Vehicle control system and vehicle control method

Abstract

This vehicle control system comprises a sensor unit, a vehicle control device having a vehicle control application, and a server. In the vehicle control device, difference data between a vehicle behavior predicted value and an actual vehicle behavior value calculated on the basis of a sensor value is calculated, and if said difference data is large, upload data is uploaded to the server and the vehicle control application and a vehicle model are updated to information delivered from the server. In the server, the vehicle model is improved on the basis of the upload data such that the vehicle behavior predicted value approaches the actual vehicle behavior value, and the vehicle control application is developed on the basis of the upload data and the improved vehicle model and the developed vehicle control application are distributed to the vehicle control device.

Description

Vehicle control system and vehicle control method

[0001] The present invention relates to a vehicle control system and a vehicle control method.

[0002] In recent years, automobiles have become increasingly intelligent, with advances in advanced driver assistance systems and the introduction of autonomous driving. Automobiles are controlled by software. Patent Document 1 discloses a technology for updating the software that controls automobiles.

[0003] Patent Document 1 states, "The software update system includes a vehicle data acquisition unit that acquires data related to the state of the vehicle, including data related to the vehicle's specified performance; an effect prediction unit that predicts, based on the data acquired by the vehicle data acquisition unit, the effect that will be obtained when update data for software used in the vehicle is applied to the vehicle to improve the specified performance; and a distribution unit that distributes the update data to the vehicle if the effect predicted by the effect prediction unit satisfies a specified standard, but refrains from distributing the update data to the vehicle if the effect predicted by the effect prediction unit does not satisfy the specified standard."

[0004] Japanese Patent Application Laid-Open No. 2023-43528

[0005] In the conventional technology described in Patent Document 1, data is uploaded in response to a request from the vehicle data acquisition unit, so there is a possibility that data that is not actually necessary may be uploaded, which is wasteful.

[0006] The present invention has been made in consideration of the above circumstances, and aims to provide a vehicle control system and a vehicle control method that can avoid unnecessary uploads by collecting and uploading only data on areas that truly require improvement.

[0007] A vehicle control system according to the present invention for solving the above problems includes a sensor unit that acquires vehicle information, a vehicle control device having a vehicle control application that calculates a control command value based on sensor values ​​from the sensor unit and a predicted value of a vehicle model, and a server. The vehicle control device includes a vehicle behavior prediction unit that calculates a vehicle behavior prediction value from environmental information calculated based on the control command value and the sensor values ​​and the predicted value of the vehicle model, a comparison unit that compares the predicted vehicle behavior value with an actual vehicle behavior value based on the sensor values ​​to calculate difference data, an upload determination unit that uploads upload data including the vehicle model, the sensor values, the control command value, and the actual vehicle behavior value to the server when the difference data is greater than a predetermined threshold, and a vehicle software update unit that updates the vehicle control application and the vehicle model to those distributed from the server. The server also includes a vehicle model improvement unit that improves the vehicle model based on the uploaded data so that the predicted vehicle behavior value approaches the actual vehicle behavior value, a vehicle control application development and verification unit that develops a vehicle control application for the vehicle control device based on the uploaded data, and a vehicle software distribution unit that distributes the improved vehicle model and the developed vehicle control application to the vehicle control device.

[0008] A vehicle control method for solving the above problems is a vehicle control method for a vehicle control system including a sensor unit that acquires vehicle information, a vehicle control device having a vehicle control app that calculates a control command value based on sensor values ​​from the sensor unit and a predicted value of a vehicle model, and a server. The vehicle control device performs the following processes: calculating a vehicle behavior prediction value from environmental information calculated based on the control command value and the sensor values ​​and the predicted value of the vehicle model; calculating difference data by comparing the vehicle behavior prediction value with an actual vehicle behavior value based on the sensor values; uploading upload data including the vehicle model, the sensor values, the control command value, and the actual vehicle behavior value to the server if the difference data is greater than a predetermined threshold; and updating the vehicle control app and the vehicle model to the vehicle control app and the vehicle model distributed from the server. The server also performs the following processes: improving the vehicle model based on the uploaded data so that the predicted vehicle behavior value approaches the actual vehicle behavior value; developing a vehicle control app for the vehicle control device based on the uploaded data; and distributing the improved vehicle model and the developed vehicle control app to the vehicle control device.

[0009] According to the present invention, by uploading data based on the difference between the information from the sensor unit that collects vehicle information and the information from the vehicle model, it is possible to collect and upload only data on parts that truly require improvement, thereby avoiding unnecessary uploads.

[0010] Problems, configurations, and effects other than those described above will become apparent from the following description of the mode for carrying out the invention (hereinafter referred to as the embodiment).

[0011] 1 is a block diagram illustrating an example of the overall configuration of a vehicle control system according to an embodiment of the present invention. FIG. 1 is a block diagram for explaining the functions of each functional unit and each program in the vehicle control system according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating an example of the processing of a vehicle behavior prediction unit in the vehicle control system according to an embodiment of the present invention. FIG. 3 is a flowchart illustrating an example of the processing of a vehicle control app in the vehicle control system according to an embodiment of the present invention. FIG. 4 is a flowchart illustrating an example of the processing of a comparison unit in the vehicle control system according to an embodiment of the present invention. FIG. 5 is a flowchart illustrating an example of the processing of an upload determination unit in the vehicle control system according to an embodiment of the present invention. FIG. 6 is a flowchart illustrating an example of the processing of a vehicle software update unit in the vehicle control system according to an embodiment of the present invention. FIG. 7 is a flowchart illustrating an example of the processing of a vehicle model improvement unit in the vehicle control system according to an embodiment of the present invention. FIG. 8 is a flowchart illustrating an example of the flow of a specific example of the processing of the vehicle model improvement unit in the vehicle control system according to an embodiment of the present invention. FIG. 9 is a flowchart illustrating an example of the processing of a vehicle control app development and verification unit in the vehicle control system according to an embodiment of the present invention. FIG. 10 is a flowchart illustrating an example of the processing of a vehicle control app development and verification unit in the vehicle control system according to an embodiment of the present invention.

[0012] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functions or configurations are designated by the same reference numerals, and redundant explanations will be omitted.

[0013] 1 is a block diagram showing an example of the overall configuration of a vehicle control system according to an embodiment of the present invention. The vehicle control system 1 according to this embodiment is configured to include an ECU (Electronic Control Unit) 10, which is an example of a vehicle control device, a server 20, a sensor unit 30, and another ECU 40.

[0014] The ECU 10 is a system mounted on a vehicle such as an automobile, and is connected to a server 20, a sensor unit 30, and other ECUs 40, and is capable of communicating via a network. The server 20 exists in the cloud and operates in conjunction with the ECU 10 mounted on the vehicle. The sensor unit 30 includes one or more sensors from various sensors for acquiring vehicle information, such as radar, LiDAR, a camera, and a gyro sensor. The sensor unit 30 outputs status information of the various sensors to the ECU 10. The other ECU 40 is an ECU that includes one or more actuators for operating the accelerator, brake, steering, etc. The other ECU 40 controls the vehicle based on input control information from the ECU 10.

[0015] [ECU (Vehicle Control Device)] The ECU 10, which is an example of a vehicle control device, includes a memory 11 and a CPU 12. The memory 11 is, for example, a random access memory (RAM), a solid state drive (SSD), or a hard disk drive (HDD), and stores programs and necessary information to be executed by the CPU 12. The CPU 12 includes a central processing unit (CPU) and an accelerator, and executes various processes according to the programs stored in the memory 11.

[0016] (Memory) The memory 11 stores a sensor value 110, a vehicle model 111, environmental information 112, a control command value 113, difference data 114, a vehicle behavior prediction value 115, an actual vehicle behavior value 116, a vehicle behavior prediction unit 117, a vehicle control application 118, a comparison unit 119, an upload determination unit 120, and a vehicle software update unit 121.

[0017] The sensor values ​​110, the environmental information 112, the control command values ​​113, the difference data 114, the vehicle behavior prediction values ​​115, and the actual vehicle behavior values ​​116 are time-series data, and are held for a certain period of time, including the latest values.

[0018] The vehicle model 111, the vehicle behavior prediction unit 117, the vehicle control application 118, the comparison unit 119, the upload determination unit 120, and the vehicle software update unit 121 are programs.

[0019] The functions of the vehicle behavior prediction unit 117, the vehicle control application 118, the comparison unit 119, the upload determination unit 120, and the vehicle software update unit 121 will be described later.

[0020] In the following description, for convenience, the program will be described as the entity that performs the operations, but the actual entity that executes the program is the CPU 12 that executes the program.

[0021] [Server] The server 20 includes a memory 21 and a CPU 22. The memory 21 is, for example, a RAM, an SSD, or an HDD, and stores programs and necessary information to be executed by the CPU 22. The CPU 22 executes each process in accordance with the programs stored in the memory 21.

[0022] (Memory) The memory 21 stores the programs of a vehicle model storage unit 211, an upload data unit 212, a vehicle model improvement unit 213, a vehicle control application development and verification unit 214, and a vehicle software distribution unit 215. The functions of each program will be described later.

[0023] In the following description, for convenience, the program may be described as the entity that performs the operations, but the actual entity that executes the program is the CPU 22 that executes the program.

[0024] [Functions of Functional Units and Programs] FIG. 2 is a block diagram for explaining the functions of the functional units and programs in the vehicle control system according to one embodiment of the present invention.

[0025] (ECU) In the ECU 10, the vehicle control application 118 calculates a control command value 113 based on the sensor value 110 of the sensor unit 30 and a predicted value of the vehicle model 111. Here, the control command value is information such as the throttle opening and steering amount.

[0026] The vehicle behavior prediction unit 117 calculates a vehicle behavior prediction value 115 from environmental information calculated based on the control command value calculated by the vehicle control application 118 and the sensor value 110, and a prediction value of the vehicle model 111. Here, the environmental information is, for example, information on the state of the road surface.

[0027] The comparison unit 119 compares the vehicle behavior predicted value 115 with the actual vehicle behavior value 116, and calculates the difference between them to obtain difference data 114. Here, the vehicle behavior predicted value is, for example, a value such as the vehicle position information or speed of the host vehicle. The actual vehicle behavior value 116 is a value based on the sensor value 110 of the sensor unit 30, i.e., a value calculated from the sensor value 110.

[0028] If the difference data 114 between the vehicle behavior prediction value 115 and the actual vehicle behavior value 116 is greater than a predetermined threshold, the upload determination unit 120 determines that uploading is necessary and uploads upload data including the vehicle model 111, the sensor value 110, the control command value 113, and the actual vehicle behavior value 116 to the server 20.

[0029] The vehicle software update unit 121 updates the vehicle control application 118 and the vehicle model 111 to the vehicle control application and the vehicle model distributed from the server 20 .

[0030] (Server) In the memory 21 of the server 20 , the vehicle model improving unit 213 updates the vehicle model 111 based on the upload data 212 from the ECU 10 so that the vehicle behavior predicted value 115 approaches the actual vehicle behavior value 116 .

[0031] The vehicle control application development and verification unit 214 updates the vehicle control application 118 based on the upload data 212 from the ECU 10 .

[0032] The vehicle software distribution unit 215 distributes the vehicle model improved by the vehicle model improvement unit 213 (improved vehicle model) and the vehicle control application developed by the vehicle control application development and verification unit 214 (developed vehicle model) to the ECU 10.

[0033] [Examples of processing by each functional unit] Next, examples of processing by each functional unit of the vehicle behavior prediction unit 117, vehicle control application 118, comparison unit 119, upload determination unit 120, and vehicle software update unit 121 in the ECU 10, as well as examples of processing by each functional unit of the vehicle model improvement unit 213, vehicle control application development and verification unit 214, and vehicle software distribution unit 215 in the server 20 will be described.

[0034] (Processing Example of Vehicle Behavior Prediction Unit) The processing of the vehicle behavior prediction unit 117 is, for example, processing of estimating to which point the vehicle will move at the next time from the control command value of the autonomous driving control app, and is realized by the vehicle behavior prediction unit 117 executing its function under the control of the CPU 12 of the ECU 10. Figure 3 is a flowchart showing an example of the processing of the vehicle behavior prediction unit 117 in the vehicle control system 1 according to one embodiment of the present invention.

[0035] The CPU 12 acquires the latest values ​​of the data of the sensor value 110, the environmental information 112, and the control command value 113 (step S11), and then predicts the vehicle behavior at the next time using the latest vehicle model stored in the vehicle model 111 based on the acquired data, and calculates the vehicle behavior prediction value (step S12). Next, the CPU 12 outputs the vehicle behavior prediction value calculated in the processing of step S12 to the vehicle behavior prediction value 115 (step S13).

[0036] In the process of step S11, the data to be acquired does not have to be the latest value, and may be, for example, time-series data for a certain period from the latest value.

[0037] (Processing Example of Vehicle Control App) The processing of the vehicle control app 118 is realized by the vehicle control app 118 executing its functions under the control of the CPU 12 of the ECU 10. Fig. 4 is a flowchart showing an example of the processing of the vehicle control app 118 in the vehicle control system 1 according to one embodiment of the present invention.

[0038] The CPU 12 acquires the latest values ​​of the sensor value 110, the environmental information 112, and the control command value 113 (step S21), and then calculates the control command value based on the acquired data (step S22). Next, the CPU 12 outputs the control command value calculated in the process of step S22 to the control command value 113 (step S23).

[0039] In the process of step S21, the acquired data does not have to be the latest value, and may be, for example, time-series data for a certain period from the latest value. In addition, in the process of step S22, the control command value may be calculated, for example, using a mathematical model based on control theory or a machine learning model. In addition, when the vehicle model is sequentially optimized by model predictive control, snapshots of the vehicle model may be taken and stored at regular intervals.

[0040] (Example of Processing by Comparison Unit) The processing by the comparison unit 119 is realized by the comparison unit 119 executing its functions under the control of the CPU 12 of the ECU 10. Fig. 5 is a flowchart showing an example of the processing by the comparison unit 119 in the vehicle control system 1 according to one embodiment of the present invention.

[0041] The CPU 12 acquires the latest vehicle behavior predicted value from the vehicle behavior predicted value 115 (step S31), and then acquires the latest actual vehicle behavior value from the actual vehicle behavior value 116 (step S32). Next, the CPU 12 calculates the difference between the latest vehicle behavior predicted value and the latest actual vehicle behavior value (step S33), and then outputs the calculated difference to the difference data 114 (step S34).

[0042] The data acquired in steps S32 and S33 does not have to be the latest value, and may be, for example, time-series data for a certain period from the latest value. Furthermore, the process of calculating the difference in step S33 may be, for example, an error or squared error between the vehicle behavior prediction value and the actual vehicle behavior value. In the case of time-series data, the average error or mean squared error may be used.

[0043] (Example of Processing by Upload Determination Unit) The processing by the upload determination unit 120 is realized by the upload determination unit 120 executing its functions under the control of the CPU 12 of the ECU 10. Fig. 6 is a flowchart showing an example of processing by the upload determination unit 120 in the vehicle control system 1 according to one embodiment of the present invention.

[0044] The CPU 12 defines an analysis period based on the differential data 114, acquires differential data within the analysis period (step S41), and then performs statistical analysis on the acquired differential data to calculate statistical values ​​(e.g., mean and variance) (step S42). Next, the CPU 12 determines whether the calculated statistical values ​​are equal to or greater than a threshold (step S43).

[0045] In the processing of step S43, the CPU 12 determines that uploading is necessary because the error is large, for example, when the average is far from 0 or the variance is larger than a threshold. That is, when the CPU 12 determines that the calculated statistical value is equal to or greater than the threshold (YES in S43), it uploads the difference data within the analysis period defined in step S41 (step S44), and then determines whether the termination condition is met (step S45).

[0046] In the process of step S44, the CPU 12 uploads the difference data within the analysis period defined in step S41 from the sensor values ​​110, the parameters of the vehicle model 111, the environmental information 112, the control command values ​​113, the difference data 114, the vehicle behavior prediction values ​​115, and the actual vehicle behavior values ​​116, with tags such as the vehicle ID and time. Furthermore, if the CPU 12 determines in the process of step S43 that the calculated statistical value is not equal to or greater than the threshold value (NO in S43), the process proceeds to step S45.

[0047] If the CPU 12 determines in the process of step S45 that the termination condition is met (YES in S45), it terminates the series of processes of the upload determination unit 120, and if it determines that the termination condition is not met (NO in S45), it returns to step S41, sets an analysis period different from the previous time, and executes the process again. Here, the termination condition is not met when, for example, there is data remaining in the differential data that has not yet been analyzed.

[0048] In the process of step S41, the differential data within the analysis period may be, for example, data for one hour, one day, or one week. Furthermore, the data may not be continuous, such as data for one week containing only nighttime data. In the process of step S42, the statistical analysis may be performed using parameters of an approximation curve of the time-series data of the differential data, or may be performed using a machine learning model (NN, SVM, Bayesian estimation). In the process of step S43, the tag attached to the uploaded data may be, for example, information identifying the user or passenger.

[0049] (Example of processing by the vehicle software update unit) The processing by the vehicle software update unit 121 is realized by the vehicle software update unit 121 executing its functions under the control of the CPU 12 of the ECU 10. Figure 7 is a flowchart showing an example of processing by the vehicle software update unit 121 in the vehicle control system 1 according to one embodiment of the present invention.

[0050] The CPU 12 downloads a new version of the vehicle model (step S51), and then downloads a new version of the vehicle control app (step S52). Next, the CPU 12 replaces the current vehicle model with the new version of the vehicle model (step S53), and then replaces the current vehicle control app with the new version of the vehicle control app (step S54).

[0051] (Example of Processing by Vehicle Model Improvement Unit) The processing by the vehicle model improvement unit 213 is realized by the vehicle model improvement unit 213 executing its functions under the control of the CPU 22 of the server 20. Fig. 8 is a flowchart showing an example of processing by the vehicle model improvement unit 213 in the vehicle control system 1 according to one embodiment of the present invention.

[0052] The CPU 22 selects a vehicle equipped with a vehicle model to be updated as a target vehicle (step S61), and then reads out each data item of the vehicle model parameters, environmental information, control command values, difference data, actual vehicle behavior values, and predicted vehicle behavior values ​​of the target vehicle from the uploaded data 212 (step S62). Next, the CPU 22 improves the vehicle model based on the data read out in the processing of step S62 (step S63), and then stores this improved vehicle model in the vehicle model storage unit 211 (step S64).

[0053] In the process of step S61, the vehicle to be selected is not limited to one vehicle, but may be, for example, a group of multiple vehicles, a user, or a user group. In addition, in the process of step S62, for example, past uploaded data or uploaded data of other vehicles may be additionally read. In addition, in the process of step S63, the parameters of the vehicle model may be optimized, or the structure of the vehicle model may be changed.

[0054] 9 schematically shows an example of the flow of specific processing by the vehicle model improvement unit 213 executed under the control of the CPU 22. The CPU 22 (1) extracts input / output data to and output from the vehicle model and comparison difference data as learning data from the upload data 212, and (2) extracts vehicle model parameters. The CPU 22 then (3) creates a base vehicle model from the vehicle model parameters, and (4) adjusts and verifies the parameters of the base vehicle model using the training data set. The verification simulator model and the vehicle control application model are output to the vehicle control application development and verification unit 214.

[0055] (Example of Processing by Vehicle Control Application Development and Verification Unit) The processing by the vehicle control application development and verification unit 214 is realized by the vehicle control application development and verification unit 214 executing its functions under the control of the CPU 22 of the server 20. Fig. 10 is a flowchart showing an example of processing by the vehicle control application development and verification unit 214 in the vehicle control system 1 according to one embodiment of the present invention.

[0056] The CPU 22 selects a vehicle equipped with a vehicle control application to be updated as a target vehicle (step S71), and then acquires an updated vehicle model linked to the target vehicle from the vehicle model storage unit 211, and reads environmental information, control command values, difference data, actual vehicle behavior values, and predicted vehicle behavior values ​​linked to the vehicle from the upload data 212 (step S72).

[0057] Next, the CPU 22 improves the vehicle control application based on the data read in step S72 (step S73), and then performs simulation verification using the improved vehicle control application, the improved vehicle model, and the uploaded data (step S74).The CPU 22 then transmits the improved vehicle control application to the vehicle software distribution unit 215 (step S75).

[0058] In the process of step S71, the vehicle to be selected is not limited to one vehicle, but may be, for example, a group of multiple vehicles, a user, or a user group.

[0059] 11 schematically shows an example of the flow of specific processing by the vehicle control application development and verification unit 214 executed under the control of the CPU 22. The CPU 22 (1) improves the vehicle control application and (2) improves the verification simulator. The CPU 22 also (3) extracts input data to the vehicle control application as verification data. The CPU 22 then (4) verifies the updated vehicle control application using the updated simulator.

[0060] (Example of Processing by Vehicle Software Distribution Unit) The processing of the vehicle software distribution unit 215 is realized by the vehicle software distribution unit 215 executing its functions under the control of the CPU 22 of the server 20. Fig. 12 is a flowchart showing an example of processing of the vehicle software distribution unit 215 in the vehicle control system 1 according to one embodiment of the present invention.

[0061] The CPU 22 selects a vehicle equipped with a vehicle control application to be updated (step S81), and then acquires a vehicle model corresponding to the vehicle control application to be updated from the vehicle model storage unit 211 (step S82).The CPU 22 then transmits the improved vehicle control application and the corresponding program data of the improved vehicle model to the target vehicle (step S83).

[0062] [Actions and Effects of the Present Embodiment] As described above, the vehicle control system (vehicle control method) according to the present embodiment records differential data between the sensor values ​​(actual vehicle behavior) of the sensor unit 30 that acquires vehicle information and the vehicle model, and uploads data based on the differential data. This makes it possible to collect and upload only data on parts that truly require improvement, thereby avoiding unnecessary uploads. Furthermore, because data is uploaded based on differential data, it is also possible to collect edge cases that would not have been anticipated in advance.

[0063] Furthermore, since the vehicle model is improved by the cloud server 20, data from multiple vehicles can be used to improve the vehicle model, which makes it possible to avoid local solutions that should not be optimized, such as in a faulty state.

[0064] Furthermore, when the sensor values ​​(actual vehicle behavior) of the sensor unit 30 are compared with the vehicle model, if the difference between them is large, it can be said that the accuracy of the vehicle model has deteriorated. Vehicle data can be collected in response to such scenes of accuracy deterioration and used to improve the vehicle model. Furthermore, since it is determined whether to upload if the difference is large, it is possible to automatically detect and collect only data related to scenes in which the accuracy of the conventional vehicle model deteriorates, including scenarios that were not anticipated in advance. By applying this method to multiple vehicles, a variety of data can be collected, and the data can be used to improve the accuracy of the vehicle model.

[0065] <Modifications> The present invention is not limited to the embodiment described above and shown in the drawings, and various modifications are possible within the scope of the invention as set forth in the claims. Furthermore, the above-described embodiment has been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to an embodiment having all of the described configurations.

[0066] Furthermore, in the above-described embodiment, the upload determination unit is configured to upload the upload data to the server when the difference data between the vehicle behavior prediction value and the actual vehicle behavior value is larger than a predetermined threshold. However, the upload determination unit may be configured to upload the upload data to the server when the difference data is statistically large. Here, the difference data being statistically large refers to, for example, when the average absolute value and variance of the difference data are large. With this configuration, the upload determination unit 120 performs statistical processing on the time-series data and determines whether or not there is a difference within a certain interval, thereby filtering outliers such as noise. As a result, unnecessary uploads based on instantaneous differences such as noise can be avoided.

[0067] Furthermore, apart from the configurations of the above-described embodiments, the processing of the vehicle model and the vehicle behavior prediction unit may be configured to operate in the background, separate from the vehicle model and the vehicle behavior prediction unit, and unrelated to vehicle control. This configuration allows for in-vehicle execution even for vehicles under development that have not yet undergone sufficient safety verification, and allows data to be uploaded based on the difference between the vehicle behavior prediction values ​​and the actual vehicle behavior values. As a result, data on difficult situations for the vehicle model under development can be uploaded using realistic actual vehicle data and an execution environment.

[0068] DESCRIPTION OF SYMBOLS 1... Vehicle control system, 10... ECU, 11... Memory, 12... CPU, 20... Server, 30... Sensor unit, 40... Other ECU, 110... Sensor value, 111... Vehicle model, 112... Environmental information, 113... Control command value, 114... Differential data, 115... Vehicle behavior predicted value, 116... Actual vehicle behavior value, 117... Vehicle behavior prediction unit, 118... Vehicle control application, 119... Comparison unit, 120... Upload determination unit, 121... Vehicle software update unit, 211... Vehicle model storage unit, 212... Upload data, 213... Vehicle model improvement unit, 214... Vehicle control application development and verification unit, 215... Vehicle software distribution unit

Claims

1. A vehicle control system comprising: a sensor unit that takes in vehicle information; a vehicle control device having a vehicle control application that calculates a control command value based on the sensor value of the sensor unit and a predicted value of a vehicle model; and a server, wherein the vehicle control device comprises: a vehicle behavior prediction unit that calculates a vehicle behavior prediction value from environmental information calculated based on the control command value and the sensor value and the predicted value of the vehicle model; a comparison unit that compares the vehicle behavior prediction value with an actual vehicle behavior value based on the sensor value to calculate difference data; an upload determination unit that, when the difference data is greater than a predetermined threshold, uploads upload data including the vehicle model, the sensor value, the control command value, and the actual vehicle behavior value to the server; and a vehicle software update unit that updates the vehicle control application and the vehicle model to a vehicle control application and a vehicle model distributed from the server, wherein the server comprises: a vehicle model improvement unit that improves the vehicle model based on the uploaded data so that the vehicle behavior prediction value approaches the actual vehicle behavior value; and a vehicle control application development and verification unit that develops a vehicle control application for the vehicle control device based on the uploaded data. a vehicle software distribution unit that distributes the improved vehicle model and the developed vehicle control application to the vehicle control device.

2. The vehicle control system according to claim 1, wherein the upload determination unit uploads upload data including the vehicle model, the sensor value, the control command value, and the actual vehicle behavior value to the server when the difference data is statistically large.

3. A vehicle control system including a sensor unit that takes in vehicle information, a vehicle control device having a vehicle control application that calculates a control command value based on the sensor value of the sensor unit and a predicted value of a vehicle model, and a server, wherein the vehicle control device executes the following processes: calculating a vehicle behavior prediction value from environmental information calculated based on the control command value and the sensor value and the predicted value of the vehicle model; calculating difference data by comparing the vehicle behavior prediction value with an actual vehicle behavior value based on the sensor value; uploading upload data including the vehicle model, the sensor value, the control command value, and the actual vehicle behavior value to the server when the difference data is greater than a predetermined threshold; and updating the vehicle control application and the vehicle model to the vehicle control application and vehicle model distributed from the server; and the server executes the following processes: improving the vehicle model based on the uploaded data so that the vehicle behavior prediction value approaches the actual vehicle behavior value; developing a vehicle control application for the vehicle control device based on the uploaded data; and distributing the improved vehicle model and the developed vehicle control application to the vehicle control device. Vehicle control method.

Images