system 10 includes an environment recognition unit 11, a vehicle state detection unit 13, a satellite positioning unit 15, a PD map storage unit 17, an automatic driving ECU 30, and a self-position estimation ECU 20. ECU is an abbreviation for Electronic Control Unit. These components may be communicatively connected to a communication bus of an in-vehicle network mounted on the vehicle EV. These nodes connected to the communication bus can communicate with each other. Certain nodes among these components may be directly electrically connected to each other by a wire harness or the like, and can communicate with each other without going through the communication bus.

[0018] The environment recognition unit 11 recognizes the external environment of the host vehicle EV. The environment recognition unit 11 includes a plurality of autonomous sensors 12. The autonomous sensors 12 include at least one of a camera 12a, a LiDAR (Light Detection and Ranging / Laser Imaging Detection and Ranging) 12b, a millimeter wave radar 13c, an ultrasonic sonar, and the like. These autonomous sensors 12 are combined and mounted on the host vehicle EV to monitor each direction of the host vehicle EV.

[0019] The autonomous sensors 12 preferably include an image sensor capable of generating an image of the surroundings of the host vehicle EV. The image sensor may be a camera 12a that sequentially captures images of the surroundings of the host vehicle EV. The camera 12a is mounted, for example, on the upper end of the front windshield of the host vehicle EV. The camera 12a outputs at least one of the image and the image recognition result as image data. The image may be a color image or a monochrome image. The image recognition result may be a result of identifying the type of object reflected in the image by, for example, semantic segmentation using a trained model by a neural network.

[0020] The image sensor may be a LiDAR 12b. The LiDAR 12b is mounted on, for example, the roof, the upper end of the front windshield, the front bumper, etc. of the vehicle. The LiDAR 12b scans a measurement area with laser light and detects the reflected light reflected by an object in the measurement area as point cloud data. The detected luminance value of each point according to the scanning position is converted into two-dimensional data as each pixel value, so that the LiDAR 12b can also output image data.

[0021] The environment recognition unit 11 may provide the detection results of each autonomous sensor 12 as sensing information directly to the self-location estimation ECU 20. The environment recognition unit 11 may provide a result of fusing the detection results of each autonomous sensor 12 to the self-location estimation ECU 20 as sensing information.

[0022] The host vehicle state detection unit 13 includes a plurality of vehicle state sensors. The vehicle state sensors may include a vehicle speed sensor, a steering angle sensor, an acceleration sensor, a yaw rate sensor, and the like. The vehicle speed sensor is a sensor that detects the traveling speed of the host vehicle EV. The steering angle sensor is a sensor that detects the steering angle of the host vehicle EV. The yaw rate sensor is a sensor that detects the angular velocity of the host vehicle EV. The shift position sensor is a sensor that detects the shift position of the transmission of the host vehicle EV. The host vehicle state detection unit 13 detects information related to the state of the host vehicle, including, for example, the vehicle speed and yaw rate, and provides the information to the host position estimation ECU 20.

[0023] The satellite positioning unit 15 is mainly configured with a GNSS receiver that receives positioning signals transmitted from, for example, multiple artificial satellites (positioning satellites). GNSS is an abbreviation for Global Navigation Satellite System. The GNSS receiver can receive positioning information from each positioning satellite of at least one satellite positioning system among satellite positioning systems such as GPS, GLONASS, Galileo, ITNSS, QZSS, and Beidou.

[0024] The PD map storage unit 17 is a database that stores PD map data that can be used in the vehicle system 10. The PD map storage unit 14 includes at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium. The PD map storage unit 17 is configured to be able to update the PD map by successively downloading the latest PD map from a map server 90 that is communicatively connected to the vehicle system via a communication device provided in the vehicle system 10.

[0025] The PD map is a map generated by the map server 90 by collecting probe data (hereinafter, referred to as PD) from a plurality of vehicles that are communicatively connected to the map server 90 and configured to be able to upload the PD to the map server 90. For example, the PD map includes road network data, lane network data, feature data, and POI (Point of Interest) data.

[0026] The road network data includes the position coordinates of the road, the length of the road, the number of lanes, information on the connected roads, etc. The lane network data is associated with the road network data, and includes the position coordinates of each lane, the length of the lane, etc. The road network data and lane network data are prepared for target roads that are predetermined to be developed in consideration of the management costs of the PD map, for example, among the actual roads. In other words, the road network data and lane network data are not prepared for undeveloped roads other than the target roads. In this way, roads that have been developed on the PD map and roads that have not been developed on the PD map coexist.

[0027] Unimproved roads are determined according to the specifications of the PD map. Unimproved roads may be, for example, expressways, roads other than national highways and trunk roads (for example, side roads to national highways), roads with two lanes or less, etc.

[0028] The feature data includes a coordinate point cloud of road edges, road marking information, and information on three-dimensional objects, such as traffic signs, commercial signs, poles, guardrails, curbs, trees, utility poles, and traffic lights installed along the roads.

[0029] The POI data is data on map elements other than those mentioned above, which are associated with the road network data and the lane network data. The map elements include static map elements such as toll booths, tunnels, branching points, and merging points. The static map elements require updating, for example, every week to every month. The map elements also include dynamic map elements such as congested sections, construction sections, broken-down vehicles, fallen objects, accident locations, and lane restrictions. The dynamic map elements require updating, for example, every few minutes to several hours.

[0030] The PD map storage unit 17 provides the latest PD map data downloaded from the map server 90 to the self-position estimation ECU 20 .

[0031] The autonomous driving ECU 30 is an electronic control device configured to be capable of executing driving tasks of automation levels 3 to 5 on behalf of the driver. The autonomous driving ECU 30 is capable of autonomously driving the host vehicle EV by autonomously controlling motion actuators such as the power train, steering, and brakes. In the autonomous driving, the autonomous driving ECU 30 refers to sensing information acquired from the environment recognition unit 11, and further to position information of the host vehicle EV acquired from the self-position estimation ECU 20, etc.

[0032] The host vehicle EV position estimation ECU 20 is an electronic control device that successively estimates the position of the host vehicle EV based on information input from the environment recognition unit 11, the host vehicle state detection unit 13, the satellite positioning unit 15, and the PD map storage unit 17. The estimation of the host vehicle EV position here includes the estimation of the road on which the host vehicle EV is traveling.

[0033] The estimation information estimated by the self-position estimation ECU 20 is provided to, for example, an automatic driving ECU and used for automatic driving. For example, when multiple roads with different speed limits run parallel to each other, such as an expressway and a general road, if the road on which the host vehicle EV is traveling is erroneously estimated, there is a possibility that the automatic driving ECU 30 will automatically drive the host vehicle EV while misidentifying the speed limit. For this reason, the self-position estimation ECU 20 is required to have estimation accuracy that suppresses malfunctions during automatic driving.

[0034] The self-location estimation ECU 20 has at least one memory and one processor. The memory may be at least one type of non-transient substantial storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium, that non-temporarily stores computer programs and data that can be read by the processor. Furthermore, a rewritable volatile storage medium, such as a random access memory (RAM), may be provided as the memory. The processor includes at least one type of core, such as a central processing unit (CPU), a graphics processing unit (GPU), or a reduced instruction set computer (RISC)-CPU.

[0035] The self-location estimation ECU 20 has a plurality of processing units that realize a location estimation function by a processor executing a computer program. Specifically, as shown in FIG. 1, the self-location estimation ECU 20 includes a reliability calculation unit 21 and a roadway estimation unit 23.

[0036] The reliability calculation unit 21 calculates road reliability for each of one or more roads that have been prepared in the PD map and that are candidate roads on which the host vehicle EV may be traveling. The road reliability here is a value indicating whether the host vehicle EV is present on the road, and is a value that quantifies the possibility of the host vehicle EV being present. The road reliability is a value calculated comprehensively from a plurality of features. The road reliability may be expressed as a probability that the host vehicle EV is traveling on the road. The plurality of features here may include at least one of the changes in the number of lanes, the position of the host vehicle EV, and the trajectory of the host vehicle EV. The plurality of features may further include other aspects, such as the degree of coincidence of targets. The targets may be signs such as billboards, road markings, and structures such as buildings and bridge piers.

[0037] The reliability calculation unit 21 sequentially calculates the reliability based on each feature. After that, the reliability calculation unit 21 substitutes the individual reliability based on each feature into a predetermined formula to calculate the overall road reliability. The formula here may be an appropriate formula such as a formula based on a parallel model or a formula based on a majority model.

[0038] Here, an example of a method for calculating the reliability based on a change in the number of lanes by the process executed by the reliability calculation unit 21 will be described with reference to the flowchart of FIG.

[0039] The processes of S10 to S13 are the processes of the recognition result. In S10, the reliability calculation unit 21 acquires sensing information for a predetermined period up to the present from the environment recognition unit 11, and acquires the recognition result of the number of lanes from this sensing information. The recognition result of the number of lanes is, for example, the result of the number of lanes of the road on which the host vehicle EV is traveling, recognized by the autonomous sensor 12 such as the camera 12a. After the process of S10, the process proceeds to S11.

[0040] In S11, the reliability calculation unit 21 temporarily stores the number of lanes acquired in S10 in a memory. After the process of S11, the process proceeds to S12.

[0041] In S12, the reliability calculation unit 21 generates time-series lane number change data from the number of lanes stored in the memory. The time-series lane number change data is data that indicates a change over time in the number of lanes of the road on which the host vehicle EV is traveling. After the process of S12, the process proceeds to S13.

[0042] In S13, the reliability calculation unit 21 acquires a PD map, and acquires information on the number of lanes of the road for which the reliability is to be calculated from the PD map. After the process of S13, the process proceeds to S14.

[0043] In S14, the reliability calculation unit 21 temporarily stores the number of lanes acquired in S13 in a memory. After the process of S14, the process proceeds to S15.

[0044] In S15, the reliability calculation unit 21 generates time-series lane number change data from the number of lanes stored in the memory. The time-series lane number change data calculated in S15 is data obtained by converting information on the number of lanes acquired from the PD map into a format that can be compared with the time-series lane number change data calculated in S12. After the process of S15, the process proceeds to S16.

[0045] In S16, the reliability calculation unit 21 compares the time-series lane number change data obtained from the recognition result with the time-series lane number change data obtained from the PD map. The comparison here may be a calculation of the difference between the data. After the process of S16, the process proceeds to S17.

[0046] In S17, the reliability calculation unit 21 calculates the reliability based on the change in the number of lanes from the comparison result (for example, the difference). The reliability has a lower value as the difference is larger, and a higher value as the difference is smaller. The series of processes ends with S17.

[0047] The order of the processes in S10 to 12 and the processes in S13 to 15 may be interchanged. The processes in S10 to 12 and the processes in S13 to 15 may be executed in parallel at the same time.

[0048] Next, an example of a method for calculating the reliability based on the position of the host vehicle EV by the process executed by the reliability calculation unit 21 will be described with reference to the flowchart of FIG.

[0049] In S20, the reliability calculation unit 21 calculates the position of the host vehicle EV based on the positioning information acquired from the satellite positioning unit 15 and the vehicle speed, yaw rate, and the like acquired from the host vehicle state detection unit 13. The position of the host vehicle EV here may be the position of the host vehicle EV expressed, for example, by latitude and longitude. After processing S20, the process proceeds to S21.

[0050] In S21, the reliability calculation unit 21 acquires road information of a PD map that is extracted as a candidate road on which the host vehicle EV is traveling based on the position of the host vehicle EV. After the process of S21, the process proceeds to S22.

[0051] In S22, the reliability calculation unit 21 compares the position of the vehicle EV with the positions of the candidate roads, and calculates the difference between the positions. After the process of S22, the process proceeds to S23.

[0052] In S23, the reliability calculation unit 21 calculates the reliability based on the position of the host vehicle EV using the difference calculated in S22. The reliability takes a lower value as the difference increases, and takes a higher value as the difference decreases. The series of processes ends with S23.

[0053] Next, an example of a method for calculating the reliability based on the trajectory of the host vehicle EV by the process executed by the reliability calculation unit 21 will be described with reference to the flowchart of FIG.

[0054] In S30, the reliability calculation unit 21 calculates the position of the host vehicle EV, similar to S20. After the process of S30, the process proceeds to S31. In S31, the reliability calculation unit 21 acquires road information of the PD map, similar to S21. After the process of S31, the process proceeds to S32.

[0055] In S32, the reliability calculation unit 21 calculates the trajectory of the host vehicle EV from the time-series information of the position of the host vehicle EV acquired in S30. The trajectory of the host vehicle EV is estimated to be the shape of the road on which the host vehicle EV is traveling. After the process of S32, the process proceeds to S33.

[0056] In S33, the reliability calculation unit 21 calculates the shape of the road from the road information acquired in S31. After the process of S33, the process proceeds to S34.

[0057] In S34, the reliability calculation unit 21 compares the road shape estimated from the trajectory of the vehicle EV with the road shape calculated from the PD map. Then, the reliability calculation unit 21 calculates the reliability based on the change in the number of lanes from the comparison result (e.g., difference). The reliability has a lower value as the difference is larger, and a higher value as the difference is smaller. A series of processes ends with S34.

[0058] In addition, for example, with respect to overlapping processes such as the relationship between S20 and S30, or the relationship between S21 and S31, the processes may be performed together in advance rather than being performed for each reliability calculation, and the results of the previously performed processes may be referred to in S20 and S30.

[0059] Here, the roads that are maintained in the PD map and are to be calculated as road reliability may be the roads extracted in S31 and S32. Alternatively, the reliability calculation unit 21 may select candidates for the roads to be calculated as road reliability based on the PD map in advance, and execute the processes of the flowcharts shown in Figs. 3 to 5 by limiting the candidates to the selected candidates. The road reliability calculation target may be only roads on which the host vehicle EV may be traveling. Roads on which there is absolutely no possibility that the host vehicle EV may physically be traveling may be excluded in advance from the calculation targets.

[0060] The traveling road estimation unit 23 compares the road reliability of each road that has been developed on the PD map and is the subject of calculation of the road reliability. Based on the comparison result, the traveling road estimation unit 23 estimates the road on which the host vehicle EV is traveling. This estimation is performed including roads that have not been developed on the PD map.

[0061] For example, as shown in Figs. 6 to 8, road 2 and road 3 are targets for calculating road reliability as roads that have been developed on the PD map. Road 1 is a road that has not been developed on the PD map. Whether or not the actual existence of road 1 is recognized by self-position estimation ECU 20 is not a major issue in the road estimation process. Traveling road estimation unit 23 may estimate the road reliability of road 1, which is a road that has not been developed on the PD map, to be the value obtained by subtracting the road reliability of road 2 and the road reliability of road 3 from 100%.

[0062] 6, the road reliability of road 2 is the highest and is greater than a preset threshold value (e.g., 50%). In this case, the traveling road estimation unit 23 outputs the result of estimating that the road on which the host vehicle EV is traveling is road 2 as the estimation result.

[0063] In the example of FIG. 7, the road reliability of all roads (i.e., roads 2 and 3) that have been developed on the PD map and are the subject of calculation of road reliability is less than 50%. In this case, the traveling road estimation unit 23 estimates that the road on which the host vehicle EV is traveling is a road that has not been developed on the PD map. When the existence of road 1 is recognized as an undeveloped road on the PD map, the traveling road estimation unit 23 may output the result of estimating that the road on which the host vehicle EV is traveling is road 1 as the estimation result. When the existence of an undeveloped road is not recognized on the PD map, the traveling road estimation unit 23 may output the result of estimating that the road on which the host vehicle EV is traveling is not developed on the PD map as the estimation result, or may output the result of "unknown" which means that estimation is not possible as the estimation result.

[0064] 8, the road reliability values ​​of the roads 2 and 3 are the same. In this case, the traveling road estimation unit 23 outputs a result that the road on which the host vehicle EV is traveling is "unknown."

[0065] In the following, some more detailed examples will be described with reference to Figs. 9 to 14. In the example of Fig. 9, three roads A1 to A3 run parallel to each other. Road A1 is a side road to roads 2 and 3, and is a road that has not been developed in the PD map. Road A1 has one lane on each side, and the number of lanes does not change. Road A2 is an expressway, and is a road that has been developed in the PD map. Road A2 is an elevated road with two lanes on each side, and is arranged to sandwich road A3 on both sides. Road A3 is a road for entering road A2, which is an expressway, from a general road, and is a road that has been developed in the PD map. Road A3 increases from one lane to two lanes before the toll gate, and returns from two lanes to one lane after the toll gate. The host vehicle EV is traveling on road A3.

[0066] FIG. 10 shows the road reliability of each of the roads A1 to A3. According to this, the road reliability of road 3 is the highest. Road A3 is characterized by a change in the number of lanes, so the reliability based on the change in the number of lanes is calculated to be high. As a result, the road reliability of road A3 is also calculated to be the highest in the overall road reliability. The traveling road estimation unit 23 outputs the result of estimating that the road on which the host vehicle EV is traveling is road A3 as an estimation result.

[0067] In the example of FIG. 11, three roads B1 to B3 run parallel to each other. Road B1 is a road that has been prepared on a PD map. Road B1 has one lane, and solid white lines are provided as road markings on both sides of the road edge. Road B2 is a road that has been prepared on a PD map. Road B2 has three lanes, and solid white lines are provided as road markings on both sides of the road edge, as well as two broken white lines (lane dividing lines) sandwiched between the solid white lines on both sides as road markings. Road B3 is a road that has been prepared on a PD map. Road B3 has two lanes, and solid white lines are provided as road markings on both sides of the road edge, as well as one broken white line (lane dividing line) sandwiched between the solid white lines on both sides as road markings.

[0068] FIG. 12 shows the road reliability of each of the roads B1 to B3. According to this, the road reliability of the road 2 is the highest. Since the road B2 is closer to the position of the host vehicle EV calculated from the positioning information from the satellite positioning unit 15 than the roads B1 and B3, the reliability based on the position of the host vehicle EV of the road B2 is calculated to be high. In addition, the matching rate between the road markings recognized by the autonomous sensor 12 such as the camera 12a and the road surface markings stored in the PD map is higher for the road B2 than for the roads B1 and B3, so the reliability based on the landmarks of the road B2 is also calculated to be high. As a result, the overall reliability of the road B2 is calculated to be the highest. Since the road reliability of B2 exceeds 50%, it is determined that there is a low possibility that the host vehicle EV is traveling on a road that is not maintained on the PD map. The traveling road estimation unit 23 outputs the result of estimating that the road on which the host vehicle EV is traveling is the road B2 as an estimation result.

[0069] In the example of FIG. 12, there are road C1 and road C2 branching off from road C1. Road C1 is a general road and has been prepared on a PD map. Road C1 is a straight road. Road C2 connects road C1, which is a general road, with an expressway and has been prepared on a PD map. Road C2 has an inclined portion C2a extending from road C1 at an angle inclined to road C1, and a parallel portion C2b extending from inclined portion C2a substantially parallel to road C1.

[0070] 13 shows the road reliability of each of the roads C1 to C2. According to this, the reliability of the road C2 is the highest. The matching rate between the curved trajectory of the host vehicle EV and the shape of the road is higher for the road C2 than for the road C1. Therefore, the reliability of the road C2 based on the trajectory of the host vehicle EV is calculated to be high.

[0071] Here, an example of a method for estimating the position of the host vehicle EV by processing executed by the self-position estimation ECU 20 will be described with reference to the flowchart of FIG. 15. In S40, the reliability calculation unit 21 calculates reliability based on each feature. In S41 after processing of S40, the reliability calculation unit 21 calculates an overall road reliability from the reliability based on each feature. In S42 after processing of S41, the traveling road estimation unit 23 compares the road reliability between roads. In S43 after processing of S42, the traveling road estimation unit 23 outputs an estimation result of the road on which the host vehicle EV is traveling based on the comparison. A series of processes ends with S43.

[0072] According to the first embodiment described above, road reliability is calculated for roads that have been prepared on a PD map. Then, the estimation of the road on which the host vehicle EV, which is a vehicle to be subjected to position estimation, is traveling based on the calculation result is performed not only on roads that have been prepared on a PD map, but also on the assumption that the vehicle to be subjected to position estimation is traveling on a road that has not been prepared on a PD map. Therefore, even if the host vehicle EV is traveling on a road that has not been prepared on a PD map, the possibility of erroneously estimating that the vehicle is traveling on a road that has been prepared on a PD map can be suppressed. Therefore, the estimation accuracy of the road on which the host vehicle EV is traveling can be improved.

[0073] Furthermore, according to the first embodiment, the road reliability is calculated by comparing at least one of the sensing information from the autonomous sensor 12 mounted on the host vehicle EV, the state of the host vehicle EV, and the positioning information received from the satellite positioning system with the information on the PD map. The road on which the host vehicle EV is traveling is estimated using such road reliability, so that the estimation accuracy is further improved.

[0074] According to the first embodiment, the reliability is calculated based on a plurality of mutually different features, and the road reliability is calculated as a comprehensive reliability based on the reliability based on each feature. By comprehensively judging the plurality of features, the estimation accuracy is further improved.

[0075] Moreover, according to the first embodiment, the recognition result of the number of lanes on which the host vehicle EV is traveling based on the sensing information is compared with information on the number of lanes of the road that has already been prepared in the PD map. When the lane number changes in this comparison match, the reliability based on the lane number change is calculated to be higher than when there is a difference in the lane number change. An overall road reliability is calculated including the reliability based on the lane number change, and the road on which the host vehicle EV is traveling is estimated using this, so that the estimation accuracy is further improved.

[0076] Moreover, according to the first embodiment, the position of the host vehicle EV calculated based on at least one of the sensing information, the state of the host vehicle EV, and the positioning information is compared with the road position on the PD map. The smaller the difference between the position of the host vehicle EV and the road position, the higher the reliability calculated based on the position of the host vehicle EV. An overall road reliability is calculated including the reliability based on the position of the host vehicle EV, and the road on which the host vehicle EV is traveling is estimated using this, so that the estimation accuracy is further improved.

[0077] Moreover, according to the first embodiment, the trajectory of the host vehicle EV, which is obtained from the position of the host vehicle EV calculated based on at least one of the sensing information, the state of the host vehicle EV, and the positioning information, is compared with the road shape of the PD map. The smaller the difference between the trajectory of the host vehicle EV and the road shape, the higher the reliability based on the trajectory of the host vehicle EV is calculated. Since an overall road reliability is calculated including the reliability based on the trajectory of the host vehicle EV, and the road on which the host vehicle EV is traveling is estimated using this, the estimation accuracy is further improved.

[0078] According to the first embodiment, when the road reliability of the specific road that has been prepared on the PD map is higher than the road reliability of other roads and higher than a preset threshold, it is estimated that the host vehicle EV is traveling on the specific road. In this way, it is possible to improve the validity of the estimation for the road that has been prepared on the PD map.

[0079] According to the first embodiment, when the road reliability of all roads that are targets of the calculation of road reliability and that have been included in the PD map are lower than a preset threshold, it is estimated that the host vehicle EV is traveling on a road that has not been included in the PD map. In this way, it is possible to improve the validity of the estimation for roads that have not been included in the PD map.

[0080] According to the first embodiment, when the road reliability of all roads that are subject to the calculation of road reliability and that have been developed in the PD map is lower than a preset threshold, an estimation result is output that the road on which the host vehicle EV is traveling cannot be estimated. In this way, it is possible to avoid outputting an erroneous estimation and to increase the validity of the estimation.

[0081] Second embodiment The second embodiment is a modification of the first embodiment. The second embodiment will be described focusing on the differences from the first embodiment.

[0082] As shown in FIG. 16, the self-position estimation ECU 120 of the second embodiment is configured to further include a reliability update unit 122 as one of a plurality of processing units that realize a position estimation function by the processor executing a computer program (e.g., a position estimation program).

[0083] The reliability update unit 122 holds the road reliability for each road calculated by the reliability calculation unit 121 until the next calculation of road reliability. Then, the reliability update unit 122 calculates an update value of road reliability for each road based on the previous value of road reliability and the current value of road reliability. For example, if the previous value of road reliability is CL, the current value of road reliability is CT, and α is a weighting value, the update value C is expressed by the following formula 1. C = (1-α) CL + α CT (Equation 1)

[0084] Here, the weight value is set according to the movement status (e.g., movement distance) of the host vehicle EV. Specifically, the memory stores a lookup table that defines the relationship between the vehicle speed and the weight value. The memory also stores a lookup table that defines the relationship between the yaw rate and the weight value. The reliability update unit 122 acquires the vehicle speed and the yaw rate from the host vehicle state detection unit 13 and refers to the lookup table. This allows the reliability update unit 122 to calculate a weight value according to the vehicle speed and a weight value according to the yaw rate.

[0085] The weight value is set according to the matching state in the comparison between at least one of the sensing information from the environment recognition unit 11, the vehicle speed and yaw rate acquired from the vehicle state detection unit 13, and the positioning information acquired from the satellite positioning unit 15, and the information on the PD map. This comparison process may use the calculation result used in the comparison process (e.g., S16, S22, S33) executed when the reliability calculation unit 121 calculates the road reliability. Alternatively, the comparison process may be executed from the beginning by the reliability update unit 122. The matching state may be, for example, the number of matching targets when a plurality of targets recognized to be close to the vehicle EV are compared with targets on the PD map. The higher the matching state, the larger the weight value is set. When the matching state is given by the number of matching targets, the larger the number of matching targets, the larger the weight value is set.

[0086] The reliability update value may be calculated by calculating an overall weight value α to be substituted into Equation 1 from a weight value according to the vehicle speed, a weight value according to the yaw rate, and a weight value according to the matching state. According to this, by taking into account the moving state (temporal position change) of the host vehicle EV, such as the vehicle speed and yaw rate, into the matching state, the weight value is set to be larger substantially the longer the state with little change in the matching state continues. Note that the overall weight value α may be the sum of the weight values, an average, or a weighted average.

[0087] The road travel estimation unit 123 of the second embodiment estimates the road on which the host vehicle EV is travelling by using the updated value of the road reliability updated by the reliability update unit 122, instead of the road reliability (i.e., the current value) calculated by the reliability calculation unit 121.

[0088] Here, an example of a reliability update method by the processes executed by the reliability calculation unit 121 and the reliability update unit 122 will be described with reference to the flowchart of FIG.

[0089] In S50, the reliability calculation unit 121 calculates the reliability based on each feature. In S51 after the process of S50, the reliability calculation unit 121 calculates the current value of the overall road reliability from the reliability based on each feature. After the process of S51, the process proceeds to S52.

[0090] In S52, the reliability update unit 122 acquires the previous value of the overall road reliability. In S53 after the process of S52, the reliability update unit 122 calculates a weight value according to the vehicle speed. In S54 after the process of S53, the reliability update unit 122 calculates a weight value according to the yaw rate. In S55 after the process of S54, the reliability update unit 122 calculates a weight value according to the matching state. In S56 after the process of S55, the reliability update unit 122 calculates an updated value of the road reliability using the weight value calculated in S52 to S55. A series of processes ends with S56.

[0091] According to the second embodiment described above, the road reliability is updated based on the previous value of the road reliability calculated last time and the current value of the road reliability calculated this time. Then, the road on which the host vehicle EV is traveling is estimated based on the updated road reliability. By taking the previous value into consideration, it is possible to suppress erroneous estimation when the current value is affected by disturbances, etc.

[0092] Moreover, according to the second embodiment, information obtained from at least one of the sensing information, the state of the host vehicle EV, and the positioning information is compared with information on the PD map. A weight value to be multiplied with the current value is set according to the time during which the matching state between the two continues in this comparison. For example, if the weight of the current value is increased when the time during which there is little change in the matching state continues for a long time, it becomes possible to more quickly obtain an estimation result in which the road on which the host vehicle EV is traveling is identified from a state in which estimation is impossible. In other words, the time from when the host vehicle EV passes a branch to when the road on which the host vehicle EV is traveling is identified can be shortened.

[0093] (Other embodiments) Although several embodiments have been described above, the present disclosure should not be construed as being limited to those embodiments, and can be applied to various embodiments and combinations within the scope not departing from the gist of the present disclosure.

[0094] As another embodiment, the functions of the self-position estimation ECU 20 and the automatic driving ECU 30 may be implemented by a single integrated ECU. In this case, the integrated ECU corresponds to the position estimation device.

[0095] In another embodiment, the position of the host vehicle EV estimated by the self-position estimation ECU 20 may be used for purposes other than the autonomous driving control of the host vehicle EV. For example, the position of the host vehicle EV may be used to provide information or a warning to the driver through an in-vehicle display or the like.

[0096] In another embodiment, the position estimation device mounted on the host vehicle EV may estimate the position of another vehicle as a position estimation target vehicle. In this case, the vehicle speed and yaw rate of the other vehicle may be obtained using V2X communication.

[0097] In another embodiment, the position estimation device may not be mounted on a vehicle, and may remotely estimate the position of the vehicle to be localized.

[0098] The controller and method described herein may be implemented by a special-purpose computer having a processor programmed to execute one or more functions embodied in a computer program. Alternatively, the device and method described herein may be implemented by a special-purpose hardware logic circuit. Alternatively, the device and method described herein may be implemented by one or more special-purpose computers configured by a combination of a processor executing a computer program and one or more hardware logic circuits. The computer program may also be stored on a computer-readable non-transitory tangible recording medium as instructions executed by the computer.

[0099] (Disclosure of technical ideas) This specification discloses a number of technical ideas described in the following paragraphs. Some paragraphs may be described in a multiple dependent form, in which the subsequent paragraph alternatively refers to the preceding paragraph. The paragraphs described in the multiple dependent form define a number of technical ideas.

[0100] <Technical philosophy 1> A position estimation device that estimates a position of a target vehicle (EV) by using a PD map generated by collecting probe data from a plurality of vehicles, a reliability calculation unit (21, 121) that calculates a road reliability indicating whether or not the location estimation target vehicle exists for one or more roads that have been developed on the PD map; A position estimation device comprising a road travel estimation unit (23, 123) that estimates the road on which the vehicle to be estimated is traveling, taking into account roads that are not yet developed on the PD map, based on the calculation result of the road reliability.

[0101] <Technical philosophy 2> The location estimation device described in Technical Idea 1, wherein the reliability calculation unit calculates the road reliability by comparing information obtained from at least one of sensing information from an autonomous sensor (12) mounted on the vehicle to be located, a state of the vehicle to be located, and positioning information received from a satellite positioning system with information on the PD map.

[0102] <Technical philosophy 3> The position estimation device described in Technical Idea 1 or 2, wherein the reliability calculation unit calculates reliability based on a plurality of mutually different features, and calculates the road reliability as an overall reliability based on the reliability based on each of the features.

[0103] <Technical philosophy 4> The feature-based reliability includes a reliability based on a lane number change; The reliability calculation unit is comparing a recognition result of the number of lanes on which the vehicle to be located is traveling based on the sensing information with information on the number of lanes of a road that has already been prepared on the PD map; A position estimation device as described in technical idea 3, which is configured to calculate a higher reliability based on the change in the number of lanes when the change in the number of lanes matches than when a difference occurs in the change in the number of lanes.

[0104] <Technical philosophy 5> the reliability based on the features includes a reliability based on a position of the vehicle to be localized; The reliability calculation unit is comparing a position of the vehicle to be estimated based on at least one of the sensing information, the state of the vehicle to be estimated, and the positioning information with a road position on the PD map; A position estimation device as described in Technical Idea 3 or 4, which is configured to calculate a higher reliability based on the position of the vehicle to be estimated, as the difference between the position of the vehicle to be estimated and the road position is smaller.

[0105] <Technical philosophy 6> the feature-based reliability includes a trajectory-based reliability of the vehicle; The reliability calculation unit is comparing a trajectory of the vehicle to be localized, which is obtained from a position of the vehicle to be localized, calculated based on at least one of the sensing information, the state of the vehicle to be localized, and the positioning information, with a road shape of the PD map; A position estimation device described in any one of technical ideas 3 to 5, which is configured to calculate a higher reliability based on the trajectory of the position estimation target vehicle, the smaller the difference between the trajectory of the position estimation target vehicle and the road shape.

[0106] <Technical philosophy 7> The position estimation device according to any one of Technical Ideas 1 to 6, wherein the driving road estimation unit estimates that the vehicle to be estimated is traveling on the specific road when the road reliability of a specific road that has been developed on the PD map is higher than the road reliability of other roads and is higher than a preset threshold value.

[0107] <Technical philosophy 8> The position estimation device described in any one of Technical Ideas 1 to 7, wherein the driving road estimation unit estimates that the vehicle to be estimated is traveling on a road that is not included in the PD map when the road reliability of all roads that are the subject of calculation of the road reliability and that are included in the PD map is lower than a preset threshold value.

[0108] <Technical philosophy 9> The position estimation device described in any one of Technical Ideas 1 to 8, wherein the driving road estimation unit outputs an estimation result that the road on which the position estimation target vehicle is traveling cannot be estimated if the road reliability of all roads that are subject to calculation of the road reliability and that have been developed in the PD map is lower than a preset threshold value.

[0109] <Technical Thought 10> The position estimation device described in any one of Technical Ideas 1 to 9, wherein the driving road estimation unit outputs an estimation result that the road on which the position estimation target vehicle is traveling cannot be estimated if the road reliability of two roads that are already maintained on the PD map and are the subject of calculation of the road reliability are the same value.

[0110] <Technical Thought 11> a reliability update unit (122) that updates the road reliability based on a previous value of the road reliability previously calculated by the reliability calculation unit and a current value of the road reliability currently calculated by the reliability calculation unit, A position estimation device described in any one of Technical Ideas 1 to 10, wherein the driving road estimation unit estimates the road on which the position estimation target vehicle is driving based on the road reliability updated by the reliability update unit.

[0111] <Technical Thought 12> At least one of the reliability calculation unit and the reliability update unit compares information obtained from at least one of sensing information from an autonomous sensor mounted on the vehicle to be estimated, a state of the vehicle to be estimated, and positioning information received from a satellite positioning system with information on the PD map; The position estimation device according to Technical Idea 11, wherein the reliability update unit sets a weighting value to be multiplied by the current value depending on the time during which the matching state between the two values ​​continues in the comparison. [Explanation of symbols]

[0112] 20: Self-position estimation ECU (position estimation device), 21, 121: Reliability calculation unit, 23, 123: Travel road estimation unit, EV: Vehicle (vehicle to be subjected to position estimation)

Claims

1. A position estimation device that estimates a position of a vehicle (EV) to be estimated by using a PD map generated by collecting probe data from a plurality of vehicles, comprising: a reliability calculation unit (21, 121) that calculates a road reliability indicating whether or not the location estimation target vehicle is present for one or more roads that have been prepared on the PD map; A position estimation device comprising a road travel estimation unit (23, 123) that estimates the road on which the vehicle to be estimated is traveling, taking into account roads that are not yet developed in the PD map, based on the calculation result of the road reliability.

2. 2. The position estimation device according to claim 1, wherein the reliability calculation unit calculates the road reliability by comparing information obtained from at least one of sensing information from an autonomous sensor (12) mounted on the vehicle to be estimated, a state of the vehicle to be estimated, and positioning information received from a satellite positioning system with information on the PD map.

3. The position estimation device according to claim 2 , wherein the reliability calculation unit calculates reliability based on a plurality of mutually different features, and calculates the road reliability as an overall reliability based on the reliability based on each of the features.

4. The feature-based reliability includes a reliability based on a lane number change; The reliability calculation unit is comparing a recognition result of the number of lanes on which the vehicle to be estimated is traveling based on the sensing information with information on the number of lanes of a road that has already been prepared on the PD map; The position estimation device according to claim 3 , configured to calculate a reliability based on the change in the number of lanes to be higher when the change in the number of lanes matches than when a difference occurs in the change in the number of lanes.

5. the reliability based on the features includes a reliability based on a position of the vehicle to be localized; The reliability calculation unit is comparing a position of the vehicle to be estimated based on at least one of the sensing information, the state of the vehicle to be estimated, and the positioning information with a road position on the PD map; 4. The position estimation device according to claim 3, configured to calculate a higher reliability based on the position of the vehicle to be estimated as the difference between the position of the vehicle to be estimated and the road position is smaller.

6. the feature-based reliability includes a trajectory-based reliability of the vehicle; The reliability calculation unit is comparing a trajectory of the vehicle to be estimated, which is obtained from a position of the vehicle to be estimated, calculated based on at least one of the sensing information, the state of the vehicle to be estimated, and the positioning information, with a road shape of the PD map; 4. The position estimation device according to claim 3, wherein the device is configured to calculate a higher reliability based on the trajectory of the vehicle whose position is to be estimated as the difference between the trajectory of the vehicle whose position is to be estimated and the road shape is smaller.

7. 2. The position estimation device according to claim 1, wherein the driving road estimation unit estimates that the position estimation target vehicle is traveling on the specific road when the road reliability of a specific road that has been developed in the PD map is higher than the road reliability of other roads and is higher than a preset threshold value.

8. 2. The position estimation device according to claim 1, wherein the driving road estimation unit estimates that the vehicle to be estimated is traveling on a road that is not included in the PD map when the road reliability of all roads that are the subject of calculation of the road reliability and that are included in the PD map is lower than a preset threshold value.

9. 2. The position estimation device according to claim 1, wherein the driving road estimation unit outputs an estimation result that the road on which the position estimation target vehicle is traveling cannot be estimated when the road reliability of all roads that are subject to calculation of the road reliability and that have been developed in the PD map is lower than a preset threshold value.

10. 2. The position estimation device according to claim 1, wherein the road estimation unit outputs an estimation result that the road on which the vehicle is traveling cannot be estimated when the road reliability of two roads that are subject to calculation of the road reliability and that have been developed on the PD map are the same value.

11. a reliability update unit (122) that updates the road reliability based on a previous value of the road reliability previously calculated by the reliability calculation unit and a current value of the road reliability currently calculated by the reliability calculation unit, The position estimation device according to claim 1 , wherein the road estimation unit estimates the road on which the vehicle is traveling, based on the road reliability updated by the reliability update unit.

12. At least one of the reliability calculation unit and the reliability update unit compares information obtained from at least one of sensing information from an autonomous sensor mounted on the vehicle to be estimated, a state of the vehicle to be estimated, and positioning information received from a satellite positioning system with information on the PD map; The position estimation device according to claim 11 , wherein the reliability update unit sets a weighting value by which the current value is multiplied depending on a duration during which the matched state between the two in the comparison continues.

13. A location estimation method executed by at least one processor for estimating a location of a target vehicle (EV) using a PD map generated by collecting probe data from a plurality of vehicles, the method comprising: Calculating road reliability indicating whether or not the location estimation target vehicle exists for one or more roads that have been developed on the PD map; and estimating a road on which the vehicle to be estimated is traveling, taking into consideration roads that are not yet developed in the PD map, based on a result of the calculation of the road reliability.

14. A position estimation program that estimates a position of a vehicle (EV) to be estimated by using a PD map generated by collecting probe data from a plurality of vehicles, At least one processor, Calculating road reliability indicating whether or not the location estimation target vehicle exists for one or more roads that have been developed on the PD map; A position estimation program configured to execute an estimation of the road on which the vehicle to be estimated is traveling, taking into account undeveloped roads in the PD map, based on the calculation result of the road reliability.

Images