Self-position estimation device

Abstract

To suppress deterioration in estimation accuracy of self-position estimation based on a detection result by a sensor for detecting a surrounding environment.SOLUTION: A self-position estimation device 100, 100a for estimating a self-position of a vehicle V1 on which a plurality of sensors 200 that detect a surrounding environment are mounted includes estimation accuracy specifying units 21-26 that respectively specify estimation accuracy when the self-position is estimated on the basis of detection results of the respective sensors, and an output self-position determination unit 50 that determines an output target self-position from self-positions estimated on the basis of the detection results of the respective sensors by using the respective specified estimation accuracy.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a self-location estimation device. [Background technology]

[0002] Conventionally, techniques have been proposed for estimating a vehicle's own position using the detection results of sensors such as cameras and millimeter-wave radar mounted on a moving object such as a vehicle. Patent Document 1 discloses a method for correcting the results of a vehicle's own position estimation by odometry, which uses the detection values ​​of internal sensors such as a steering angle sensor and a wheel speed sensor, based on the detection results of external sensors such as a front camera and a front radar. Patent Document 2 also discloses a method for estimating a vehicle's own position by so-called sensor fusion, which combines the detection results of a front camera and the detection results of a millimeter-wave radar.

[0003] The accuracy of self-position estimation using sensors that detect the surrounding environment, such as cameras and millimeter-wave radar, can vary depending on the time of day and weather. Patent Document 1 specifies the correction accuracy of the forward camera, forward radar, etc. according to changes in the usage environment, such as the time of day and weather, and if the accuracy is low, the accuracy of the odometry is further specified. If the odometry accuracy is low, the collision avoidance assistance mode is set to a lower level. Patent Document 2 acquires temperature information, time information, etc., and determines whether the environment is adverse, such as fog, afternoon sun, or heavy rain, based on the acquired information. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-139400 A [Patent Document 2] Japanese Patent Publication No. 2022-014729 Summary of the Invention [Problem to be solved by the invention]

[0005] In Patent Document 1, when the correction accuracy of a sensor that detects the surrounding environment, such as a forward camera or a forward radar, is low, the accuracy of odometry is merely determined and the mode of collision avoidance assistance is changed, and a decrease in the accuracy of the estimation of the vehicle's own position cannot be suppressed. Furthermore, in Patent Document 2, the accuracy of the sensor that detects the surrounding environment is merely determined, and a decrease in the accuracy of the estimation of the vehicle's own position using the sensor cannot be suppressed. Therefore, a technology that can suppress a decrease in the accuracy of the estimation of the vehicle's own position based on the detection results of the sensor that detects the surrounding environment is desired. [Means for solving the problem]

[0006] As one aspect of the present disclosure, there is provided a self-position estimation device (100, 100a) for estimating a self-position of a vehicle (V1) equipped with a plurality of sensors (200) for detecting a surrounding environment. The self-position estimation device includes estimation accuracy specification units (21-26) that specify estimation accuracies when estimating the self-position based on detection results from the sensors, and an output self-position determination unit (50) that uses the specified estimation accuracies to determine an output self-position, which is the self-position to be output, from the self-positions estimated based on the detection results from the sensors.

[0007] According to this form of self-position estimation device, the estimation accuracy when estimating self-position based on the detection results of each sensor is identified, and the identified estimation accuracy is used to determine the output self-position, which is the self-position to be output, from among the self-positions estimated based on the detection results of each sensor, thereby suppressing a decrease in the estimation accuracy of self-position estimation based on the detection results of sensors that detect the surrounding environment. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a block diagram illustrating a schematic configuration of a self-position estimation device according to an embodiment of the present disclosure. [Figure 2] 10 is a flowchart showing the procedure of an output self-position determination process in the first embodiment. [Figure 3] 10 is a flowchart showing the detailed procedure of S125. [Figure 4] 10 is a flowchart showing the detailed procedure of S125 in the second embodiment. [Figure 5] FIG. 10 is a block diagram showing a schematic configuration of a self-position estimation device according to a third embodiment. [Figure 6] 11 is a flowchart showing the procedure of an output self-position determination process in the third embodiment. [Figure 7] FIG. 11 is an explanatory diagram showing an example of the setting contents of an estimation accuracy table in the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] A. First embodiment: A1.Device configuration: 1 is mounted on a vehicle V1 and is used to estimate the vehicle's own position. In addition to the self-position estimation device 100, the vehicle V1 is also equipped with a sensor group 200 consisting of a plurality of sensors.

[0010] The sensor group 200 includes a GNSS 210, a periphery vicinity camera 220, a forward camera 230, a radar 240, a sonar 250, a LiDAR 260, and a periphery camera 270. The periphery vicinity camera 220, the forward camera 230, the radar 240, the sonar 250, and the periphery camera 270 can be said to be sensors that detect the surrounding environment.

[0011] The GNSS 210 is a group of devices that make up a global positioning satellite system: GNSS (Global Navigation Satellite System). In this embodiment, the GNSS 210 is a group of devices that make up a GPS (Global Positioning System), and includes a receiver for GPS satellite signals. Note that instead of GPS, any other type of GNSS may be used, such as QZSS (Quasi-Zenith Satellite System), GLONASS (Global Navigation Satellite System), or Galileo.

[0012] The periphery vicinity camera 220 captures an image of the periphery of the vehicle V1 and the vicinity of the vehicle. The imaging range of the periphery vicinity camera 220 is set in advance to capture an image of the road surface within a radius of 5 m (meters) centered on the vehicle V1. The periphery vicinity camera 220 is disposed, for example, on the front grill or side mirror of the vehicle V1. The front camera 230 captures an image of the area in front of the vehicle V1 at a predetermined angle of view. The front camera 230 is disposed, for example, on the upper end of the windshield on the interior side of the vehicle, the front grill, the rooftop, etc.

[0013] The radar 240 outputs a search wave of a predetermined wavelength, such as a millimeter wave or a quasi-millimeter wave, and receives and analyzes the reflected wave to detect the relative position and relative speed of the obstacle with respect to the vehicle V1. The radar 240 is disposed on the front grille, front bumper, or the like of the vehicle V1. The sonar 250 outputs a sound wave as the search wave, and receives and analyzes the reflected wave to detect the relative position and relative speed of the obstacle with respect to the vehicle V1. The sonar 250 is disposed on the front grille, front bumper, or the like of the vehicle V1. The LiDAR 260 outputs a pulsed laser light as the search wave, and receives and analyzes the reflected wave to detect the relative position and relative speed of the obstacle with respect to the vehicle V1.

[0014] Similar to the periphery proximity camera 220, the periphery camera 270 captures images of the periphery of the vehicle V1. However, unlike the periphery proximity camera 220, the periphery camera 270 is configured to be able to capture images of a relatively wide range with high resolution. For example, it has the ability to capture images of two objects that are several tens to several hundreds of meters apart in a manner that allows them to be distinguished. Similar to the periphery proximity camera 220, the periphery camera 270 is disposed on the front grille, side mirrors, etc. of the vehicle V1.

[0015] In this embodiment, the self-location estimation device 100 is configured as an ECU (Electronic Control Unit) including a CPU 110, a ROM 120, and a RAM 130. The self-location estimation device 100 can exchange data with each of the devices 210 to 270 constituting the sensor group 200 via a network such as a CAN (Controller Area Network) provided within the vehicle V1. The ROM 120 may be configured, for example, as a rewritable EEPROM. The CPU 110 loads a program pre-stored in the ROM 120 into the RAM 130 and executes it, thereby functioning as a GNSS position estimation unit 40, a first position estimation unit 11, a second position estimation unit 12, a third position estimation unit 13, a fourth position estimation unit 14, a fifth position estimation unit 15, a sixth position estimation unit 16, a sensor fusion unit 30, and an output self-location determination unit 50.

[0016] The GNSS position estimation unit 40 receives signals output from the GNSS 210 and estimates the position of the vehicle V1 (position of the GNSS 210). As described above, in this embodiment, the GNSS 210 constitutes a GPS, and the GNSS position estimation unit 40 estimates the position (latitude and longitude) of the vehicle V1 using signals output from GPS satellites and received by the GNSS 210. The position estimated by the GNSS position estimation unit 40 may generally include an error of approximately several tens of centimeters to several meters. Information on the self-position estimated by the GNSS position estimation unit 40 is output to the output self-position determination unit 50. Self-position estimation by the GNSS position estimation unit 40 is repeatedly performed at predetermined intervals. For example, it may be performed once every 100 msec (milliseconds).

[0017] The first position estimation unit 11 estimates the self-position of the vehicle V1 using the captured image output from the periphery proximity camera 220. In this embodiment, the first position estimation unit 11 compares the road surface pattern included in the captured image output from the periphery proximity camera 220 with map information pre-stored in a storage device (not shown) mounted on the vehicle V1, and identifies matching locations to estimate the self-position of the vehicle V1. The "road surface pattern" refers to a pattern formed by white lines on the road surface, various signs, manholes, gutters, roadsides, etc. The "map information" stores these road surface patterns in association with position information. In addition to the road surface pattern, the map information also stores point cloud information indicating features such as buildings and guardrails in association with position information. The first position estimation unit 11 includes a first estimation accuracy specification unit 21. The first estimation accuracy specification unit 21 specifies the estimation accuracy of the self-position estimation performed by the first position estimation unit 11. The determination of the estimation accuracy by the first estimation accuracy determination unit 21 and the other estimation accuracy determination units 22 to 26 will be described later.

[0018] The second position estimation unit 12 estimates the self-position of the vehicle V1 using the captured image output from the front camera 230. In this embodiment, when the second position estimation unit 12 receives the captured image output from the front camera 230, it extracts feature points contained in the captured image. Examples of the extracted feature points include edge portions and points consisting of pixels whose brightness is significantly different from that of surrounding pixels. The second position estimation unit 12 then compares the feature point pattern with map information and identifies matching locations to estimate the self-position of the vehicle V1. The second position estimation unit 12 includes a second estimation accuracy specification unit 22. The second estimation accuracy specification unit 22 specifies the estimation accuracy of the self-position estimation performed by the second position estimation unit 12.

[0019] The third position estimation unit 13 estimates the self-position of the vehicle V1 by using the detection result output from the radar 240. Specifically, the third position estimation unit 13 compares the radar point cloud detected by the radar 240 with the map information and identifies a matching location to estimate the self-position of the vehicle V1. The third position estimation unit 13 has a third estimation accuracy specifying unit 23. The third estimation accuracy specifying unit 23 specifies the estimation accuracy of the self-position estimation by the third position estimation unit 13.

[0020] The fourth position estimation unit 14 compares the point cloud output from the sensor fusion unit 30 with map information and identifies matching locations to estimate the self-position of the vehicle V1. The sensor fusion unit 30 receives as input the captured image output from the peripheral proximity camera 220 and the sonar point cloud, which is the detection result output from the sonar 250. The sensor fusion unit 30 identifies (estimates) the point cloud constituting an object using the input captured image and sonar point cloud. By estimating the point cloud using two types of detection results in this way, it is possible to accurately identify (estimate) a point cloud that is likely to actually exist. The fourth position estimation unit 14 includes a fourth estimation accuracy specification unit 24. The fourth estimation accuracy specification unit 24 specifies the estimation accuracy of the self-position estimation performed by the fourth position estimation unit 14.

[0021] The fifth position estimation unit 15 compares the LiDAR point cloud output from the LiDAR 260 with map information and identifies matching locations to estimate the self-position of the vehicle V1. Specifically, the fifth position estimation unit 15 compares the radar point cloud detected by the LiDAR 260 with map information and identifies matching locations to estimate the self-position of the vehicle V1. The fifth position estimation unit 15 has a fifth estimation accuracy specifying unit 25. The fifth estimation accuracy specifying unit 25 specifies the estimation accuracy of the self-position estimation by the fifth position estimation unit 15.

[0022] The sixth position estimation unit 16 estimates the self-position of the vehicle V1 using the captured image output from the peripheral camera 270. In this embodiment, when the sixth position estimation unit 16 receives the captured image output from the peripheral camera 270, it extracts feature points contained in the captured image, similar to the front camera 230. Then, it compares the pattern of feature points with map information and identifies matching locations to estimate the self-position of the vehicle V1. The sixth position estimation unit 16 includes a sixth estimation accuracy specification unit 26. The sixth estimation accuracy specification unit 26 specifies the estimation accuracy of the self-position estimation performed by the sixth position estimation unit 16.

[0023] In this embodiment, the estimation accuracy in self-position estimation using each of the sensors 220 to 270 is set in advance as a fixed value in the ROM 120. In this embodiment, the estimation accuracy is set in advance as a deviation (distance) of the estimated self-position from the actual position, which is specified by an experiment or the like. For example, the self-position estimation is actually performed multiple times using each of the sensors 210 to 270, and the deviation is specified by comparing with the actual position each time. Then, a statistical value such as the average value or maximum value of the deviation may be specified as the estimation accuracy. Specifically, it is set in advance as follows. Peripheral camera: 10cm (centimeters) Front camera: 1m (meter) Radar: 1m Sonar: 30cm LiDAR: 20cm Peripheral camera: 50cm

[0024] In this embodiment, each of the estimation accuracy specifying units 21 to 26 refers to these values ​​set in the ROM 120 to specify the estimation accuracy of the self-position estimation performed by each of the position estimation units 11 to 16.

[0025] The output self-position determiner 50 determines a self-position to be output (hereinafter referred to as "output self-position") from among the self-positions of the vehicle V1 estimated based on the detection results of the sensors 210 to 270. The self-position estimation device 100 outputs the estimated self-position of the vehicle V1. The output destination at this time is, for example, another ECU mounted on the vehicle V1. Examples of such other ECUs include an ECU (ECU for safe driving assistance) that performs collision detection and collision avoidance operations by automatic braking and automatic steering, and an ECU that controls automatic driving when the vehicle V1 performs automatic driving. These ECUs perform collision detection, automatic driving, and the like, using the self-position output from the self-position estimation device 100. The output self-position determiner 50 then determines which of the detection results of the sensors 210 to 270 is used to output the estimated self-position as the self-position used for such purposes. A specific method of determination will be described later. The output self-position determining unit 50 is configured to be able to exchange data with the GNSS position estimating unit 40 and each of the position estimating units 11-16.

[0026] A2. Output self-positioning process: The output self-position determination process shown in FIG. 2 is a process for determining an output self-position, and is executed by the self-position estimation device 100 when the self-position estimation device 100 is powered on.

[0027] In step S50, the output self-position determiner 50 determines whether or not the current execution of the output self-position determination process is the first execution since the self-position estimation device 100 is powered on. Hereinafter, the notation "step S" will be abbreviated to simply "S." If it is determined that this is the first execution (S50: YES), the process proceeds to S105.

[0028] In S105, upon receiving information on the estimated self-position from the GNSS position estimation unit 40, the output self-position determination unit 50 transmits this information to each of the position estimation units 11-16.

[0029] In S110, each of the position estimation units 11 to 16 estimates its own position. At this time, when matching with map information, each of the position estimation units 11 to 16 narrows the matching range in the map information by using the self-position estimated by the GNSS 210. With this configuration, it is possible to prevent a significant decrease in estimation accuracy caused by a match with a road surface pattern or point cloud pattern that happens to exist in a location far away.

[0030] In S115, each of the estimation accuracy specifying units 21 to 26 specifies the estimation accuracy of the self-position by each of the position estimation units. Specifically, as described above, each of the estimation accuracy specifying units 21 to 26 specifies the estimation accuracy by reading out the estimation accuracy set in the ROM 120.

[0031] In S120, each of the position estimation units 11 to 16 transmits information on the estimated self-position and the identified estimation accuracy to the output self-position determination unit 50. If it is determined that the estimated self-position is significantly different from the estimated self-position based on the GNSS 210 transmitted in S105, as described above, the estimated self-position and the estimation accuracy may not be transmitted.

[0032] In S125, the output self-position determiner 50 determines an output self-position from among the estimated self-positions estimated by the position estimators 11-16, using the estimation accuracy received from each of the position estimators 11-16.

[0033] As shown in FIG. 3, S125 more specifically includes S205 and S210. In S205, the output self-location determiner 50 identifies a sensor that has been able to estimate its self-location. In S210, the output self-location determiner 50 determines, as the output self-location, a self-location estimated based on the detection result of the sensor with the highest estimation accuracy among the sensors that have been able to estimate its self-location. "Self-location estimation has been able to be ... Furthermore, for example, if the driving environment is, for example, inside a tunnel on a highway, and only walls are visible in the captured image of the surrounding area, making it impossible to extract feature points, and a matching location cannot be identified even when compared with map information, the detection result cannot be output to the output self-position determination unit 50, and it is determined that "self-position estimation is not possible."

[0034] For example, if the self-position can be estimated for all sensors 210-270, the self-position estimated based on the detection result by the peripheral proximity camera 220 (10 cm), which has the highest accuracy among the above-mentioned preset estimation accuracies, is determined as the output self-position. Also, if the self-position cannot be estimated using the peripheral proximity camera 220 due to a malfunction, snowfall, or the like, the self-position estimated based on the detection result by the LiDAR 260 (20 cm), which has the highest accuracy among the sensors other than the peripheral proximity camera 220, is determined as the output self-position and output to another ECU. After completion of S210, the process returns to S50 as shown in FIG. 2.

[0035] Once the initial output self-location determination is completed as described above, from the second time onward, it is determined in S50 that this is not the first execution (S50: NO). In this case, the output self-location determination unit 50 transmits the previously determined output self-location to each of the position estimation units 11-16 (S100). Then, in this case, in the above-mentioned S110, each of the position estimation units 11-16 performs matching by narrowing the matching range in the map information using the previously determined output self-location notified in S100. This increases the possibility of a match through matching, and improves the estimation accuracy of the self-location.

[0036] According to the self-position estimation device 100 of the first embodiment described above, the estimation accuracy when estimating the self-position based on the detection results of each sensor 210 to 270 is identified, and the output self-position is determined from among the self-positions estimated based on the detection results of each sensor 210 to 270 using each of the identified estimation accuracies, thereby making it possible to suppress a decrease in the estimation accuracy of the self-position estimation based on the detection results of sensors that detect the surrounding environment.

[0037] B. Second embodiment: The configuration of the self-location estimation device 100 of the second embodiment is the same as the configuration of the self-location estimation device 100 of the first embodiment, so the same components are given the same reference numerals and detailed description thereof will be omitted.

[0038] The output self-position determination process of the second embodiment differs from the output self-position determination process of the first embodiment in the detailed procedure of S125, but the other procedures are the same. As shown in Fig. 4, S125 of the second embodiment differs from S125 of the first embodiment in that S220, S225, S230, S235, S240, S245, and S250 are executed instead of S210.

[0039] After S205 is completed, in S220, the output self-position determiner 50 identifies an estimated self-position based on the detection result of the sensor with the highest estimation accuracy among the sensors that were able to estimate the self-position. This S220 differs from S210 in the first embodiment only in that the output self-position is not determined as the final output self-position.

[0040] In S225, the output self-position determining unit 50 assigns "1" to the variable N. The variable N indicates the order of high estimation accuracy.

[0041] In S230, the output self-location determiner 50 determines whether the estimated self-location (hereinafter referred to as "estimated self-location (N)") estimated by the Nth sensor is within the estimation error range of the N+1th sensor. "Within the estimation error range of the N+1th sensor" refers to a circular range having a predetermined radius centered on the self-location estimated based on the detection result of the N+1th sensor. This "predetermined radius" is set in advance for each sensor and stored in the ROM 120. Specifically, in this embodiment, the "predetermined radius" is set to the same numerical value as the specific estimation accuracy set for each of the sensors 210 to 270 described above. For example, it is determined whether the estimated self-location estimated by the surrounding vicinity camera 220, which is the sensor with the first estimation accuracy, is within the estimation error range of the LiDAR 260, which is the sensor with the second estimation accuracy. In this case, the estimation error range of the LiDAR 260 refers to a circular range having a radius of 20 cm centered on the estimated self-location estimated by the LiDAR 260. Therefore, it is determined whether or not the estimated self-position estimated based on the detection result of the peripheral vicinity camera 220 is included within this range.

[0042] If it is determined in S230 that the estimated self-position (N) is within the estimation error range of the (N+1)th sensor (S230: YES), the output self-position determiner 50 determines the self-position estimated based on the detection result of the sensor having the Nth estimation accuracy as the output self-position (S235). Therefore, for example, if the estimated self-position estimated based on the detection result of the peripheral proximity camera 220 is contained within a circular range with a radius of 20 cm centered on the estimated self-position estimated based on the detection result of the LiDAR 260, the estimated self-position estimated based on the detection result of the peripheral proximity camera 220 is determined as the output self-position. If the estimated self-position (N) is contained within the estimation error range of the sensor having the (N+1)th estimation accuracy, there is a high possibility that the estimated self-position (N) indicates the correct position. Therefore, in this case, in this embodiment, the estimated self-position (N) is determined as the output self-position.

[0043] In the above-mentioned S230, if it is determined that the estimation accuracy of the estimated self-position (N) is not within the estimation error range of the (N+1)th sensor (S230: NO), the output self-position determination unit 50 determines whether the determination in the above S230 has been completed for all estimated self-positions (self-positions estimated based on the detection results of all sensors 210 to 270) (S240).

[0044] If it is determined that the determination in S230 above has not been completed for all estimated self-positions (S240: NO), the output self-position determination unit 50 increments the variable N by 1 (S250). After S250, the above-mentioned S230 is executed again. Therefore, when S230 is executed for the second time, it is determined whether the self-position estimated based on the detection result by the LiDAR 260, which is the second sensor, whose estimation accuracy is within the estimation error range of the sonar 250, which is the third sensor, whose estimation accuracy is within the estimation error range. The estimation error range of the sonar 250 means a circular range with a radius of 30 cm, centered on the self-position estimated based on the detection result by the sonar 250.

[0045] In the above-mentioned S240, if it is determined that the determination of S230 has been completed for all estimated self-positions (S240: YES), the output self-position determiner 50 does not determine an output self-position (S245). After the above-mentioned S235 or S245 is completed, S125 is completed and the process returns to S50.

[0046] As is clear from the above description of S205 to S250, in the second embodiment, if the condition "the estimation accuracy of the estimated self-position (1) is within the estimation error range of the second sensor" is satisfied, the self-position estimated based on the detection result by the sensor (periphery nearby camera 220) having the first estimation accuracy is determined as the output self-position. This "the estimation accuracy of the estimated self-position (1) is within the estimation error range of the second sensor" corresponds to the predetermined determination condition in the present disclosure.

[0047] The self-location estimation device 100 of the second embodiment described above achieves the same effects as the self-location estimation device 100 of the first embodiment. In addition, when a predetermined determination condition is satisfied, the self-location estimated based on the detection result by the sensor (periphery vicinity camera 220) having the highest estimation accuracy among the multiple sensors 210 to 270 is determined as the output self-location, thereby further suppressing a decrease in the estimation accuracy of the self-location estimation. Furthermore, in the second embodiment, the "predetermined determination condition" includes a condition that the self-location estimated based on the detection result by the sensor (periphery vicinity camera 220) having the highest estimation accuracy is included within the estimation error range of the self-location estimated based on the detection result by the sensor (LiDAR 260) having the second highest estimation accuracy. Therefore, even if a low-accuracy self-position is estimated based on the detection result of surrounding vicinity camera 220, which is the sensor with the highest estimation accuracy, due to, for example, snowfall or the like making it impossible to capture a clear image of the road surface, if such self-position falls outside the estimation error range of the self-position estimated based on the detection result by the sensor (LiDAR 260) with the second highest estimation accuracy, it is possible to prevent the self-position estimated based on the detection result by surrounding vicinity camera 220 from being determined as the output self-position. Furthermore, if the determination condition is not met, the self-position estimated based on the detection result by the sensor with the second or later highest estimation accuracy is determined as the output self-position, thereby preventing the output self-position from being not determined.

[0048] C. Third embodiment: 5 differs from the self-location estimation device 100 of the first embodiment in that the CPU 110 also functions as the environmental information acquisition unit 60. Other configurations of the self-location estimation device 100a of the third embodiment are similar to those of the self-location estimation device 100 of the first embodiment, and therefore the same components are denoted by the same reference numerals and detailed description thereof will be omitted.

[0049] The environmental information acquisition unit 60 acquires information about the driving environment of the vehicle V1 (hereinafter referred to as "driving environment information"). The environmental information acquisition unit 60 transmits the acquired driving environment information to each estimation accuracy determination unit 21-26. In this embodiment, the "driving environment information" includes information about the weather at the location where the vehicle V1 is driving and information about the time period when the vehicle V1 is driving. The "information about the weather at the location where the vehicle V1 is driving" refers to weather information that may affect the detection of at least one of the sensors 210-270, such as whether it is raining or snowing. Rain or snow affects the detection results of the cameras (the surrounding vicinity camera 220, the front camera 230, and the surrounding camera 270). Therefore, the accuracy of the vehicle's location estimated based on the detection results of these sensors may be low. Information such as "raining" or "snowing" may be determined from the detection results of a rain sensor mounted on the vehicle V1 or the operation status of the windshield wipers, or may be acquired by communicating with a server device installed outside the vehicle V1. The above-mentioned "information related to the driving time period" refers to information related to a time period that may affect the detection of at least one of the sensors 210 to 270, such as nighttime or a time period in the evening when the afternoon sun is strong. For example, at night, the accuracy of object detection based on captured images decreases due to the low amount of light. Information such as "nighttime or a time period in the evening when the afternoon sun is strong" may be determined, for example, from the detection results of an illuminance sensor mounted on the vehicle V1 or time information provided by a clock or time server (not shown).

[0050] The output self-location determination process of the third embodiment shown in Fig. 6 differs from the output self-location determination process of the first embodiment in that S107 is additionally executed and S115a is executed instead of S115. Since the other steps in the output self-location determination process of the third embodiment are the same as those in the output self-location determination process of the first embodiment, the same steps are denoted by the same reference numerals and detailed descriptions thereof will be omitted.

[0051] After the above-mentioned S105 is completed, the environment information acquisition unit 60 acquires the traveling environment information and transmits it to each of the estimation accuracy specification units 21 to 26 (S107). After the completion of S107, the above-mentioned S110 is executed.

[0052] After completion of S110, in S115a, each of the estimation accuracy specifying units 21 to 26 specifies the estimation accuracy adjusted in accordance with the traveling environment information acquired by the environment information acquiring unit 60. In this embodiment, each of the estimation accuracy specifying units 21 to 26 adjusts the estimation accuracy by referring to an estimation accuracy table stored in advance in the ROM 120.

[0053] As shown in FIG. 7 , the estimation accuracy table includes a sensor type field, a normal time field, a rain and snow field, and a night field. The sensor type field indicates each sensor 210-270. The normal time field indicates the estimation accuracy under normal conditions, i.e., in this embodiment, conditions excluding rain and snow and night. The rain and snow field indicates the estimation accuracy under rain or snow. The night field indicates the estimation accuracy under night. The estimation accuracy of the vehicle's own location based on the detection results of the surrounding vicinity camera 220, the front camera 230, the LiDAR 260, and the surrounding camera 270 is lower during rain, snow, and night than during normal conditions. In contrast, the estimation accuracy of the vehicle's own location based on the detection results of the radar 240 and the sonar 250 is equal under normal conditions, during rain or snow, and at night. These estimation accuracies are determined in advance through experiments and simulations and set in the estimation accuracy table. As is clear from the estimation accuracy table, under normal circumstances, the sensor with the highest estimation accuracy is the periphery proximity camera 220. In contrast, during rain and snow, or at night, the sensor with the highest estimation accuracy is the sonar 250.

[0054] In S115a, for example, when the driving environment information indicates that it is snowing, first estimation accuracy specifying unit 21 specifies the estimation accuracy of the self-estimation based on the detection result of surrounding vicinity camera 220 as 1 m. Also, for example, when the driving environment information indicates that it is nighttime, fourth estimation accuracy specifying unit 24 specifies the estimation accuracy of the self-estimation based on the detection result of sonar 250 as 30 cm, the same as in normal times.

[0055] After completion of S115a described above, S120 and S125 are executed. Therefore, under normal circumstances, the output self-position is determined to be the self-position estimated based on the detection result by the surrounding vicinity camera 220. During rain, snow, and at night, the output self-position is determined to be the self-position estimated based on the detection result by the sonar 250.

[0056] The self-location estimation device 100a of the third embodiment described above has the same effects as the self-location estimation device 100 of the first embodiment. In addition, each estimation accuracy specifying unit 21-26 adjusts the estimation accuracy according to the driving environment indicated by the driving environment information, and the output self-location determiner 50 uses each adjusted estimation accuracy to determine an output self-location from among self-locations estimated based on the detection results of each sensor 210-270. Therefore, by using the estimation accuracy adjusted according to the driving environment, it is possible to determine an output self-location more accurately.

[0057] In addition, since the driving environment information includes at least one of information regarding the weather at the driving location of the vehicle V1 and information regarding the driving time period of the vehicle V1, the output self-position can be determined more accurately by utilizing the estimation accuracy adjusted according to the weather or driving time period.

[0058] D. Other Embodiments: (D1) In the second embodiment, the determination in S230 is executed for all estimated self-positions (self-positions estimated based on the detection results of all sensors 210 to 270) until it is determined that the estimated self-position (N) has an estimation accuracy within the estimation error range of the (N+1)th sensor, but the present disclosure is not limited to this. For example, if the self-position estimated based on the detection result of the sensor with the third highest estimation accuracy is not within the estimation error range of the fourth sensor, S245 may be executed so as not to determine an output self-position.

[0059] (D2) In each embodiment, in S110, each of the position estimation units 11 to 16 narrows the matching range in the map information using the estimated self-position by the GNSS 210 to perform matching, but the present disclosure is not limited to this. This narrowing of the matching range may be omitted. In this case, the GNSS 210 and the GNSS position estimation unit 40 may be omitted.

[0060] (D3) In the third embodiment, the driving environment information includes both information about the weather at the driving location of the vehicle V1 and information about the driving time of the vehicle V1, but one of these pieces of information may be omitted. Also, instead of or in addition to at least one of these pieces of information, any other type of information about the driving environment of the vehicle V1 may be used as the driving environment information.

[0061] (D4) The self-localization device 100, 100a and the method thereof described herein may be implemented by a special-purpose computer configured by configuring a processor and memory programmed to execute one or more functions embodied in a computer program. Alternatively, the self-localization device 100, 100a and the method thereof described herein may be implemented by a special-purpose computer configured by configuring a processor with one or more dedicated hardware logic circuits. Alternatively, the self-localization device 100, 100a and the method thereof described herein may be implemented by one or more special-purpose computers configured by combining a processor and memory programmed to execute one or more functions with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored in a computer-readable non-transitory tangible recording medium as instructions executed by a computer.

[0062] The present disclosure is not limited to the above-described embodiments and can be realized in various configurations without departing from the spirit thereof. For example, the technical features in each embodiment corresponding to the technical features in the embodiments described in the Summary of the Invention section can be appropriately replaced or combined to solve some or all of the above-described problems or achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be deleted as appropriate. The present disclosure may be realized, for example, in the following forms.

[0063] [Form 1] A self-position estimation device (100, 100a) for estimating the self-position of a vehicle (V1) equipped with a plurality of sensors (200) for detecting a surrounding environment, an estimation accuracy specifying unit (21 to 26) for specifying an estimation accuracy when estimating the self-position based on the detection results of each of the sensors; an output self-position determination unit (50) that determines an output self-position, which is the self-position to be output, from the self-positions estimated based on the detection results of the sensors by using the identified estimation accuracy; A self-location estimation device comprising: [Form 2] The self-location estimation device according to aspect 1, A self-position estimation device in which the output self-position determination unit determines, when predetermined determination conditions are met, the self-position estimated based on the detection results of the sensor having the highest estimation accuracy among the multiple sensors as the output self-position. [Form 3] In the self-location estimation device according to aspect 2, A self-position estimation device, wherein the determination condition includes a condition that the self-position estimated based on the detection result by the sensor having the highest estimation accuracy is within an estimation error range of the self-position estimated based on the detection result by the sensor having the second highest estimation accuracy. [Form 4] In the self-location estimation device according to aspect 3, A self-position estimation device in which, if the determination condition is not satisfied, the output self-position determination unit determines the self-position estimated based on the detection result by the sensor having the second or subsequent highest estimation accuracy as the output self-position. [Form 5] In the self-location estimation device according to any one of the first to fourth aspects, The vehicle further includes an environmental information acquisition unit (60) that acquires driving environment information, which is information about the driving environment of the vehicle; the estimation accuracy specifying unit adjusts the estimation accuracy in accordance with the traveling environment indicated by the traveling environment information; The output self-position determination unit determines the output self-position from among the self-positions estimated based on the detection results of each of the sensors, using each of the adjusted estimation accuracies. [Form 6] In the self-location estimation device according to aspect 5, The driving environment information includes at least one of information about the weather at the location where the vehicle is driving and information about the time period when the vehicle is driving. [Explanation of symbols]

[0064] 200... sensor (sensor group), V1... vehicle, 100, 100a... self-position estimation device, 21 to 26... estimation accuracy specification unit, 50... output self-position estimation device, 60... environmental information acquisition unit

Claims

1. A self-position estimation device (100, 100a) for estimating the self-position of a vehicle (V1) equipped with a plurality of sensors (200) for detecting a surrounding environment, an estimation accuracy specifying unit (21 to 26) for specifying an estimation accuracy when estimating the self-position based on the detection results of each of the sensors; an output self-position determination unit (50) that determines an output self-position, which is the self-position to be output, from the self-positions estimated based on the detection results of the sensors by using the specified estimation accuracy; A self-location estimation device comprising:

2. The self-location estimation device according to claim 1 , A self-position estimation device in which the output self-position determination unit determines, when predetermined determination conditions are met, the self-position estimated based on the detection results of the sensor having the highest estimation accuracy among the multiple sensors as the output self-position.

3. The self-position estimation device according to claim 2, A self-position estimation device, wherein the determination condition includes a condition that the self-position estimated based on the detection results by the sensor with the highest estimation accuracy is within the estimation error range of the self-position estimated based on the detection results by the sensor with the second highest estimation accuracy.

4. The self-position estimation device according to claim 3, A self-position estimation device in which, if the determination condition is not satisfied, the output self-position determination unit determines the self-position estimated based on the detection result by the sensor having the second or subsequent highest estimation accuracy as the output self-position.

5. 5. The self-location estimation device according to claim 1, The vehicle further includes an environment information acquisition unit (60) that acquires driving environment information, which is information about the driving environment of the vehicle; the estimation accuracy specifying unit adjusts the estimation accuracy in accordance with the traveling environment indicated by the traveling environment information; The output self-position determination unit determines the output self-position from among the self-positions estimated based on the detection results of each of the sensors, using each of the adjusted estimation accuracies.

6. The self-position estimation device according to claim 5, The driving environment information includes at least one of information about the weather at the location where the vehicle is driving and information about the time period when the vehicle is driving.

Images