Vehicle data collection device and vehicle data collection method

Abstract

To provide a vehicle data collection device capable of collecting data from start to end of a scenario with neither excess nor shortage when the scenario is detected.SOLUTION: A data collection device comprises: a data storage section which stores time-series data indicating behaviors of one or more other vehicles which are acquired by an environment detection device loaded on a vehicle; a scenario detection section which detects a start of a scenario when an end of the scenario defined as a data collection target is detected; and a collection target data determination section which determines time-series data of a collection target from the time-series data stored in the data storage section based on timings of the start and the end of the scenario.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a vehicle data collection device and a vehicle data collection method for collecting data on a vehicle and other vehicles traveling around the vehicle. [Background technology]

[0002] Due to the rapid social changes and technological advances in recent years, there is growing momentum for using vehicle data to solve social issues, and vehicle manufacturers are considering vehicle data collection systems. The purposes of collecting vehicle data are varied, such as updating autonomous driving AI models and maps.

[0003] Let us take the example of updating an autonomous driving AI model. The data required to update an autonomous driving AI model is data for situations that are generally considered dangerous, such as when the vehicle is close to another vehicle.

[0004] In other words, data on situations that could lead to car accidents is needed, especially data on accidents caused by collisions between cars, which are said to account for more than 85% of car accidents.

[0005] Hereinafter, events that are the subject of data collection and are defined by the behavior of the subject vehicle or events that are defined by the behavior of the subject vehicle and the behavior of other vehicles are defined as "scenarios." Scenarios particularly include traffic scenarios that are defined by the behavior of the subject vehicle (such as turning at an intersection, changing lanes, or a situation where an obstacle such as a stone has fallen in front of the subject vehicle) or traffic scenarios that are defined by the behavior of the subject vehicle and the behavior of other vehicles (such as cutting in, cutting out, overtaking, a following vehicle suddenly approaching, or a preceding vehicle suddenly braking).

[0006] A technique for collecting data related to this scenario is disclosed in Patent Document 1. In Patent Document 1, when the end point of a specified scenario is detected, data of a fixed length going back a fixed time from the end point is collected. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Publication No. 2020-140657 Summary of the Invention [Problem to be solved by the invention]

[0008] However, the technology in Patent Document 1 collects and stores data only for a fixed period of time going back from the end point of the scenario, so if the length of the scenario exceeds the fixed period of time, data from the start of the scenario cannot be collected, and if the length of the scenario is less than the fixed period of time, unnecessary data from before the start of the scenario is collected.

[0009] In view of the above-mentioned problems, an object of the present invention is to provide a vehicle data collection device that, when a scenario is detected, can collect just the right amount of data from the start to the end of the scenario. [Means for solving the problem]

[0010] In order to solve the above problems, the vehicle data collection device of the present invention includes a data storage unit that stores time series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle, a scenario detection unit that detects the start of a scenario when it detects the end of a scenario defined as a data collection target, and a collection target data determination unit that determines the time series data to be collected from the time series data stored in the data storage unit based on the timing of the start and end of the scenario. [Effects of the Invention]

[0011] According to the invention, when the end of a scenario is detected, the start of the scenario is detected retroactively and only data from the start to the end of the scenario is collected, making it possible to collect just the right amount of data from the start to the end of the scenario. Further features related to the present invention will become apparent from the description of the present specification and the accompanying drawings. In addition, the problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a diagram illustrating an example of a hardware configuration including a data collection device according to a first embodiment of the present invention. [Figure 2] FIG. 1 is a block diagram showing the functional configuration of a data collection device according to a first embodiment of the present invention. [Figure 3] FIG. 10 is a diagram showing another vehicle cutting in on the own vehicle; [Figure 4] FIG. 10 is a diagram for explaining a situation in which a scenario occurs between a plurality of other vehicles. [Figure 5] FIG. 3 is a diagram showing an example of the data structure of data stored in a candidate data storage unit 106. [Figure 6] 4 is a flowchart showing a process executed by the data collecting device according to the first embodiment of the present invention. [Figure 7] FIG. 2 is a diagram for explaining the effect of the first embodiment of the present invention. [Figure 8] FIG. 10 is a block diagram showing the functional configuration of a data collection device according to a second embodiment of the present invention. [Figure 9] FIG. 10 is a diagram for explaining processing in the second embodiment of the present invention. [Figure 10] 10 is a flowchart showing a process executed by a data collecting device according to a second embodiment of the present invention. [Figure 11] FIG. 10 is a diagram for explaining determination of the end point of a lane change. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that hereinafter, the vehicle data collection device will be simply referred to as a data collection device.

[0014] [Example 1] 1 is a block diagram showing the hardware configuration of a vehicle including a data collection device according to Example 1. The vehicle 1 includes at least a data collection device 100, a camera 15, and a sensor 16. The data collection device 100 includes a CPU (Central Processing Unit) 11 that calculates parameters required for vehicle control based on data, a ROM (Read Only Memory) 12 that is a storage device from which data can be read, a RAM (Random Access Memory) 13 that can both write and read data, and a communication module 14 that converts the calculation results into a communication protocol and transmits the converted results.

[0015] The data collection device 100 is also connected to an external server 18 via a communication path 17. The communication path 17 may physically include multiple communication buses, and the standards of each communication bus may be the same or different. These communication bus standards include CAN (registered trademark), LIN (registered trademark), FlexRay (registered trademark), Ethernet (registered trademark), etc.

[0016] The data collection device 100 receives the data that has passed through the communication path 17. Then, in response to reception, the received data is expanded in the RAM 13. Then, the CPU 11 performs calculations based on the data written in the RAM 13 and the data read from the ROM 12. A program is pre-installed in the data collection device 100, and the CPU 11 executes the program to perform the processing described below. Note that the data collection device 100 may be configured to be self-contained within the vehicle, without being connected to the external server 18.

[0017] The vehicle 1 is also equipped with a camera 15 and a sensor 16. A plurality of cameras 15 and sensors 16 are provided, and they acquire information about the host vehicle, other vehicles traveling around the host vehicle, and the surrounding environment of the host vehicle (roads, signs, etc.). The camera 15 may be, for example, a stereo camera, and the sensor 16 may be, for example, a millimeter-wave radar or a LiDAR (Light Detection and Ranging). Note that, hereinafter, these cameras 15 and sensors 16 may be collectively referred to as an external environment detection device.

[0018] 2 is a functional block diagram showing the functional configuration of the data collection device 100. As shown in FIG. 2, the data collection device 100 according to this embodiment includes a termination algorithm group 101, a start algorithm selection unit 102, a start detection data selection unit 103, a start algorithm group 104, a candidate data extraction unit 105, a candidate data storage unit 106, and a data storage unit 107.

[0019] The termination algorithm group 101 detects the end point of a scenario when a scenario occurs between the vehicle itself or between vehicles. Specific methods for detecting a scenario will be described later. The termination algorithm group 101 further includes multiple end detection algorithms. An end detection algorithm is set for each scenario, and when a certain end detection algorithm (in this embodiment, an end detection algorithm for detecting cut-in of another vehicle) detects the end of a scenario in which another vehicle cuts in, it transmits detected scenario information (the type of scenario detected) and a detected vehicle ID (an ID for identifying the other vehicle that caused the scenario) to the start algorithm selection unit 102 and the start detection data selection unit 103, respectively. Note that this termination algorithm group 101 is always running, and is always able to detect the end of a scenario.

[0020] The start algorithm selection unit 102 selects the corresponding scenario ID based on the received detected scenario information. That is, it transmits to the start algorithm group 104 the type of scenario detected by the end algorithm group 101 (in this embodiment, cut-in by another vehicle).

[0021] The start detection data selection unit 103 acquires measurement data of other vehicles from the data storage unit 107, which temporarily stores data acquired by the external environment detection device, and transmits the data to the start algorithm group 104. Note that the other vehicles are other vehicles identified by the detected vehicle ID received from the end detection algorithm, and the measurement data has a length that goes back a sufficient time from the end of the scenario. The start detection data selection unit 103 may also acquire driving data of the host vehicle and data related to the host vehicle's surrounding environment from the data storage unit 107 and transmit the data to the start algorithm group 104.

[0022] The start algorithm group 104 detects the start time of a scenario occurring between the host vehicle or between the host vehicle and another vehicle. Like the termination algorithm group 101, the start algorithm group 104 includes multiple start detection algorithms set for each scenario. When the start algorithm group 104 receives a scenario ID from the start algorithm selection unit 102, it selects a start detection algorithm according to the scenario ID. In this embodiment, a start detection algorithm for detecting the start time of the other vehicle's cut-in is selected. Then, using the selected algorithm, data received from the start detection data selection unit 103 is analyzed to detect the time when the other vehicle starts to cut-in. Specifically, the time when the other vehicle traveling in a traveling lane adjacent to the host vehicle's traveling lane comes alongside the host vehicle. Information on the detected start time is transmitted to the candidate data extraction unit 105. Note that the start algorithm selection unit 102, the start detection data selection unit 103, and the start algorithm group 104 may be activated in response to receiving information about a scenario from the termination algorithm group 101 to reduce load. The start algorithm group 104 corresponds to the scenario detection unit in the claims.

[0023] To summarize one function of the initiation algorithm group 104 in this embodiment, the scenario includes a traffic scenario defined by the behavior of other vehicles, and when the end of the traffic scenario is detected, the initiation algorithm group (scenario detection unit) 104 detects the start of the scenario by determining the behavior of other vehicles corresponding to the type of traffic scenario. Also, the scenario may include a traffic scenario defined by the behavior of the vehicle itself, and when the end of the traffic scenario is detected, the initiation algorithm group (scenario detection unit) 104 may detect the start of the scenario by determining the behavior of the vehicle itself corresponding to the type of traffic scenario.

[0024] The candidate data extraction unit 105 receives information about the start time from the start algorithm group 104 and also receives the end timing from the end algorithm group 101, and therefore extracts the data from the data storage unit 107 from the start time to the end time of the relevant scenario (cut-in of other vehicles) without excess or deficiency, and transmits it to the candidate data storage unit 106. The candidate data extraction unit 105 corresponds to the collection target data determination unit in the claims.

[0025] The candidate data storage unit 106 stores the travel data of other vehicles received from the candidate data extraction unit 105 as data to be collected.

[0026] The data storage unit 107 has a function of constantly receiving other vehicle data from the external environment detection device and temporarily storing the data. In many cases, multiple other vehicles (e.g., other vehicle 20a and other vehicle 20b) are traveling around the traveling host vehicle, and the data of each other vehicle is stored independently and transmitted to the termination algorithm group 101, etc. The data storage unit 107 may temporarily store not only the traveling data of the other vehicles, but also the traveling data of the host vehicle and data related to the surrounding environment of the host vehicle. Furthermore, the data storage unit 107 may temporarily store the traveling data of the host vehicle and data related to the surrounding environment of the host vehicle instead of the traveling data of the other vehicles.

[0027] As described above, the data collection device of this embodiment includes a data storage unit 107 that stores time-series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle, a start algorithm group (scenario detection unit) 104 that detects the start of a scenario when the end of a scenario defined as a data collection target is detected, and a candidate data extraction unit (collection target data determination unit) 105 that determines time-series data to be collected from the time-series data stored in the data storage unit 107 based on the start and end timing of the scenario. The data collection device of this embodiment also includes a data storage unit 107 that stores time-series data indicating the behavior of the host vehicle acquired by the external environment detection device mounted on the vehicle, a start algorithm group (scenario detection unit) 104 that detects the start of a scenario when the end of a scenario defined as a data collection target is detected, and a candidate data extraction unit (collection target data determination unit) 105 that determines time-series data to be collected from the time-series data stored in the data storage unit 107 based on the start and end timing of the scenario.

[0028] Next, an example of the scenario described above will be explained. Figure 3 is a diagram showing how another vehicle cuts in on the host vehicle. As shown in Figure 3, it is assumed that the host vehicle 19 is traveling in the left driving lane, and another vehicle 20 traveling in the right driving lane accelerates to overtake the host vehicle 19 and cuts in on the left driving lane.

[0029] At this time, it is assumed that the end of the cut-in is detected at time t1. At this point, the start of the cut-in has not been detected. This is because, in a situation where the other vehicle 20 is traveling in the right-hand driving lane, it is not possible to determine whether the other vehicle 20 will cut in or continue traveling in the right-hand driving lane.

[0030] Therefore, as described above, when the end algorithm group 101 detects the end of the cut-in, the start algorithm selection unit 102, the start detection data selection unit 103, and the start algorithm group 104 are activated, and the driving data of the other vehicle 20 is analyzed from the data storage unit 107 to detect the start time t2 of the cut-in (the time when the other vehicle 20 comes alongside the host vehicle 19).

[0031] Then, the candidate data extraction unit 105 extracts point cloud information acquired by the sensor and video information acquired by the camera of other vehicles 20 between times t2 and t1 from the data storage unit 107 and stores them in the candidate data storage unit 106.

[0032] In this embodiment, by performing such processing, it is possible to obtain data from the start to the end of a cut-in scenario without excess or deficiency.

[0033] FIG. 4 is a diagram for explaining a situation in which different scenarios occur between a plurality of other vehicles. As shown in Figure 4, assume a situation in which the host vehicle 19 is traveling in the left driving lane, a first other vehicle 20a is also traveling in the left driving lane, and a second other vehicle 20b is traveling in the adjacent right driving lane.

[0034] Here, suppose that the preceding vehicle 20a first executes a cutout to change lanes into the right-hand driving lane, and the end of the cutout is detected by the end algorithm group 101 at time t3. Then, as in the case described with reference to Fig. 3, the start algorithm selection unit 102, the start detection data selection unit 103, and the start algorithm group 104 are activated, analyze the driving data of the other vehicle 20 from the data storage unit 107, and detect the start time t4 of the cutout (the time when the other vehicle 20a starts to change lanes).

[0035] At the same time, another vehicle 20b traveling parallel in the right lane starts cutting in, and the end of the cutting in is detected by the end algorithm group 101 at time t5. Then, as in the case described in Fig. 3, the start algorithm selection unit 102, the start detection data selection unit 103, and the start algorithm group 104 are activated, analyze the traveling data of the other vehicle 20 from the data storage unit 107, and detect the start time t6 of the cutting in (the time when the other vehicle 20b comes alongside the host vehicle 19).

[0036] As described above, when scenarios occur between multiple other vehicles at approximately the same time, it is difficult to process them simultaneously because data transfer becomes complicated. Therefore, in this embodiment, when multiple scenarios occur simultaneously, a priority order is set. Specifically, a scenario with a higher risk is given priority. A high-risk scenario refers to a scenario in which another vehicle approaches the host vehicle. In the case of FIG. 4, the other vehicle 20a that cuts out moves away from the host vehicle 19, while the other vehicle 20b that cuts in moves closer to the host vehicle 19. Therefore, it is determined that the risk is high, and the traveling data of the other vehicle 20b is analyzed with priority.

[0037] In other words, when there are multiple scenarios defined as data collection targets, the candidate data extraction unit (collection target data determination unit) 105 can determine the scenario with the highest risk as the collection target with priority. In other words, of the multiple scenarios defined as data collection targets, the collection target data determination unit 105 can determine the scenario including the behavior of another vehicle approaching the host vehicle with priority over a scenario including the behavior of another vehicle moving away from the host vehicle as the collection target.

[0038] Other high-risk scenarios include sudden braking by a leading vehicle and a sudden approach by a following vehicle. It is preferable to determine priorities for these multiple scenarios in advance, and if multiple scenarios occur simultaneously, to deal with the scenarios with the highest priority first.

[0039] Here, a specific example of a method for detecting the end of a scenario will be described with reference to FIG.

[0040] FIG. 11 is a diagram showing how the host vehicle 19 changes lanes, moving from the left lane to the right lane. That is, the scenario here is a lane change by the host vehicle (as mentioned above, in this case, it is assumed that there is a fallen object or an accident ahead in the left lane in which the host vehicle 19 was originally traveling). Distances 23-27 between the center of the host vehicle 19 and the lane center lines (center line 21 of the left lane and center line 22 of the right lane) are defined as offsets. The offsets take positive values ​​when the host vehicle 19 is located to the left of the lane center line, and take negative values ​​when the host vehicle 19 is located to the right of the lane center line.

[0041] As shown in Figure 11(a), when the vehicle 19 gradually moves from the left lane to the right lane, the offset changes discontinuously from a negative value to a positive value at a certain moment (offset 25 to 26, the moment of 30 seconds in Figure 11(b)). This moment is determined to be the end of the lane change, i.e., the end of the scenario.

[0042] In other words, in this example, the data storage unit 107 also stores time series data indicating the behavior of the vehicle 19 acquired by the external environment detection device, the scenario is a lane change, which is a traffic scenario defined by the behavior of the vehicle 19, and the initiation algorithm group 104 detects the start of a lane change based on the surrounding environment of the vehicle 19 when the end of the traffic scenario (lane change) is detected.

[0043] If the subject vehicle 19 is assumed to be the other vehicle 20, this can also be used as a method for detecting the cut-in and cut-out end points.

[0044] Fig. 5 is a diagram showing an example of the data structure of data stored in the candidate data storage unit 106. As shown in Fig. 5, the data stored in the candidate data storage unit 106 includes a number for identifying the data, a scenario ID, a detected vehicle ID, a start time of the scenario, and an end time of the scenario. Note that the data related to the traffic scenario defined by the behavior of the host vehicle 19 does not necessarily include the detected vehicle ID.

[0045] However, the data stored by the candidate data storage unit 106 is not limited to this, and may also store, for example, the vehicle's current driving speed, weather, road surface conditions, and the type of external environment detection device (camera, sensor) from which the data is collected.

[0046] FIG. 6 is a flowchart showing the processing executed by the data collecting device in this embodiment. First, in step S101, the termination algorithm group 101 detects the end of the scenario. Next, in step S102, the start algorithm selection unit 102 selects a start algorithm based on the detection scenario information received from the termination algorithm group 101. Next, in step S103, the start detection data selection unit 103 selects start detection data from the data storage unit 107 to collect information about the host vehicle or other vehicles that caused the scenario. Note that the order of steps S102 and S103 may be reversed, or they may be performed simultaneously.

[0047] Next, in step S104, the start algorithm group 104 uses a start detection algorithm selected according to the scenario ID received from the start algorithm selection unit 102 to detect the start point of the scenario in the driving data of the host vehicle or other vehicles received from the start detection data selection unit 103. Finally, the candidate data extraction unit 105 extracts driving data of the host vehicle or other vehicles from the start to the end of the scenario from the data storage unit 107 based on the end timing received from the end algorithm group 101 and the start timing received from the start algorithm group 104, and transmits this to the candidate data storage unit 106, which stores it.

[0048] The effects of the present embodiment described above will be explained with reference to FIG. In this example, the data acquisition time for each scenario differs, as shown in Figure 7. Specifically, the time from the start to the end of data acquisition is long for cut-in / cut-out, and becomes shorter for turning at an intersection, changing lanes when merging / branching, and overtaking.

[0049] Furthermore, for example, the time it takes for a cut-in / cut-out varies depending on speed, lane width, vehicle width, number of lanes, etc. Similarly, the time from the start to the end of a scenario varies depending on the speed and presence or absence of oncoming vehicles when turning at an intersection, the speed and merging distance when changing lanes when merging / branching, and the speed and presence or absence of third parties when overtaking.

[0050] According to this embodiment, it is possible to accurately grasp the time from the start to the end of a scenario, which varies depending on the type of scenario and other factors even for the same type of scenario, and to collect just the right amount of data. In other words, the candidate data extraction unit (collection target data determination unit) 105 changes the length of the time-series data to be collected based on the type of traffic scenario. On the other hand, in conventional systems, data is collected for a fixed period of time prior to the end of the scenario, regardless of the type of scenario, which means that there is an excess or deficiency of data for each scenario.

[0051] [Example 2] Next, a data collection device according to a second embodiment of the present invention will be described. In the following, a description of parts that overlap with the first embodiment will be omitted.

[0052] The data collecting device 100 according to the embodiment further includes a backup start detection algorithm 108 in addition to the configuration described in the first embodiment. The backup start detection algorithm 108 performs the function of setting a backup start timing for data temporarily stored in the data storage unit 107 and transmitting a deletion signal to delete part of the data. The backup start detection algorithm 108 corresponds to the backup start timing setting unit in the claims.

[0053] The processing in this embodiment will be described with reference to FIG. Figure 9 shows a case where another vehicle 20 cuts into the driving lane in which the host vehicle 19 is driving, and Figure 9(a) shows a situation where the preliminary start detection algorithm 108 of this embodiment is not used, and Figure 9(b) shows a situation where the preliminary start detection algorithm 108 of this embodiment is used.

[0054] 9(a), the data storage unit 107 needs to store all driving data before time t7 until the end algorithm group 101 of the vehicle 19 detects the end of the cut-in at time t7. Therefore, if a scenario does not occur for a long time and it takes time to detect the end of the scenario, the required capacity will increase, which will increase costs.

[0055] On the other hand, when a backup start detection algorithm 108 is provided as in this embodiment, the backup start detection algorithm 108 first sets a backup start timing at an arbitrary time t10. If the end of any scenario is not detected for a predetermined period of time thereafter, the backup start timing is reset at time t9, and the data storage unit 107 is instructed to delete the data collected between times t10 and t9.

[0056] In other words, the backup start detection algorithm (backup start timing setting unit) 108 in this embodiment detects the occurrence of a specified event based on the time series data stored in the data storage unit, and if the end of the scenario is not detected for a specified time after the occurrence of the specified event is detected, deletes the time series data stored in the data storage unit from the time the specified event occurs until the specified time has elapsed.

[0057] By doing so, when the end of cut-in is detected at time t8 thereafter, data only from time t9 to t8 is temporarily stored in the data storage unit 107, thereby reducing the capacity required for the data storage unit 107. Note that an example of the timing for setting the backup start timing is the timing when another vehicle 20 comes alongside the host vehicle 19.

[0058] 10 is a flowchart showing the processing executed by the data collecting device in this embodiment. The flowchart in this embodiment differs from the flowchart in the first embodiment described in FIG. 6 in that steps S106 to S108 are provided instead of step S101.

[0059] In this embodiment, first, the preparatory start detection algorithm 108 sets the preparatory start timing at an arbitrary timing (for example, when another vehicle 20 comes alongside the host vehicle 19, or when the following vehicle 20 comes within a predetermined distance, etc.).

[0060] Then, in step S107, the preliminary start detection algorithm 108 determines whether the end of the scenario is detected within a predetermined time. If the end of the scenario is detected within the predetermined time, the process is the same as step S101 in Figure 6, and steps S102 and onwards are executed.

[0061] If the end of the scenario is not detected within the predetermined time, the process proceeds to step S108, where the backup start timing is reset as described with reference to FIG. 9, and the data storage unit 107 is instructed to delete some of the data.

[0062] As described above, according to this embodiment, in addition to the effects of the first embodiment, the amount of data temporarily stored in the data storage unit 107 can be reduced, and therefore cost reduction can be expected.

[0063] Although the above mainly describes the cut-in of other vehicles as a scenario, scenarios also include those caused by the behavior of the vehicle itself (cut-in / cut-out of the vehicle itself, turning at an intersection, etc.). Since such scenarios caused by the vehicle itself may also be worth collecting as data, it is possible to store time-series data indicating the behavior of the vehicle itself acquired by the external environment detection device in the data storage unit 107, and collect data in the same way as in the case of scenarios caused by other vehicles.

[0064] In this case, however, it is preferable to set a scenario that considers the proximity of the vehicle to the surrounding environment (obstacles, accidents, etc.) in addition to the relationship between the vehicle and other vehicles.

[0065] The above-described embodiment of the present invention provides the following advantageous effects.

[0066] (1) The vehicle data collection device of the present invention includes a data storage unit that stores time series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle, a scenario detection unit that detects the start of a scenario when the end of a scenario defined as a data collection target is detected, and a collection target data determination unit that determines the time series data to be collected from the time series data stored in the data storage unit based on the timing of the start and end of the scenario.

[0067] With the above configuration, when the end of a scenario is detected, the start of the scenario is detected retroactively and only data from the start to the end of the scenario is collected, making it possible to collect just the right amount of data from the start to the end of the scenario.

[0068] (2) The scenario includes a traffic scenario defined by the behavior of other vehicles, and when the end of the traffic scenario is detected, the scenario detection unit detects the start of the scenario by determining the behavior of other vehicles corresponding to the type of traffic scenario. A specific example of the scenario in (1) is, for example, one defined by the behavior of other vehicles.

[0069] (3) The data storage unit also stores time-series data indicating the behavior of the vehicle acquired by the external environment detection device, the scenario includes at least a traffic scenario defined by the behavior of the vehicle, and the scenario detection unit detects the start of a scenario based on the surrounding environment of the vehicle when the end of the traffic scenario is detected. Another specific example of a scenario in (1) is one based on such behavior of the vehicle.

[0070] (4) The system further includes a backup start timing setting unit that detects the occurrence of a predetermined event based on the time-series data stored in the data storage unit, and deletes the time-series data stored in the data storage unit from the occurrence of the predetermined event until the passage of the predetermined time if the end of the scenario is not detected within a predetermined time after the occurrence of the predetermined event. This makes it possible to reduce the amount of data temporarily stored in the data storage unit and suppress cost increases.

[0071] (5) The collection target data determination unit changes the length of the time-series data to be collected based on the type of scenario, making it possible to collect just the right amount of data for each scenario.

[0072] (6) When there are multiple scenarios defined as data collection targets, the collection target data determination unit determines the scenario with the highest risk as the data collection target in priority order.

[0073] (7) Specifically, the collection target data determination unit determines, among multiple scenarios defined as data collection targets, scenarios including behavior of another vehicle approaching the subject vehicle as collection targets in preference to scenarios including behavior of another vehicle moving away from the subject vehicle.

[0074] The more dangerous the scenario, the more important the data is, so this type of operation is preferable.

[0075] (8) The vehicle data collection device according to the present invention includes a data storage unit that stores time series data indicating the behavior of the vehicle acquired by an external environment detection device mounted on the vehicle, a scenario detection unit that detects the start of a scenario when the end of a scenario defined as a data collection target is detected, and a collection target data determination unit that determines the time series data to be collected from the time series data stored in the data storage unit based on the timing of the start and end of the scenario.

[0076] With the above configuration, when the end of a scenario defined by the behavior of the vehicle is detected, the start of the scenario is detected retroactively and only data from the start to the end of the scenario is collected, making it possible to collect just the right amount of data from the start to the end of the scenario.

[0077] (9) The scenario includes at least a traffic scenario defined by the behavior of the vehicle, and the scenario detection unit detects the start of a scenario based on the surrounding environment of the vehicle when the end of the traffic scenario is detected. Another specific example of the scenario in (8) is one defined by the behavior of the vehicle.

[0078] (10) The vehicle data collection method of the present invention is a vehicle data collection method executed by a computer, and includes a data storage step of storing time series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle in a data storage unit, a scenario detection step of detecting the start of a scenario when the end of a scenario defined as a data collection target is detected, and a collection target data determination step of determining time series data to be collected from the time series data stored in the data storage unit based on the timing of the start and end of the scenario.

[0079] With the above configuration, when the end of a scenario is detected, the start of the scenario is detected retroactively and only data from the start to the end of the scenario is collected, making it possible to collect just the right amount of data from the start to the end of the scenario.

[0080] (11) The vehicle data collection method of the present invention is a computer-executed vehicle data collection method, and includes a data storage step of storing time series data indicating the behavior of the vehicle, acquired by an external environment detection device mounted on the vehicle, in a data storage unit; a scenario detection step of detecting the start of a scenario when the end of a scenario defined as a data collection target is detected; and a collection target data determination step of determining time series data to be collected from the time series data stored in the data storage unit based on the timing of the start and end of the scenario.

[0081] With the above configuration, when the end of a scenario defined by the behavior of the vehicle is detected, the start of the scenario is detected retroactively and only data from the start to the end of the scenario is collected, making it possible to collect just the right amount of data from the start to the end of the scenario.

[0082] The present invention is not limited to the above-described embodiments, and various modifications are possible. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to embodiments including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment. It is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to delete part of the configuration of each embodiment, or to add or replace other configurations. [Explanation of symbols]

[0083] 19: Vehicle itself; 20: Other vehicles; 101: End algorithm group; 104: Start algorithm group (scenario detection unit); 105: Candidate data extraction unit (collection target data determination unit); 107: Data storage unit; 108: Backup start detection algorithm (backup start timing setting unit)

Claims

1. a data storage unit that stores time-series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle; a scenario detection unit that detects the start of a scenario when detecting the end of the scenario defined as a data collection target; a collection target data determination unit that determines time series data to be collected from the time series data stored in the data storage unit based on the start and end timings of the scenario, A vehicle data collection device characterized by:

2. 2. The vehicle data collection device according to claim 1, the scenario includes a traffic scenario defined by the behavior of the other vehicle; when the end of the traffic scenario is detected, the scenario detection unit detects the start of the scenario by determining the behavior of the other vehicle corresponding to the type of the traffic scenario. A vehicle data collection device characterized by:

3. 2. The vehicle data collection device according to claim 1, The data storage unit also stores time-series data indicating the behavior of the host vehicle acquired by the external environment detection device, the scenario includes a traffic scenario defined by at least a behavior of the host vehicle; the scenario detection unit detects the start of the scenario based on a surrounding environment of the vehicle when the end of the traffic scenario is detected; A vehicle data collection device characterized by:

4. 2. The vehicle data collection device according to claim 1, a backup start timing setting unit that detects the occurrence of a predetermined event based on the time series data stored in the data storage unit, and, if the end of the scenario is not detected for a predetermined time after the occurrence of the predetermined event is detected, deletes the time series data from the occurrence of the predetermined event until the elapse of the predetermined time from the time series data stored in the data storage unit; A vehicle data collection device characterized by:

5. 2. The vehicle data collection device according to claim 1, the collection target data determination unit changes the length of the time series data to be collected based on the type of the scenario. A vehicle data collection device characterized by:

6. 2. The vehicle data collection device according to claim 1, When there are multiple scenarios defined as data collection targets, the collection target data determination unit determines the scenario with the highest risk as the data collection target with priority. A vehicle data collection device characterized by:

7. 7. A vehicle data collection device according to claim 6, the collection target data determination unit determines, among the plurality of scenarios defined as data collection targets, a scenario including a behavior of another vehicle approaching the host vehicle as a collection target in preference to a scenario including a behavior of another vehicle moving away from the host vehicle. A vehicle data collection device characterized by:

8. a data storage unit that stores time-series data indicating the behavior of the vehicle, the time-series data being acquired by an external environment detection device mounted on the vehicle; a scenario detection unit that detects the start of a scenario when detecting the end of the scenario defined as a data collection target; a collection target data determination unit that determines time series data to be collected from the time series data stored in the data storage unit based on the start and end timings of the scenario, A vehicle data collection device characterized by:

9. 9. A vehicle data collection device according to claim 8, the scenario includes a traffic scenario determined by a behavior of the host vehicle; the scenario detection unit detects the start of the scenario based on a surrounding environment of the vehicle when the end of the traffic scenario is detected; A vehicle data collection device characterized by:

10. 1. A computer-implemented vehicle data collection method, comprising: a data storage step of storing time series data indicating the behavior of one or more other vehicles acquired by an external environment detection device mounted on the vehicle in a data storage unit; a scenario detection step of detecting the start of a scenario defined as a data collection target when the end of the scenario is detected; a collection target data determination step of determining time series data to be collected from the time series data stored in the data storage unit based on timings of the start and end of the scenario, A vehicle data collection method comprising:

11. 1. A computer-implemented vehicle data collection method, comprising: a data storage step of storing time series data indicating the behavior of the host vehicle, the time series data being acquired by an external environment detection device mounted on the vehicle, in a data storage unit; a scenario detection step of detecting the start of a scenario defined as a data collection target when the end of the scenario is detected; a collection target data determination step of determining time series data to be collected from the time series data stored in the data storage unit based on timings of the start and end of the scenario, A vehicle data collection method comprising:

Images