JP7220050B2 - Data collection device, data collection system and data collection method - Google Patents

Description

The disclosed embodiments relate to data collection devices, data collection systems and data collection methods.

2. Description of the Related Art Conventionally, a data collection device is known that collects road information from an in-vehicle device mounted on each vehicle. Such a data collection device collects road information for a desired position by selecting vehicles for which road information is to be collected based on the position information of each vehicle (see, for example, Patent Document 1).

JP 2018-055581 A

However, the conventional technology described above merely collects road information and cannot evaluate the suitability of parameters relating to vehicle control.

Specifically, in recent years, with the remarkable development of automated driving technology, at the development site of the vehicle control model that controls the automated driving of the vehicle, information on the behavior of the vehicle is collected from the vehicle actually traveling on the road, and the collected information It is required to evaluate the suitability of parameters related to vehicle control based on.

One aspect of the embodiments has been made in view of the above, and aims to provide a data collection device, a data collection system, and a data collection method that can evaluate the suitability of parameters related to vehicle control. .

A data collection device according to an aspect of an embodiment includes a control unit that collects vehicle data about a vehicle from an in-vehicle device mounted in each vehicle. Further, the control unit determines a vehicle exhibiting a desired behavior based on the collected vehicle data, and based on the determined determination result, the vehicle data for evaluation, which is used for evaluating parameters related to vehicle control. Extract data collection conditions, analyze the collected evaluation data based on the collection conditions , update the vehicle control model by machine learning based on the analyzed analysis results, and update the updated vehicle control model. of vehicles.

According to one aspect of the embodiments, the suitability of parameters for vehicle control can be evaluated.

FIG. 1A is a schematic explanatory diagram (part 1) of a data collection method according to an embodiment; FIG. 1B is a schematic explanatory diagram (part 2) of the data collection method according to the embodiment; FIG. 1C is a schematic explanatory diagram (part 3) of the data collection method according to the embodiment; FIG. 1D is a schematic explanatory diagram (part 4) of the data collection method according to the embodiment; FIG. 1E is a schematic explanatory diagram (No. 5) of the data collection method according to the embodiment; FIG. 1F is a schematic explanatory diagram (No. 6) of the data collection method according to the embodiment; FIG. 1G is a schematic explanatory diagram (No. 7) of the data collection method according to the embodiment; FIG. 1H is a schematic explanatory diagram (No. 8) of the data collection method according to the embodiment; FIG. 2 is a block diagram showing a configuration example of the data collection system according to the embodiment. FIG. 3A is a diagram (part 1) illustrating an example of extraction processing. FIG. 3B is a diagram (part 2) illustrating an example of the extraction process; FIG. 4 is a diagram illustrating a processing sequence executed by the data collection system according to the embodiment;

Hereinafter, embodiments of a data collection device, a data collection system, and a data collection method disclosed in the present application will be described in detail with reference to the accompanying drawings. In addition, this invention is not limited by embodiment shown below.

First, an overview of the data collection method according to the embodiment will be described with reference to FIGS. 1A to 1H. 1A to 1H are schematic explanatory diagrams (part 1) to (part 8) of the data collection method according to the embodiment. 1A to 1H, the data collection system 1 to which the data collection method according to the embodiment is applied will be described as an example.

As shown in FIG. 1A, the data collection system 1 according to the embodiment includes a data collection device 10 and in-vehicle devices 100-1, 100-2 mounted on vehicles V-1, V-2, V-3, . , 100-3 . . . and the user terminal 200. In the following description, the vehicle in general is referred to as "vehicle V", and the in-vehicle device as a whole is referred to as "in-vehicle device 100".

Moreover, below, the vehicle V shall be an automatic driving vehicle, and automatic operation control shall be carried out by the vehicle control model 102c each mounted in each vehicle V. FIG. The vehicle control model 102c is, for example, a learning model generated by analyzing vehicle data collected from the vehicle V and using machine learning such as deep learning based on the analysis results.

The data collection device 10 is configured as a cloud server that provides a cloud service via a network N such as the Internet or a mobile phone network, for example. , collects vehicle data from each in-vehicle device 100 and provides it to data users.

The in-vehicle device 100 is, for example, a drive recorder having various sensors such as a camera, an acceleration sensor, and a GPS (Global Positioning System) sensor, a storage device, a microcomputer, and the like. Data is taken from vehicle V.

The in-vehicle device 100 also uploads the collected vehicle data to the data collection device 10 as necessary. By using the drive recorder as the in-vehicle device 100 in this way, the in-vehicle components mounted on the vehicle V can be made more efficient. It should be noted that the in-vehicle device 100 and the drive recorder may be configured separately without using them together.

The user terminal 200 is a terminal used by a data user, and is for example a notebook PC (Personal Computer), a desktop PC, a tablet terminal, a PDA (Personal Digital Assistant), a smart phone, an eyeglass type, or a watch type information processing device. Examples include a wearable device, which is a terminal.

A data user is, for example, a developer who develops automatic driving technology based on vehicle data provided from the data collection device 10 . The data collection device 10 provides the data user with a user interface (hereinafter referred to as “UI”) screen that can be accessed via the user terminal 200 .

The data user designates vehicle data collection conditions via the UI screen, as shown in FIG. 1A (step S1). Then, the data collection device 10 that has received this distributes the collection conditions to each vehicle V, for example, in a file format (step S2).

The collection conditions include various parameters related to collection of vehicle data, as shown in FIG. 1B. Various parameters are, for example, "target vehicle", "data type", "collection trigger condition", "collection period", etc., as shown in FIG.

The "collection target vehicle" is the identification information of the vehicle V to be collected. Such a "collection target vehicle" corresponds to an example of a "collection target vehicle condition" which is one of the collection conditions. Further, the "data type" is the type of data to be collected, such as the accelerator opening rate, for example. The "collection trigger condition" is a condition that triggers collection, and is, for example, when the vehicle speed exceeds a predetermined speed. A "collection period" is a period during which data is collected. These "data type", "collection trigger condition", "collection period", etc. correspond to an example of "collection data condition" which is one of the collection conditions.

Then, each in-vehicle device 100 uploads the vehicle data collected in each vehicle V to the data collection device 10 constantly or based on an upload request from the data collection device 10 (step S3), and the data collection device 10 Accumulate (step S4). Then, the data user browses or downloads the vehicle data accumulated in the data collection device 10 via the aforementioned UI screen, for example (step S5), and uses it for analysis for development.

Although not shown in FIG. 1A, the vehicle data uploaded by the in-vehicle device 100 includes tag data T and actual data R. As shown in FIG.

More specifically, a series of flows until vehicle data is provided to data users in the data collection system 1, including this point, will be described with reference to FIGS. 1C to 1E. As shown in FIG. 1C, the data user first designates collection conditions using the user terminal 200 connected to the data collection device 10 .

Also, at this time, the data collecting device 10 generates data for generating tag data T that is added to the actual data R to be collected and has characteristics as index data used for searching the actual data R and for grasping the outline of the actual data R. to generate The data for generating the tag data T is generated based on the operation of the data user while using the program and data for generation stored in the user terminal 200 or the data collection device 10 .

Then, the specified collection conditions and the data for generating the generated tag data T are stored in the data collection device 10, distributed to the vehicle V that is the object of data collection, and also stored in the in-vehicle device 100. be done.

Next, each in-vehicle device 100 monitors the output data of various sensors, and stores the actual data R in the storage device when an event that satisfies the stored collection condition occurs. Further, each in-vehicle device 100 generates and stores tag data T corresponding to the actual data R based on the stored data for generation of the tag data T and the actual data R. FIG.

Note that the tag data T is not data that is a simple excerpt of a part of the actual data R. When a data user refers to the tag data T, he/she can grasp the outline of the actual data R and judge whether the actual data R is necessary or not. It is preferable that the information is converted to meta-information to some extent.

Then, each in-vehicle device 100 uploads the tag data T to the data collection device 10, and the data collection device 10 stores the tag data T. FIG. Note that the actual data R is not uploaded to the data collection device 10 at this time.

Then, when the data user connects to the data collection device 10 to check the data collection status or collect the actual data R using the user terminal 200, meta information based on the tag data T collected by the data collection device 10 is generated. is displayed on the user terminal 200 . At the same time, a UI screen for performing an operation of collecting actual data R corresponding to each tag data T is displayed.

Then, as shown in FIG. 1D, when the data user designates the tag data T corresponding to the actual data R to be collected using the user terminal 200, the actual data is sent to the corresponding on-vehicle device 100 via the data collection device 10. "Instruction data" specifying R is transmitted. In addition, hereinafter, such instruction data may be referred to as an "upload request".

After that, as shown in FIG. 1E, the designated actual data R is uploaded from each in-vehicle device 100 to the data collection device 10 and stored in the data collection device 10 . Then, the data user accesses the actual data R stored in the data collection device 10 using the user terminal 200, and browses or downloads the actual data R. FIG.

From the viewpoint of the storage capacity of the in-vehicle device 100, the actual data R uploaded to the data collection device 10 and the tag data T corresponding thereto are deleted from the in-vehicle device 100 after the upload to the data collection device 10 is completed. is preferred.

By the way, based on the tag data T collected by the data collection device 10 as described above, it is possible to determine how many versions of the vehicle control model 102c each vehicle V is equipped with, and where it is currently traveling. , the behavior of the vehicle V indicated by the CAN (Controller Area Network) data.

Therefore, in the data collection method according to the embodiment, first, based on the situation of each vehicle V, vehicles V exhibiting behaviors to be evaluated regarding automatic driving control are extracted. Then, along with the extracted vehicle V, a vehicle V similar to the vehicle V, such as the same vehicle type, the same version of the vehicle control model 102c, the number of passengers, the year of manufacture, the mileage, the weather, etc. Vehicles with similar running environments and the like are selected as vehicles to be collected for vehicle data for desired evaluation of the vehicle control model 102c.

Then, a new collection condition including these as collection target vehicles is specified, vehicle data is collected from the corresponding vehicle V based on the collection condition, the collected vehicle data is analyzed, and the vehicle data is analyzed based on the analysis result. We decided to update the control model 102c.

Specifically, as shown in FIG. 1F, in the data collection method according to the embodiment, for example, the user terminal 200 or the like is used to virtually grasp the situation of each vehicle V based on the tag data T that is constantly collected. Keep Further, here, regarding the automatic driving control, a predetermined determination condition is set in advance so that the vehicle V exhibiting the behavior to be evaluated can be determined. It does not matter whether the behavior to be evaluated is normal or abnormal, but for the sake of convenience of explanation, the vehicle V extracted under such determination conditions will be described as being "abnormal".

Here, it is assumed that the vehicle V-2 is determined to be abnormal as shown in FIG. 1F under these determination conditions (see the exclamation mark in the figure).

Then, in the data collection method according to the embodiment, as shown in FIG. 1G, the data collection device 10 extracts collection target vehicles based on the vehicle V-2 (step S11). For example, FIG. 1G shows an example in which a vehicle type "B", which is the same vehicle type as the vehicle V-2, and the same version "2.0" of the vehicle control model 102c are extracted as collection target vehicles.

Then, for example, the data collection device 10 designates a collection condition including each vehicle V that has become a collection target vehicle as a collection target vehicle condition (step S12). Such collection conditions may be the same for each vehicle V, or may be set individually so that vehicle data with variations can be collected. Of course, the data user may specify using the user terminal 200 instead of the data collection device 10 . The collection conditions are, for example, collection conditions set according to the type of abnormality, which are stored in a table format, and the table is searched according to the type of abnormality determined above. As a result of the search, collection conditions linked to the matched abnormality type are extracted. That is, the collection conditions can be registered in advance as preset information corresponding to the type of abnormality, and by using this, it is possible to simplify the designation of the collection conditions.

Then, the data collection device 10 distributes collection conditions to each vehicle V (step S13), and each in-vehicle device 100 uploads vehicle data based on the collection conditions (step S14), and the data collection device 10 is accumulated (step S15).

Then, as shown in FIG. 1H, the data collection device 10 analyzes the vehicle data accumulated in step S15 (see FIG. 1G) (step S16). Such an analysis is to analyze the contents such as when, where, what kind of control the vehicle control model 102c performed, and why. In addition, it analyzes what kind of learning data set should be provided for updating the vehicle control model 102c, in other words, how parameters relating to the control of the vehicle V should be set.

Then, the data collection device 10 updates the vehicle control model 102c based on the analysis result (step S17). Then, the data collection device 10 distributes the vehicle control model 102c to each vehicle V (step S18), and the in-vehicle device 100, for example, rewrites the vehicle control model 102c to a new one (step S19).

As described above, in the data collection method according to the embodiment, the vehicle V exhibiting the desired behavior regarding vehicle control is determined based on the vehicle data indicating the general condition of each vehicle V. FIG. Then, based on the result of the determination, conditions for collecting vehicle data used for evaluating parameters related to vehicle control are extracted.

Then, vehicle data for evaluation is collected based on the extracted collection conditions, and the collected vehicle data is analyzed.

Therefore, according to the data collection method according to the embodiment, it is possible to evaluate the suitability of the parameters related to the control of the vehicle V. Hereinafter, a configuration example of the data collection system 1 according to the embodiment will be described more specifically.

FIG. 2 is a block diagram showing a configuration example of the data collection system 1 according to the embodiment. It should be noted that FIG. 2 shows only the constituent elements necessary for explaining the features of the embodiment, and omits the description of general constituent elements.

In other words, each component illustrated in FIG. 2 is functionally conceptual and does not necessarily need to be physically configured as illustrated. For example, the specific form of distribution/integration of each block is not limited to the one shown in the figure. It is possible to integrate and configure.

In addition, in the description using FIG. 2, the description of components that have already been described may be simplified or omitted.

As shown in FIG. 2 , the data collection system 1 according to the embodiment includes a data collection device 10 , an in-vehicle device 100 and a user terminal 200 .

First, the data collection device 10 will be described. The data collection device 10 includes a communication section 11 , a storage section 12 and a control section 13 .

The communication unit 11 is implemented by, for example, a NIC (Network Interface Card) or the like. The communication unit 11 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the in-vehicle device 100 and the user terminal 200 via the network N.

The storage unit 12 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. It stores the judgment condition information 12b and the vehicle control model 102c.

Collected data DB 12a accumulates vehicle data collected from each in-vehicle device 100 according to collection conditions specified by user terminal 200 and such collection conditions. In other words, the collected data DB 12a includes collection conditions and past results of vehicle data collected thereby. The vehicle data referred to here includes the tag data T and the actual data R described above.

The determination condition information 12b is information about determination conditions for determining a vehicle V exhibiting a desired behavior regarding automatic driving control based on the tag data T indicating the outline of each vehicle V. FIG. Therefore, the determination condition information 12b includes data types for identifying the behavior of the vehicle V, threshold values, and the like.

The vehicle control model 102c is a learning model for automatically controlling each vehicle V, as already described, and is updated by the updating unit 13f, which will be described later.

The control unit 13 is a controller, and various programs stored in a storage device inside the data collection device 10 are stored in the RAM as a work area by, for example, a CPU (Central Processing Unit) or MPU (Micro Processing Unit). It is realized by executing as Also, the control unit 13 can be implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

The control unit 13 includes a collection unit 13a, a determination unit 13b, an extraction unit 13c, a distribution unit 13d, an analysis unit 13e, and an update unit 13f, and implements the information processing functions and actions described below. or run.

The collection unit 13a collects vehicle data uploaded from the in-vehicle device 100 based on collection conditions that are specified in advance and distributed, and stores the vehicle data in the collection data DB 12a.

The determination unit 13b determines a vehicle V exhibiting a behavior to be evaluated regarding automatic driving control based on the tag data T and the determination condition information 12b stored in the collected data DB 12a. Further, the determination unit 13b notifies the extraction unit 13c of the determination result.

The extraction unit 13c extracts a vehicle V to be a collection target vehicle for collecting vehicle data for evaluation of the vehicle control model 102c based on the determination result of the determination unit 13b. In addition, the extraction unit 13c sets a collection condition including the vehicle V, which is the extracted collection target vehicle, as a collection target vehicle condition.

The extraction unit 13c sets the collection condition based on, for example, parameters for evaluation of the vehicle control model 102c provided in advance. Further, the extracting unit 13c displays, for example, a list of vehicles V to be collection target vehicles on the UI screen of the user terminal 200, and sets collection conditions based on designation from the data user via the UI screen.

Here, an example of extraction processing by the extraction unit 13c will be described with reference to FIGS. 3A and 3B. 3A and 3B are diagrams (Part 1) and (Part 2) showing an example of extraction processing.

It is assumed here that the vehicle V-2 has been determined to be abnormal by the determination unit 13b, as already shown in FIG. 1F. In such a case, as shown in FIG. 3A, the extracting unit 13c extracts, for example, a vehicle of the same vehicle type and the same version of the vehicle control model 102c as the vehicle V-2 as a vehicle to be collected.

As shown in the figure, even if the model and version are the same, the situation of each vehicle V, such as "number of passengers", "year of manufacture", "mileage", and even "weather" at the current position, can be determined. may be narrowed down to those that are closer to each other, as they are usually different in various factors. For example, when narrowing down to similar items such as "number of passengers", "year of manufacture", "mileage" and "weather", at least vehicles V-2 and V-52 are selected here.

Then, as shown in FIG. 3B, collection data conditions are specified for each of the vehicles V-2 and V-52. Here, an example is shown in which all data types to be collected are "accelerator opening rate" and points to be collected are all "YYYY".

However, in order not to collect only similar vehicle data from vehicles V-2 and V-52 selected as similar collection target vehicles and bias the evaluation, the collected data conditions are set, for example, by the distance to be collected. It is preferable to specify different times, periods, etc. This can contribute to improving the versatility of the vehicle control model 102c.

Returning to the description of FIG. 2, the distribution unit 13d will be described. The distribution unit 13d distributes the collection conditions set by the extraction unit 13c to each vehicle V via the communication unit 11 in a file format, for example. The vehicle data for evaluation of the vehicle control model 102c based on the collection conditions distributed by the distribution unit 13d is collected by the collection unit 13a and stored in the collected data DB 12a.

The analysis unit 13e analyzes the vehicle data for evaluation of the vehicle control model 102c collected by the collection unit 13a, and notifies the update unit 13f of the analysis result. The update unit 13f performs, for example, additional learning based on the analysis result of the analysis unit 13e to update the vehicle control model 102c.

The vehicle control model 102c updated by the update unit 13f is delivered to each vehicle V to be updated (for example, the vehicle control model 102c is of the same version) by the delivery unit 13d.

Next, the in-vehicle device 100 will be described. The in-vehicle device 100 includes a communication section 101 , a storage section 102 and a control section 103 . In addition, as described above, the in-vehicle device 100 is connected to various sensors 150 such as a camera, an acceleration sensor, and a GPS sensor.

Like the communication unit 11, the communication unit 101 is implemented by, for example, a NIC. The communication unit 101 is wirelessly connected to the network N, and transmits and receives information to and from the data collection device 10 via the network N. The communication unit 101 also receives output data from various sensors 150 .

The storage unit 102 is implemented by, for example, a semiconductor memory device such as a RAM and a flash memory, and stores collection condition information 102a, vehicle data information 102b, and a vehicle control model 102c in the example of FIG.

The collection condition information 102a is information including the collection conditions distributed from the data collection device 10. FIG. The vehicle data information 102b is information including vehicle data collected by the collecting unit 103c, which will be described later. Vehicle data includes tag data T and actual data R described above. The vehicle control model 102c is a vehicle control model distributed from the data collection device 10. FIG.

Like the control unit 13, the control unit 103 is a controller, and is realized by, for example, executing various programs stored in a storage device inside the in-vehicle device 100 using a RAM as a work area by a CPU, MPU, or the like. be done. Also, the control unit 103 can be realized by an integrated circuit such as an ASIC or FPGA, for example.

The control unit 103 includes an acquisition unit 103a, a detection unit 103b, a collection unit 103c, and an upload unit 103d, and implements or executes information processing functions and actions described below.

The acquisition unit 103a acquires the collection conditions distributed from the data collection device 10 and stores them in the collection condition information 102a. The acquisition unit 103a also acquires a new vehicle control model 102c distributed from the data collection device 10 and rewrites the old vehicle control model 102c.

The detection unit 103b monitors the output data from the various sensors 150 and detects the occurrence of an event that serves as a trigger under the collection conditions.

For example, the detection unit 103b causes the collection unit 103c to collect vehicle data when detecting the occurrence of an event that triggers collection of vehicle data under the collection condition.

The collecting unit 103c collects vehicle data based on the output data of the various sensors 150 and stores the vehicle data in the vehicle data information 102b when the detection unit 103b detects the occurrence of a trigger for collecting vehicle data. Further, the collection unit 103c stops collection of vehicle data when the detection unit 103b detects generation of a trigger for stopping collection of vehicle data.

The upload unit 103d uploads the tag data T to the data collection device 10 all the time. Further, the upload unit 103 d uploads the actual data R to the data collection device 10 when an upload request for the actual data R is received from the data collection device 10 via the communication unit 101 . Note that the upload unit 103d may upload the actual data R and the like for evaluation of the vehicle control model 102c to the data collection device 10 immediately.

Next, a processing sequence executed by the data collection system 1 according to the embodiment will be described using FIG. FIG. 4 is a diagram showing a processing sequence executed by the data collection system 1 according to the embodiment.

First, the in-vehicle device 100 constantly transmits tag data T indicating the general condition of each vehicle V (step S101). Based on this, the data collection device 10 determines a desired behavior, for example, a desired abnormality (step S102).

Then, based on the determination result of step S102, the data collection device 10 extracts the collection target vehicle of the vehicle data for evaluation of the vehicle control model 102c (step S103), and designates collection conditions including the collection target vehicle ( step S104).

Then, the data collection device 10 distributes collection conditions to each vehicle V (step S105). The in-vehicle device 100 stores the distributed collection conditions (step S106). Then, the in-vehicle device 100 collects vehicle data (step S108) based on the occurrence of an event that triggers the collection of vehicle data (step S107).

The in-vehicle device 100 then uploads the collected vehicle data to the data collection device 10 (step S109), and the data collection device 10 collects the vehicle data (step S110).

Then, the data collection device 10 analyzes the vehicle data collected for evaluation of the vehicle control model 102c (step S111), and updates the vehicle control model 102c based on the analysis results (step S112).

Then, the data collection device 10 distributes the updated vehicle control model 102c to the in-vehicle device 100 (step S113), and the in-vehicle device 100 acquires it and rewrites the model (step S114).

By the way, although the vehicle control model 102c was mentioned as an example and the case where the suitability of the parameter regarding automatic driving control was evaluated until now was demonstrated, if it is about vehicle control, it will not be restricted to automatic driving control.

Therefore, the present embodiment can also be applied to the case of evaluating the suitability of parameters related to vehicle control during manual operation, such as parameters of control programs that operate in each ECU (Electronic Control Unit) installed in the vehicle V. can be done. In this case, data users are not limited to automatic driving technology, but are developers of vehicle control technology including manual driving.

As described above, the data collection device 10 according to the embodiment includes the collection unit 13a, the determination unit 13b, the extraction unit 13c, and the analysis unit 13e. The collection unit 13a collects vehicle data related to the vehicle V from the in-vehicle device 100 mounted on each vehicle V. FIG. Based on the vehicle data collected by the collection unit 13a, the determination unit 13b determines the vehicle V exhibiting the desired behavior. The extraction unit 13c extracts conditions for collecting evaluation data, which is vehicle data for evaluating parameters related to vehicle control, based on the determination result of the determination unit 13b. The analysis unit 13e analyzes the evaluation data collected by the collection unit 13a based on the collection conditions.

Therefore, according to the data collection device 10 according to the embodiment, it is possible to evaluate the suitability of parameters related to operational control.

Further, the determination unit 13b determines a vehicle V exhibiting an abnormal behavior based on the vehicle data.

Therefore, according to the data collection device 10 according to the embodiment, it is possible to evaluate the suitability of the parameters related to the driving control of the vehicle V exhibiting abnormal behavior.

The extraction unit 13c also extracts a vehicle V similar to the vehicle V determined to exhibit the desired behavior by the determination unit 13b as a collection target vehicle condition, which is one of the collection conditions.

Therefore, according to the data collection device 10 according to the embodiment, vehicles V in situations similar to the vehicle V exhibiting the above behavior can be collectively collected, and whether or not similar events occur under similar situations It is possible to collect a large amount of vehicle data for evaluating such points.

Further, the extracting unit 13c extracts a different collection data condition for the collection target vehicle with respect to the collection data condition, which is one of the collection conditions.

Therefore, according to the data collection device 10 according to the embodiment, it is possible to prevent the evaluation from being biased due to collection of only similar vehicle data from vehicles V selected as similar collection targets. .

Further, the parameters relate to the vehicle control model 102c in automatic driving control, and the data collection device 10 according to the embodiment updates the vehicle control model 102c by machine learning based on the analysis result of the analysis unit 13e. 13f, and a distribution unit 13d that distributes the vehicle control model 102c updated by the update unit 13f to the vehicle V to be updated.

Therefore, according to the data collection device 10 according to the embodiment, the vehicle control model 102c can be appropriately updated, and the vehicle control model 102c can be quickly updated when there is a parameter whose set value is found to be incompatible with the vehicle control model 102c. It is possible to deliver to the vehicle V which is different. That is, it can contribute to improving the safety of automatic operation control.

Further effects and modifications can be easily derived by those skilled in the art. Therefore, the broader aspects of the invention are not limited to the specific details and representative embodiments so shown and described. Accordingly, various changes may be made without departing from the spirit or scope of the general inventive concept defined by the appended claims and equivalents thereof.

1 data collection system 10 data collection device 12a collection data DB
12b determination condition information 13a collection unit 13b determination unit 13c extraction unit 13d distribution unit 13e analysis unit 13f update unit 100 in-vehicle device 102a collection condition information 102b vehicle data information 102c vehicle control model 103a acquisition unit 103b detection unit 103c collection unit 103d upload unit 150 Various sensors 200 User terminal R Actual data T Tag data V Vehicle

Claims (7)

Equipped with a control unit that collects vehicle data related to the vehicle from the in-vehicle device installed in each vehicle,
The control unit
Determining a vehicle exhibiting the desired behavior based on the collected vehicle data;
extracting conditions for collecting evaluation data, which is vehicle data for evaluating parameters related to vehicle control, based on the determined determination result;
analyzing the evaluation data collected based on the collection conditions ;
Update the vehicle control model by machine learning based on the analyzed analysis results,
Distribute the updated vehicle control model to the vehicle to be updated,
data collection device. The control unit
Determining a vehicle exhibiting anomalous behavior based on the vehicle data ;
A data collection device according to claim 1 . The control unit
extracting a vehicle similar to the vehicle determined to exhibit the desired behavior as a collection target vehicle condition, which is one of the collection conditions ;
3. A data collection device according to claim 1 or 2. The control unit
extracting a different collected data condition for a vehicle to be collected for the collected data condition which is one of the collection conditions ;
4. A data collection device according to claim 1, 2 or 3. The parameters relate to a vehicle control model in automatic driving control,
A data collection device according to any one of claims 1 to 4. Equipped with an in-vehicle device mounted on each vehicle and a data collection device,
The data collection device is
collecting vehicle data about the vehicle from the in-vehicle device;
Determining a vehicle exhibiting the desired behavior based on the collected vehicle data;
extracting conditions for collecting evaluation data, which is vehicle data for evaluating parameters related to vehicle control, based on the determined determination result;
analyzing the evaluation data collected based on the collection conditions;
Update the vehicle control model by machine learning based on the analyzed analysis results,
Distribute the updated vehicle control model to the vehicle to be updated,
data collection system. A data collection method performed by a data collection device,
Collecting vehicle data about the vehicle from an on-board device installed in each vehicle;
Determining a vehicle exhibiting the desired behavior based on the collected vehicle data;
Extracting conditions for collecting evaluation data, which is vehicle data for evaluating parameters related to vehicle control, based on the determined determination result;
analyzing the evaluation data collected based on the collection conditions ;
updating the vehicle control model by machine learning based on the analyzed analysis results;
distributing the updated vehicle control model to the vehicle to be updated;
data collection methods , including;

Images