JP2017129039A - Vehicle control device and vehicle control parameter learning system - Google Patents

Abstract

In a vehicle control apparatus that updates control parameters by learning, a technique is provided that can update a learning value within an appropriate range even if the learning condition is not yet satisfied. When an abnormal learning value is reset, the vehicle control device according to the present invention acquires the learning value of another vehicle from a server that aggregates the learning values of the other vehicles and resets the learning value after resetting the learning value. Used as the initial value of. [Selection] Figure 1

Description

  The present invention relates to a technique for learning control parameters of a vehicle control device.

  Conventionally, various learning methods have been proposed as a technique for an electronic control device that controls an engine included in a vehicle to learn control parameters. Here, learning is to update a control parameter suitable for individual differences inherent in the vehicle over time so as to match the vehicle starting from an initial value. As a result, the vehicle is controlled using the optimum control parameter for each individual vehicle.

  In the following Patent Document 1, the basic fuel injection amount determined according to the intake air amount and the engine speed is calculated based on the engine coolant noise, the throttle valve opening, the air-fuel ratio of the exhaust gas output from the air-fuel ratio sensor, and the target air-fuel ratio. A learning control method that absorbs the deviation is disclosed.

  In the technique described in Patent Document 2 below, a multiplication term for the basic fuel injection amount is learned in the normal operation state, and an addition term for the basic fuel injection amount is learned in the idle operation state. An allowable limit is set for at least one of the multiplication term and the addition term. This prevents the learning value from becoming an abnormal value and the air-fuel ratio from shifting to an abnormal state even when the control circuit fails.

  In the technique described in Patent Document 3 below, a value that is not normally obtained as a learning value is stored when the engine is not yet learned for each operating state of the engine, and a value corresponding to this is stored when the engine is already learned. Is determined. The learning value is rewritten at once for the unlearned driving state, and the learning value is rewritten little by little for the already learned driving state.

  On the other hand, there is a technique for acquiring and updating learning parameters from outside the vehicle after the vehicle is shipped. Patent Document 4 below discloses a technology for rewriting parameters of an in-vehicle ECU (Electronic Control Unit) after shipment by connecting a learning server and a vehicle via a wireless line and transmitting learning parameters from the learning server to the vehicle. is suggesting.

JP-A-57-32844 JP 58-150058 A JP 58-176440 A Japanese Patent Laying-Open No. 2015-135552

  In the conventional learning control as described in Patent Document 1, for example, an air-fuel ratio feedback based on a sensor output is performed for each operation region during steady engine operation with respect to a learning value table using engine speed and load as parameters. Based on the deviation between the average value of the correction coefficient for a predetermined time and the target value, a learning value as a correction value for realizing the fuel injection amount at the stoichiometric air-fuel ratio is plotted and used as a control variable. However, in this method, only the portion at the time of steady engine operation is learned from the learning value table having the engine speed and the load as components, so there is an operation region that is not sufficiently learned depending on the driving conditions of the automobile. Therefore, there is a case where the learning control does not progress according to the initial target.

  In the technique described in Patent Document 2, when the learning value is updated, the learning value is replaced within the allowable limit only when the learning value exceeds the allowable limit. Therefore, it can suppress that a learning value exceeds an allowable range. On the other hand, the learning value may reach an abnormal value due to accumulation of abnormal learning states due to an abnormal sensor output value or a CPU (Central Processing Unit) runaway. Since the learning value is updated when a predetermined update condition is established during steady engine operation, if the learned value becomes abnormal except during steady engine operation, it takes a relatively long time to return to the normal range. . Then, for example, if the learned value once becomes an abnormal value and then the acceleration operation and the deceleration operation are repeated and the engine is operated without being in a steady operation state, the abnormal learned value is used for engine control, causing engine trouble. May occur.

  In the technique described in Patent Document 3, as in Patent Document 2, the learning value is rewritten only when the learning condition is satisfied. Therefore, as in Patent Document 2, it takes a relatively long time for the abnormal learning value to return to the normal range. For example, even if an abnormal learning value is once initialized and returned to the normal range, it takes time to return to the learning state immediately before the initialization again.

  The present invention has been made in view of the above problems, and in a vehicle control device that updates control parameters by learning, the learning value can be updated within an appropriate range even if the learning conditions are not yet satisfied. The purpose is to provide technology.

  When an abnormal learning value is reset, the vehicle control device according to the present invention acquires the learning value of another vehicle from a server that aggregates the learning values of other vehicles, and uses this as the initial value of the learning value after reset. Use.

  According to the vehicle control device of the present invention, the vehicle can be controlled using an appropriate learning value even when the learning timing has not yet been reached after the learning value is reset.

1 is an overall view of a control parameter learning system. 1 is a block diagram showing a configuration of a vehicle control device 100. FIG. 5 is a diagram illustrating an example of a control parameter 150. FIG. It is a figure which illustrates the part which is not learned among learning value maps. 5 is a flowchart for explaining a procedure for the vehicle 11 to transmit a learned control parameter to the server 16. It is a flowchart explaining the procedure in which the server 16 collects control parameters from each vehicle. 4 is a flowchart illustrating a procedure for distributing control parameters from a server 16 to a vehicle 200. It is a flowchart explaining the procedure which the vehicle control apparatus 100 acquires the control parameter of another vehicle from the server 16, and uses this as its own control parameter.

  FIG. 1 is an overall view of a control parameter learning system according to the present invention. In this system, the server 16 updates the learned control parameters from each vehicle and provides them to other vehicles, thereby updating the unlearned control parameters to an appropriate learned state. The following description is based on the assumption that the vehicle 11 provides the learned control parameter and the vehicle 200 receives this via the server 16.

  The vehicle 11 transmits the learned control parameter to the server 16 via the wireless line 12, the relay station 13, the wireless line 14, and the base station 15. The server 16 classifies the received control parameters according to the classifications described later, and holds each of the classifications. The vehicle 200 receives control parameters from the server 16 via the base station 17, the radio line 18, the relay station 19, and the radio line 20, and stores the control parameters as control parameters of its own vehicle control device (ECU) 100. .

  The server 16 can be implemented as a computer having a program that implements a process for collecting control parameters from each vehicle, a CPU that executes the program, a storage device that stores the control parameters, and the like, and has equivalent functions. It can also be implemented using hardware such as a circuit device.

  FIG. 2 is a block diagram illustrating a configuration of the vehicle control device 100. The vehicle control device 100 is a device that controls the operation of the vehicle 200, and includes a CPU 110, a control unit 120, a communication unit 130, and a storage device 140.

  The control unit 120 is a functional unit that performs a control calculation for controlling the vehicle 200. The control unit 120 performs a control calculation using the control parameter 150 and further updates the parameter by reflecting a parameter unique to the vehicle 200 to the control parameter 150. The control unit 120 can be configured by a circuit device that implements the function, or can be configured by the CPU 110 executing software that implements an equivalent function. The following description is based on the assumption that the CPU 110 executes the control unit 120 as shown in FIG. However, for convenience of description, the control unit 120 may be described as an operation subject.

  The communication unit 130 communicates with the outside of the vehicle using an appropriate communication interface. For example, the server 16 is requested to transmit the control parameter, and the control parameter is received from the server 16 as a response. The control unit 120 stores the received control parameter as its own control parameter 150 in the storage device 140.

  FIG. 3 is a diagram illustrating an example of the control parameter 150. Here, a learning value map generally used when recording a learning result on the ignition timing of the engine is illustrated. The horizontal axis is the engine speed, and the vertical axis is the engine load. By querying the map using these two parameters as keys, the corresponding learning value can be obtained. Since the ideal engine ignition timing has characteristics specific to the vehicle, the control unit 120 learns the parameters specific to the vehicle 200 by, for example, the methods described in Patent Documents 1 to 3, and uses the result as a new control parameter 150. As overwritten can be updated.

  FIG. 4 is a diagram illustrating an unlearned portion of the learning value map. In actual traveling, it is difficult to evenly learn all the regions indicated by the learning value map as shown in FIG. 3, and there is a possibility that the learning frequency of the regions indicated by the diagonal lines in FIG. Therefore, regarding the control parameters of these unlearned areas, it is considered that the control parameters of other vehicles are acquired from the server 16 and diverted to obtain the same state as the vehicle control device 100 has learned by itself. It is done.

  Since it is desirable for the vehicle control device 100 to learn individual differences inherent to the vehicle 200 in principle, it is not appropriate to divert the control parameters of other vehicles immediately after the vehicle 200 is shipped. Therefore, as a timing suitable for the vehicle control device 100 to acquire control parameters of another vehicle, for example, a time point when the learned control parameter reaches an abnormal value such as exceeding a preset allowable range can be considered. When the learning result reaches an abnormal value, the vehicle control device 100 generally initializes the learned control parameter, so that the previous learning result becomes blank, and it takes time until a sufficient learning result is obtained again. Because.

  An unlearned region shown in FIG. 4 is suitable for using the control parameter acquired from another vehicle as the initial value when initializing the learning value map. However, also for the learned region, a control parameter acquired from another vehicle may be used as an initial value in view of the fact that an appropriate time is required until the learning region has been initialized once and sufficient learning experience has passed. Which region is acquired from another vehicle can be appropriately determined based on various conditions such as vehicle characteristics.

  FIG. 5 is a flowchart illustrating a procedure in which the vehicle 11 transmits the learned control parameter to the server 16. This flowchart can be implemented periodically by the vehicle control device provided in the vehicle 11, for example. Hereinafter, for convenience of description, the vehicle 11 is assumed to be an operation subject. Hereinafter, each step of FIG. 5 will be described.

(FIG. 5: Step S501)
The vehicle 11 determines whether or not a condition for transmitting the control parameter to the server 16 is satisfied. If it is established, the process proceeds to step S502, and if it is not established, this flowchart is terminated. As a transmission condition in this step, for example, the following conditions are assumed on the assumption that the vehicle 11 holds a control parameter that can be transmitted and that a transmission requirement such as a communication line is not interrupted is satisfied. Things can be considered.

(FIG. 5: Step S501: Example 1 of transmission conditions)
For example, the server 16 can periodically request the vehicle 11 to transmit the currently held control parameter. The vehicle 11 can make the transmission condition to receive the request. For example, a case where the vehicle 11 and the server 16 are fixedly connected during maintenance of the vehicle 11 is suitable for using this condition.

(FIG. 5: Step S501: Example 2 of transmission conditions)
For example, when the ignition switch is turned on or off, the vehicle 11 can transmit the currently held control parameter to the server 16. For example, when the data size of the control parameter is large and the communication load becomes excessive when frequently transmitted, it is desirable to use this transmission condition.

(FIG. 5: Step S501: Example 3 of transmission conditions)
The vehicle 11 can transmit the currently held control parameter to the server 16 every time the control parameter is updated. Among the control parameters, those having a relatively small data size and those having a relatively large fluctuation range and a large influence on the control of the vehicle can be transmitted only according to the present transmission condition.

(FIG. 5: Step S502)
The vehicle 11 transmits the currently held control parameter to the server 16. At this time, (a) the vehicle type or model number of the vehicle 11, (b) the distance traveled by the vehicle 11, (c) the number of years since the vehicle 11 was shipped, (d) the type of optional parts mounted on the vehicle And (e) the type and range of the control parameter to be transmitted to the server 16 are transmitted together.

(FIG. 5: Step S502: Supplement 1)
When the vehicle types and model numbers of the vehicles are the same, the control parameters that have been learned by any vehicle are considered to be generally useful for other vehicles. Therefore, the server 16 classifies the control parameters for each vehicle type (or model number, the same applies hereinafter), and provides learned control parameters provided from vehicles of the same vehicle type. Therefore, in this step, the above (a) is also transmitted, and the server 16 classifies and holds the control parameters according to the information. The above (b) to (e) are also transmitted for the same reason, and the server 16 can classify the control parameters by classifying them.

(FIG. 5: Step S502: Supplement 2)
Depending on the types of optional parts that the vehicle 11 has, there are some that affect the running performance of the vehicle 11, and such parts are considered to also affect the learning result of the control parameter. For example, an optional part that affects the vehicle weight (such as a loading platform installed on the roof) is an example. Therefore, in this step, the information can be transmitted as an attribute of the vehicle 11.

(FIG. 5: Step S502: Supplement 3)
The vehicle 11 does not necessarily have to transmit all the control parameters held by itself to the server 16 and can transmit only necessary ones. Furthermore, even if the control parameters are the same, some of the array elements (for example, shaded portions in FIG. 4) are configured as array data as in the learning value maps illustrated in FIGS. Can also be transmitted to the server 16. Therefore, in this step, in order to specify which control parameter or which range is to be transmitted, the above (e) can also be transmitted.

  FIG. 6 is a flowchart illustrating a procedure for the server 16 to collect control parameters from each vehicle. The server 16 collects control parameters from each vehicle by executing this flowchart every time the control parameters are received from any of the vehicles. Hereinafter, each step of FIG. 6 will be described.

(FIG. 6: Step S601)
The server 16 determines whether or not a condition for collecting the received control parameters is satisfied. If established, the process proceeds to step S602, and if not established, this flowchart is terminated. The aggregation condition in this step includes, for example, that the received control parameter and vehicle attribute information (information such as the vehicle type transmitted together with the control parameter in step S502) are not damaged / deficient.

(FIG. 6: Step S602)
The server 16 classifies the control parameters according to the vehicle attribute information received in step S601, and updates the control parameters of the corresponding classification according to the received values. Since learning results may be received for the same control parameter from a plurality of vehicles, it is not always necessary to retain the last received value. For example, (a) For the same control parameter, an average value is held based on the number of transmission source vehicles, (b) A weighted average with a large weight is given to any value before and after the update, (c) Predetermined An example in which the updated value is obtained in consideration of the correction value is considered.

  FIG. 7 is a flowchart illustrating a procedure for distributing control parameters from the server 16 to the vehicle 200. For example, when the server 16 receives a request for distributing control parameters from the vehicle 200 (that is, the vehicle control device 100), the server 16 can execute this flowchart. Hereinafter, each step of FIG. 7 will be described.

(FIG. 7: Step S701)
The server 16 determines whether a condition for distributing the control parameter to the vehicle 200 is satisfied. More specifically, it is necessary that the server 16 holds control parameters that can be transmitted, and that transmission requirements such as the communication line is not interrupted are satisfied. If the condition is satisfied, the process proceeds to step S702. If the condition is not satisfied, for example, a distribution error is returned to the vehicle 200, and then the present flowchart is ended.

(FIG. 7: Step S702)
The server 16 specifies the control parameter to be transmitted according to the attribute information notified when the vehicle 200 requests to transmit the control parameter. The attribute information here is the same as that transmitted when the vehicle 11 transmits the control parameter in step S502.

(FIG. 7: Step S703)
Server 16 transmits the control parameter specified in step S <b> 702 to vehicle 200.

  FIG. 8 is a flowchart illustrating a procedure in which the vehicle control device 100 acquires control parameters of another vehicle from the server 16 and uses them as its own control parameters. This flowchart is executed by the control unit 120 included in the vehicle control device 100. Hereinafter, each step of FIG. 8 will be described.

(FIG. 8: Step S800)
The control unit 120 periodically checks whether or not an abnormal value has been reached, for example, based on whether or not the control parameter 150 exceeds a preset allowable range. When the abnormal value has been reached, the control parameter 150 is initialized and this flowchart is started. The range to be initialized does not necessarily need to cover the entire range of all the control parameters 150. The individual control parameters that have reached an abnormal value, and among those array elements that have reached an abnormal value (for example, the hatched lines in FIG. 4). Part) can also be initialized.

(FIG. 8: Step S800: Supplement)
The control unit 120 may initialize the control parameter 150 even when the control parameter 150 does not necessarily reach the abnormal value. For example, initialization associated with maintenance work can be considered. The control unit 120 can implement this flowchart even in such a case.

(FIG. 8: Step S801)
The control unit 120 determines whether a condition for acquiring the control parameter of the other vehicle from the server 16 is satisfied. Specifically, the communication unit 130 needs to be in a state where it can receive the control parameter. If the condition is satisfied, the process proceeds to step S802. If the condition is not satisfied, for example, the flowchart is ended after setting the flowchart again after a predetermined time.

(FIG. 8: Step S802)
The control unit 120 transmits a request for transmitting the control parameter to the server 16 together with the attribute information of the vehicle 200. The attribute information here is for specifying a control parameter acquired from the server 16, and is the same as in steps S502 and S702.

(FIG. 8: Step S802: Supplement)
The control unit 120 can request the server 16 to transmit only control parameters that need to be acquired from other vehicles and their array elements. For example, only a portion (for example, the shaded portion in FIG. 4) that has not been learned at the start of this flowchart in the learned value map of the engine ignition timing can be acquired from the server 16. Alternatively, when it is desirable to reach a state equivalent to already learned earlier, the control parameter or its array element that has been learned at the start of this flowchart can also be acquired from the server 16.

(FIG. 8: Step S803)
The control unit 120 receives the control parameter requested from the server 16 via the communication unit 130.

(FIG. 8: Step S804)
The control unit 120 stores the received control parameter as a new control parameter 150 in the storage device 140. Thereafter, the control unit 120 controls the vehicle 200 using the new control parameter 150.

<Summary of the present invention>
When the vehicle control device 100 according to the present invention is initialized when the control parameter 150 reaches an abnormal value or the like, the vehicle control device 100 acquires the control parameter of the other vehicle from the server 16 and uses this as the initial value. Thereby, after initializing the control parameter 150, it is possible to quickly reach a state equivalent to the learned state.

<Modification of the present invention>
The present invention is not limited to the above embodiment, and includes various modifications. For example, the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to one having all the configurations described.

  3 to 4, the learning value map of the engine ignition timing is illustrated as an example of the control parameter 150. However, the learned parameter of the other vehicle is acquired from the server 16 in the same manner even for other control parameters and array elements. be able to.

  Communication means between the server 16 and each vehicle is not necessarily limited to wireless communication, and may be wired communication. For example, when transmitting and receiving control parameters during vehicle maintenance, it is conceivable that the vehicle communicates with the server 16 via a wired connection.

  In the above description, it has been described that when the control parameter 150 is initialized, the control parameter of the other vehicle is acquired from the server 16. However, when a situation equivalent to the initialization occurs, the server 16 similarly controls the other vehicle. Parameters may be acquired. For example, immediately after replacing parts (e.g., exhaust catalyst) of the vehicle 200, it is considered that the control parameter changes discontinuously. Therefore, it is assumed that this is a situation similar to the initialization, and processing similar to the present invention can be performed. Conceivable. As a means for the vehicle control apparatus 100 to recognize whether or not a part has been replaced, for example, it can be considered to give a signal to that effect from the outside of the vehicle control apparatus 100 during maintenance.

  16: server, 100: vehicle control device, 110: CPU, 120: control unit, 130: communication unit, 140: storage device, 150: control parameter, 200: vehicle.

Claims (7)

A vehicle control device for controlling the operation of a vehicle,
A storage unit for storing a first control parameter unique to the vehicle;
A control unit for controlling the vehicle using the first control parameter and updating the first control parameter and overwriting the storage unit;
A communication unit that communicates with a server that aggregates second control parameters unique to other vehicles to obtain the second control parameters;
With
The control unit determines whether the first control parameter has reached an abnormal value;
When the control unit determines that the first control parameter has reached an abnormal value, the control unit initializes the first control parameter and acquires the second control parameter from the server via the communication unit. The vehicle control device, wherein the second control parameter is stored in the storage unit as a new initial value of the first control parameter. The first control parameter and the second control parameter are configured as array data having one or more dimensions,
The control unit stores, in the storage unit, the second control parameter acquired through the communication unit as a new initial value for an array element of the first control parameter that has not yet been updated from the initial value. The vehicle control device according to claim 1. The control unit stores, in the storage unit, the second control parameter acquired via the communication unit as a new initial value for one of the array elements of the first control parameter that has been updated from the initial value. The vehicle control device according to claim 2. The communication unit is
Vehicle type or model number of the vehicle,
The distance traveled by the vehicle,
The number of years since the vehicle was shipped,
Of the types of optional parts mounted on the vehicle, those that affect the running performance of the vehicle,
A range of array elements of the second control parameter acquired from the server;
The vehicle according to claim 2, wherein the second control parameter corresponding to the key is acquired by inquiring the second control parameter corresponding to the server using at least one of the key as a key. Control device. The vehicle control device according to any one of claims 1 to 4,
The server for aggregating the second control parameters;
A second vehicle for transmitting the second control parameter to the server;
A vehicle control parameter learning system comprising: The second vehicle, when an ignition switch of the second vehicle is turned off, transmits the second control parameter held by the second vehicle at that time to the server. Item 6. The vehicle control parameter learning system according to Item 5. The vehicle control parameter learning system according to claim 5, wherein the second vehicle transmits the updated second control parameter to the server every time the second control parameter is updated.

Images