JP2014043791A - Vehicle control system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a vehicle control system capable of calculating learning values, such as an amount of correction required for controlling a control target device mounted on a vehicle, without performing a special control process for learning.SOLUTION: In the vehicle control system, learning values are calculated not in a vehicle-side control device 200 but in a data center 100, on the basis of information sent from the vehicle-side control device 200. Accordingly, the vehicle-side control device 200 does not need to perform any special control process for calculating the learning values, and can perform ordinary control for the control target device. In the data center 100, the learning values, such as an amount of correction and a control gain, are calculated using a computer with higher performance compared with the vehicle-side control device 200.

Description

  The present invention relates to a vehicle control system that includes a vehicle-side control device and a data center that can communicate with each other, and controls a device to be controlled mounted on the vehicle.

  For example, Patent Document 1 discloses a fuel injection control device for a diesel engine as an example of vehicle control. In this fuel injection control device, single injection is performed at the time of non-injection when the command injection amount to the injector is less than or equal to zero, the amount of change in engine speed (rotation speed increase δx) that increases due to the single injection, and single injection The product of the engine rotational speed ω0 at the time of execution is calculated as the torque proportional amount Tp. In a diesel engine, since the injection amount and the generated torque are proportional, the actual injection amount is estimated from the generated torque calculated from the torque proportional amount Tp. Then, the correction amount of the injection amount is learned by detecting the difference between the estimated actual injection amount and the command injection amount as an injection amount deviation.

JP 2005-36788 A

  In the above-described apparatus, the control process for learning the correction amount can be executed only when a predetermined learning condition (when no command injection amount is zero) is satisfied. For this reason, it may be considered that an excessive time is required until an optimum learning value is obtained. Further, in the control process for learning the correction amount, there is a problem in that the fuel consumption is increased and the engine speed is fluctuated because a certain amount of fuel is actually injected. Further, when the ECU is initialized by replacing the ECU or the like, new learning must be performed. Until the learning is completed, the engine cannot be operated in an optimum state, and emission and the like are deteriorated. There is a risk of inviting.

  The present invention has been made in view of the above-described points, and the learning value such as a correction amount necessary for controlling the control target device mounted on the vehicle can be obtained without performing a special control process for learning. An object of the present invention is to provide a vehicle control system capable of calculating.

To achieve the above object, a vehicle control system according to the present invention comprises a vehicle-side control device (200) and a data center (100) that can communicate with each other,
The vehicle side control device
A sensor unit (14, 23, 24, 28, 360) for detecting information on the operation state of the control target device;
A control unit (220, 330) that calculates a control amount using an arithmetic expression including a learning value determined by learning control based on information detected by the sensor unit, and controls the control target device according to the calculated control amount; With
The vehicle-side control device periodically transmits information necessary for calculating a learning value to the data center,
The data center includes a learning control unit (110) that performs learning control based on information periodically transmitted from the vehicle-side control device, and calculates a learning value suitable for a control target device mounted on the vehicle. ,
When the learning control unit calculates the learning value, the data sensor transmits the learning value to the vehicle-side control device of the corresponding vehicle. When the vehicle-side control device receives the learning value from the data center, the data sensor uses the received learning value. Then, the control amount is calculated.

  As described above, in the vehicle control system according to the present invention, the learning value is calculated based on the information detected by the sensor unit in the data center, not in the vehicle-side control device. For this reason, in order to calculate a learning value, the vehicle side control apparatus does not need to perform any special control process, and can perform normal control with respect to a control object apparatus. Then, in the data center, using a high-performance computer compared to the vehicle-side control device, for example, based on a transfer function obtained by system identification of the device to be controlled, correction amount, control gain, etc. It is also possible to calculate a learning value. Thereby, compared with the case where a learning value is calculated in the vehicle side control apparatus like the past, it becomes possible to calculate a learning value with higher accuracy.

  Note that the reference numerals in the parentheses merely show an example of a correspondence relationship with a specific configuration in an embodiment described later in order to facilitate understanding of the present invention, and limit the scope of the present invention. It is not intended.

  Further, the features of the present invention other than the features described above will be apparent from the description of embodiments and the accompanying drawings described later.

1 is a diagram illustrating an overall configuration of a vehicle control system according to an embodiment. It is a figure which shows an example of the specific structure of the engine which engine control ECU controls. It is a flowchart which shows the process which engine control ECU performs. It is a flowchart which shows the process performed in a data center. It is a figure which shows the relationship between fuel injection quantity and an engine speed. FIG. 6 is a signal flow diagram of the system shown in FIG. 5. It is the figure which showed the structure for controlling an electric motor in the vehicle side control system of 2nd Embodiment. It is the figure which showed motor control ECU as a functional block. It is the figure which showed the PWM pattern in the relationship between the rotation speed of an electric motor, and a torque. It is a figure which shows the table of the control gain of P term and I term. It is a flowchart which shows the learning process of the control gain of P term and I term performed in a vehicle side control apparatus. It is a figure for demonstrating a slope ratio. It is a figure which shows the example of a division of a slope ratio. It is a figure for demonstrating a slope gradient. It is a figure which shows the example of a division of a slope gradient. It is a figure which shows the management table which manages the information regarding the drive state of an electric motor corresponding to the traveling pattern divided. It is a flowchart which shows the process performed in a data center. It is a flowchart which shows the process performed in a vehicle side control apparatus.

Hereinafter, a vehicle control system according to an embodiment of the present invention will be described in detail with reference to the drawings.
(First embodiment)
The vehicle control system according to the first embodiment is embodied as an engine mounted on a vehicle, which is a control target device, and controls the operating state of the engine. As shown in FIG. 1, the vehicle control system according to the present embodiment mainly includes a data center 100 and a vehicle-side control device 200.

  The data center 100 performs a transmission / reception process with the communication module 210 of each vehicle or calculates a learning correction amount δ as a learning value by executing a learning process described later, and is calculated for each vehicle. And a database 120 that stores the learned correction amount δ in association with information (vehicle specifying information) for specifying each vehicle. Note that as the vehicle identification information, in addition to information unique to each vehicle such as a chassis number and a registration number, information such as an engine type and a travel distance of each vehicle is used. The vehicle specifying information is transmitted from the vehicle side control device 200 to the data center 100 together with sensor information for learning control.

  Thus, by storing the learning correction amount δ for each vehicle in the database 120 of the data center 100, for example, even when the engine control ECU 220 is replaced in the corresponding vehicle, The engine control ECU 220 can immediately acquire the learning correction amount δ from the data center 100. In addition, for a vehicle that has newly started communication with the data center 100, even before the calculation of the learning correction amount δ is completed in the server 110, based on the vehicle identification information, for other vehicles equipped with the same type engine Can be transmitted as a default value.

  The learning correction amount δ as a default value may be calculated by averaging learning correction values δ for a plurality of other vehicles equipped with the same type of engine. Further, in addition to the condition of the engine of the same type, the learning correction amount δ for other vehicles used as the default value may be further narrowed down under the condition that the travel distance is approximate. In this way, narrowing down using a plurality of conditions makes it easier to select other vehicles of the same type that use a more approximate learning correction amount δ.

  The vehicle-side control device 200 includes a communication module 210 for communicating with the data center 100 and various control ECUs 220 to 240 including the engine control ECU 220 connected to the communication module 210 via a LAN. Therefore, each control ECU220-240 can communicate with the data center 100 via the communication module 210, and can also communicate between ECUs.

  Next, an example of a specific configuration of an engine to be controlled by the engine control ECU 220 will be described with reference to FIG. In FIG. 2, an air cleaner 13 is provided at the most upstream portion of the intake pipe 12 of the direct injection engine 11 which is an internal combustion engine of the direct injection type, and an intake air amount is detected downstream of the air cleaner 13. An air flow meter 14 is provided. A throttle valve 16 whose opening is adjusted by a motor 15 and a throttle opening sensor 17 for detecting the opening of the throttle valve 16 are provided on the downstream side of the air flow meter 14.

  A surge tank 18 is provided downstream of the throttle valve 16, and an intake pipe pressure sensor 19 for detecting the intake pipe pressure is provided in the surge tank 18. The surge tank 18 is provided with an intake manifold 20 that introduces air into each cylinder of the engine 11, and controls the in-cylinder airflow strength (swirl flow strength and tumble flow strength) in the intake manifold 20 of each cylinder. An airflow control valve 31 is provided.

  An injector 21 that directly injects fuel into the cylinder (combustion chamber) is attached to the upper part of each cylinder of the engine 11. A spark plug 22 is attached to the cylinder head of the engine 11 for each cylinder, and the air-fuel mixture in the cylinder is ignited by spark discharge of each spark plug 22. Further, the intake valve 37 and the exhaust valve 38 of the engine 11 are provided with variable valve timing devices 39 and 40 for varying the opening / closing timing, respectively.

  A knock sensor 32 for detecting knocking and a cooling water temperature sensor 23 for detecting cooling water temperature are attached to the cylinder block of the engine 11. A crank angle sensor 24 that outputs a pulse signal every time the crankshaft rotates a predetermined crank angle is attached to the outer peripheral side of the crankshaft (not shown). Based on the output signal of the crank angle sensor 24, the crank angle and the engine speed are detected.

The exhaust pipe 25 of the engine 11 is provided with an upstream catalyst 26 and a downstream catalyst 27 for purifying the exhaust gas, and an exhaust gas sensor 28 (for detecting the air-fuel ratio of the exhaust gas upstream of the upstream catalyst 26). O ₂ sensor or the like) is provided. In the present embodiment, a three-way catalyst for purifying CO, HC, NOx, etc. in the exhaust gas near the stoichiometric air-fuel ratio is provided as the upstream side catalyst 26, and a NOx occlusion reduction type catalyst is provided as the downstream side catalyst 27. .

  An EGR pipe 33 for returning a part of the exhaust gas to the intake side is connected between the downstream side of the upstream catalyst 26 in the exhaust pipe 25 and the surge tank 18 of the intake pipe 12. Is provided with an EGR valve 34 for controlling the exhaust gas recirculation amount (EGR amount). Further, the accelerator sensor 36 detects the amount of depression of the accelerator pedal 35 (accelerator opening).

  Outputs of the various sensors described above are input to the engine control ECU 220. This engine control ECU 220 is mainly composed of a microcomputer, and executes various engine control programs stored in a built-in ROM, thereby determining the fuel injection amount and fuel injection timing of the injector 21 according to the engine operating state. Various engines such as fuel injection control for controlling, ignition timing control for controlling the ignition timing of the spark plug 22, EGR control for controlling the exhaust gas recirculation amount, and VVT control for controlling the opening / closing timing of the intake valve 37 and the exhaust valve 38 Execute control.

  Here, the various controls described above are executed by driving mechanical and electrical components such as the injector 21, the ignition coil 22, the EGR valve 34, the intake valve 37, and the exhaust valve 38. For these mechanical and electrical parts, for example, the control characteristics change (aging) over time. For this reason, in the various engine controls described above, a control method called learning control is generally employed to compensate for changes in control characteristics due to aging of components used for control.

  Conventionally, this learning control is also performed by an engine control ECU mounted on the vehicle, and thus various problems have occurred. For example, as described in the column of the prior art, learning control can be performed only when a predetermined learning condition is satisfied, and it takes an excessive amount of time until an optimum learning value is obtained or the learning is performed. For control, an increase in fuel consumption and fluctuations in the engine speed were caused, and when the ECU was initialized by replacing the ECU or the like, it was necessary to newly learn.

  Therefore, in this embodiment, learning control is not performed by the engine control ECU 220 mounted on the vehicle, but is performed by the data center 100 (the server 110). Thus, for the purpose of calculating the learning correction amount δ, the vehicle-side control device 200 does not need to execute any special control process, and can perform normal control on the engine. In the data center 100, a transmission obtained by identifying a system relating to, for example, an engine and / or mechanical and electrical parts using a computer having a higher performance than the ECU of the vehicle-side control device 200. A learning value such as a correction amount can be calculated based on the function (model). Thereby, compared with the case where the learning value is calculated by the vehicle-side control device 200 as in the prior art, it becomes possible to calculate the learning value with higher accuracy in a shorter time.

  Below, the process performed by the engine control ECU 220 and the process performed by the data center 100 will be described by taking fuel injection control performed by the engine control ECU 220 as an example.

  FIG. 3 is a flowchart showing a process executed by engine control ECU 220. The process shown in this flowchart is repeatedly executed, for example, at intervals of several tens of ms.

First, in step 100, detection signals output from various sensors are captured. Specifically, a detection signal indicating the engine speed Ne from the crank angle sensor 24, a detection signal indicating the intake air amount Qa from the air flow meter 14, a detection signal indicating the air-fuel ratio λ from the O ₂ sensor, and cooling from the cooling water temperature sensor 23 Each detection signal indicating the water temperature is captured.

In the subsequent step S110, the captured engine rotation speed Ne, intake air amount Qa, O ₂ sensor output λ, and cooling water temperature T are transmitted to the data center 100. However, the engine control ECU 220 does not transmit the detection result of the sensor to the data center 100 every time, but periodically transmits the detection result of the sensor to the data center 100 at intervals of several seconds, for example. This is because the learning correction amount δ does not change in units of time such as several tens of ms, and is intended to prevent the communication load of the engine control ECU 220 from becoming excessive.

  In step S120, the basic fuel injection amount BQf is calculated as a function of the engine speed Ne and the intake air amount Qa using a predetermined arithmetic expression, a map, or the like. At this time, the basic fuel injection amount BQf is calculated so as to obtain the theoretical air-fuel ratio. In the subsequent step S130, the correction coefficient K (T) is calculated using the cooling water temperature T as an argument. In step S140, the fuel injection amount Qf is calculated using the basic fuel injection amount BQf, the correction coefficient K (T), and the increase / decrease correction value λPI (t) determined in time series.

The O ₂ sensor output λ outputs a binary value when the air-fuel ratio is richer than the stoichiometric air-fuel ratio and when it is lean. When the O ₂ sensor output λ indicates that it is richer than the stoichiometric air-fuel ratio, the formula b is used so that the increase / decrease correction value λPI (t) decreases the fuel injection amount Qf. On the other hand, when the O ₂ sensor output λ indicates that it is lean, the expression a is used so that the increase / decrease correction value λPI (t) increases the fuel injection amount Qf. The increase / decrease correction value λPI (t) is determined such that the absolute value thereof decreases as time elapses. For example, from time t1 to time tn, λPI (t1) = BQf, λPI (t2) = 0.9BQf, λPI (t3) = 0.8BQf,..., ΛPI (tn) = 0.4BQf, The value is set to decrease.

  In step S150, the fuel injection amount Qf calculated in step S140 is transmitted to the data center 100. The transmission of the fuel injection amount Qf is performed in synchronization with the transmission of the detection result by the sensor in step S110 described above or collectively with the detection result by the sensor.

  In step S160, when the learning process in the data center 100 is completed and the learning correction amount δ of the fuel injection amount is transmitted from the data center 100, the learning correction amount δ is received. As described above, when communication with a certain vehicle is started, the data center 100 is equipped with the same type of engine until calculation of the learning correction amount δ suitable for fuel control of the vehicle is completed. The learning correction amount δ stored for the vehicle is transmitted as a default value. Thus, before the learning correction amount δ based on the detection result of the sensor and the calculated fuel injection amount Qf is calculated in the data center 100, the engine control ECU 220 uses the learning correction amount δ close to the appropriate value to perform fuel injection. The quantity Qf can be calculated. Then, after the calculation of the learning correction amount δ is completed, the data center 100 transmits the learning correction amount δ to the vehicle-side control device 200. This transmission is repeated every time the learning correction amount δ is updated.

  In step S170, the fuel injection amount Qf is corrected using the learning correction amount δ transmitted from the data center. The engine control ECU 220 corrects the fuel injection amount Qf using the latest learning correction amount δ received in step S160 until a new learning control amount δ is received from the data center 100.

  Next, processing executed in the data center 100 will be described based on the flowchart of FIG. First, in step S200, the detection result of the sensor and the calculated fuel injection amount Qf are received from the vehicle-side control device 200. In subsequent step S210, learning control is executed based on the sensor detection result and the calculated fuel injection amount Qf received in step S200, and learning for correcting a deviation due to a secular change or the like of the fuel injection amount in the injector 21 is performed. A correction amount δ is calculated.

  For example, this learning control is performed by executing arithmetic processing according to a learning algorithm using artificial intelligence or a neural network, or by executing a kind of simulation called so-called system identification, thereby transferring a transfer function representing the system. It is carried out by asking. The calculation processing for performing such learning control is enormous, but in this embodiment, learning control is performed by the server 110 that is a high-performance computer installed in the data center 100. Therefore, the learning correction amount δ can be obtained without imposing a load on the vehicle-side control device 200.

  Hereinafter, as an example of learning control, a method for obtaining a transfer function (model) representing the system by system identification and calculating the learning correction amount δ using the obtained transfer function will be described. In the following example, system identification is performed using the least square method.

  When the engine is operated in a predetermined state, the engine speed changes due to the change in the fuel injection amount. Therefore, as shown in FIG. 5, the relationship between the fuel injection amount and the engine rotation speed can be expressed using a transfer function G1 (z) having the fuel injection amount as an input and the engine rotation speed as an output.

  Then, data of the fuel injection amount change ΔQf and the engine speed change ΔNe are sampled N times. This is expressed as an input data sequence {u (i)} = {ΔQf} and an output data sequence {y (i)} = {ΔNe} (where Ii = 1, 2, 3,..., N).

As shown in FIG. 5, since the system has one input and one output, the transfer function G1 (z) of this system can be expressed as Equation 1 below.
(Equation 1) G1 (z) = B (Z ⁻¹ ) / A (Z ⁻¹ )
When this is modified, the following formula 2 is obtained.
(Expression 2) G1 (z) = (b0 + b1 · Z ⁻¹ +... + Bn · Z ⁻ⁿ ) /
(1 + a1 · Z ⁻¹ + a2 · Z ⁻² +... + An · Z ⁻ⁿ )
Z ⁻¹ is a unit transition operator and means Z ⁻¹ · x (k) = x (k−1).

If the parameters a1 to an and b0 to bn in Expression 2 are determined from the input / output data series {u (i)} and {y (i)}, the system transfer function G1 (z) is obtained. In system identification by the method of least squares, the parameters a1 to an and b0 to bn are determined so that the evaluation function J of Equation 3 below is minimized.

  In the present embodiment, each parameter is obtained with n = 2. In addition, when each parameter in Formula 2 is obtained based on the fuel injection amount Qa and the engine speed Ne transmitted from the vehicle side, the obtained transfer function G1 (z) is the fuel injection of the engine at that time. It represents the state of the system. This transfer function G1 (z) is stored in the database 120 of the data center 100, and is used thereafter to calculate the learning correction amount δ. Hereinafter, a method for calculating the learning correction amount δ using the transfer function G1 (z) will be described.

The signal flow diagram of the system shown in FIG. 5 is as shown in FIG. 6. When [x1 (k), x2 (k)] ^T is taken as the state variable quantity, the state / output equation is expressed by the following Equation 4. Can be expressed as:

  This means that if the fuel injection amount change ΔQf is input to the input u (k) and Equation 4 is sequentially calculated, the engine rotation speed change ΔNe is calculated as y (k). To do.

In actual control, the calculated change amount ΔNe of the engine speed is set as дNe, and the correction amount δ of the fuel injection amount Qf is set so that the difference from the actually observed change amount ΔNe of the engine speed becomes zero. Just ask. That is, ΔQf is obtained so that | ΔNe (k) −дNe (k) | → 0, and this may be used as the learning correction amount δ (k). More specifically, the learning correction amount δ (k) is obtained from the following formulas 5 and 6.
(Expression 5) (ΔNe (k) −дNe (k)) = G1 (z) · δ (k)
(Equation 6) δ (k) = G1 (z) ⁻¹ · (ΔNe (k) −дNe (k))
In step S220, it is determined whether or not the calculation of the learning correction amount δ has been completed by the learning control in step S210. If it is determined that the calculation of the learning correction amount δ is complete, the process proceeds to step S230, and the calculated learning correction amount δ is transmitted to the vehicle-side control device. Thereafter, the calculated learning correction amount δ is stored in the database 120 in association with vehicle specifying information for specifying the vehicle.

On the other hand, when it is determined that the calculation of the learning correction amount δ is not completed, the process proceeds to step S240, and the learning correction amount δ stored for another vehicle equipped with the same type engine is transmitted as a default value. .
(Second Embodiment)
Next, a vehicle control system according to a second embodiment of the present invention will be described. The vehicle control system according to the present embodiment is embodied as a device to be controlled using an electric motor as a drive source mounted on a vehicle and controlling the operating state of the electric motor.

  Similarly to the vehicle control system of the first embodiment, the vehicle control system according to this embodiment includes a vehicle-side control device and a data center. FIG. 7 is a diagram illustrating a configuration for controlling the electric motor in the vehicle-side control system.

  As shown in FIG. 6, the vehicle-side control system has a battery 310 that is charged with the generated power when supplying electric power to the electric motor 350 or when the electric motor 350 operates as a generator during regenerative braking. doing.

  Battery control ECU 320 calculates the remaining capacity of battery 310 and notifies motor control ECU 330 that controls the driving state of the electric motor. Based on the notified remaining capacity, the motor control ECU 330 provides the vehicle occupant with information indicating the remaining battery level and the travelable distance, or in a hybrid vehicle, determines the ratio between the engine output and the electric motor output. To do.

  Further, the battery control ECU 320 monitors the state of the battery 310 and, when an abnormality occurs, executes a measure for dealing with the abnormal state. For example, when the temperature of the battery 310 rises above a predetermined temperature, the battery control ECU 320 drives a fan (not shown) to reduce the temperature of the battery 310, or suppresses further temperature rise. A power usage restriction command is transmitted to the motor control ECU 330 so as to limit the amount of power supplied by the battery 310 to a predetermined power or less.

  The inverter 340 is controlled by the motor control ECU 330 and performs a DC-AC conversion between the battery 310 and the electric motor 350 to generate a rotating magnetic field in the electric motor or charge the battery with the generated electric power of the electric motor. Or convert it to possible DC power.

  The motor control ECU 330 controls the driving state of the electric motor 350. The motor control ECU 330 will be described in detail with reference to the functional block diagram of FIG.

  The motor control ECU 330 receives a torque command based on the driver's driving operation or the like, and converts the torque command into a current command using a predefined map. This current command is compared with a detection current detected by the current sensor 360 and actually flowing through the electric motor 350. The motor control ECU 330 executes vector control for determining an energization mode (switching mode of the inverter 340) as to which phase of the electric motor 350 is energized in order to bring the detected current close to the current command.

  Further, the motor control ECU 330 determines the PWM pattern (sine wave PWM in the sine wave region, over modulation PWM in the over modulation region, rectangular wave region in accordance with the modulation degree from the relationship between the rotational speed and torque of the electric motor shown in FIG. And the motor voltage determined by the PWM pattern is output to the PWM modulator. At this time, in this embodiment, PI control is adopted as a control law, and control gains (feedback gains) of the proportional term P and the integral term I are determined by adjustment, and the actual motor voltage is obtained.

  Here, in the PI control law in motor control, there is a trade-off relationship that if the P term is increased, the response is improved, but the efficiency is lowered. Therefore, the control gains of the P term and the I term are adjusted. The control gains of the P term and the I term are adjusted to a traveling pattern (for example, 20% on a slope with a slope, and then 60% on a flat road, Thereafter, the optimum value varies depending on, for example, 20% of the downhill road and regenerative power is stored in the battery on the downhill. Considering such various traveling patterns, man-hours for determining the control gains of the P term and the I term of the PI control law become enormous, and therefore a moderate control gain should be used assuming a standard traveling pattern. Did not get.

  However, after actually traveling in a certain traveling pattern, the control gains of the P term and the I term can be updated in the vehicle-side control device so as to suit the traveling pattern. For example, as shown in FIG. 10, in order to facilitate the setting of the control gains for the P term and the I term, the control gain I (k) for the I term is linked to the control gain P (k) for the P term (for example, , I term = P term / 10), and combinations of some P terms and I-term control gains P (k) and I (k) are prepared as a table. And then. By executing the processing shown in the flowchart of FIG. 11, the control gains P (k) and I (k) of the P term and the I term are updated so as to suit the travel pattern actually traveled.

  The process shown in the flowchart of FIG. 11 will be described below. First, in step S300, it is determined whether or not the vehicle has traveled for a certain time or a certain distance. If “YES” is determined in this determination processing, the process proceeds to step S310, and it is determined whether or not the motor drive efficiency is smaller than a predetermined value (for example, 50%). The motor drive efficiency is calculated based on the ratio of torque generated with respect to the applied voltage.

  If it is determined in step S310 that the motor drive efficiency is greater than the predetermined value, the process proceeds to step S320, and the control gain P (k in the table is increased in order to increase the control gain P (k) of the P term and improve the responsiveness. ), A variable k for designating a combination of I (k) is incremented. Thereby, a combination of control gains P (k) and I (k) in which the control gain P (k) of the P term is larger by one step is designated by the variable k. However, since combinations of control gays P (k) and I (k) are prepared only up to k = n, it is determined in step S330 whether or not the variable k is n or more. At this time, if it is determined that the variable k is greater than or equal to n, the process proceeds to step S350, where the variable k is set to the maximum value n. On the other hand, when it is determined that the variable k is less than n, the process proceeds to step S340, and the variable k is set to a value as it is.

  In step S360, the control gains P (k) and I (k) designated by the set variable k are set as control gains for the P and I terms.

  On the other hand, if it is determined in step S310 that the motor drive efficiency is smaller than the predetermined value, the process proceeds to step S370 to reduce the control gain P (k) of the P term to reduce the responsiveness. Decrement k. Thus, a combination of control gains P (k) and I (k) in which the control gain P (k) of the P term is one step smaller is designated by the variable k. However, combinations of control gays P (k) and I (k) are prepared only up to k = 0. For this reason, in step S380, it is determined whether the variable k is 0 or less. At this time, if it is determined that the variable k is 0 or less, the process proceeds to step S400, and the variable k is set to 0, which is the minimum value. On the other hand, if it is determined that the variable k is greater than 0, the process proceeds to step S390, where the variable k is set to a value as it is.

  In step S410, the control gains P (k) and I (k) designated by the set variable k are set as control gains for the P and I terms.

  However, the learning control of the control gains P (k) and I (k) for the P term and I term as described above is simple, and is not necessarily limited to the control gain P (k suitable for the actual traveling pattern of the vehicle. ), I (k) cannot always be set. Therefore, in the present embodiment, information on travel patterns and motor drive efficiencies in each vehicle is collected in the data center 100, and control gains P (k), I ( k) is saved. Based on the information transmitted from each vehicle, the control gain P (k) inferior to the control gains P (k) and I (k) stored in the data center 100 in the traveling pattern of the corresponding vehicle, When it is considered that I (k) is used, the control gains P (k) and I (k) stored in the data center 100 are transmitted to the corresponding vehicle. On the vehicle side, the motor control ECU 330 controls the electric motor 350 using the received control gains P (k) and I (k). This makes it possible to control the electric motor 350 mounted on each vehicle in a more efficient state without sacrificing responsiveness.

  Hereinafter, processing in the data center 100 and processing in the vehicle-side control device 200 for executing the above-described control will be described with reference to FIGS.

  First, in the data center 100, the traveling pattern of the vehicle is classified according to the slope ratio and slope slope. When the electric motor is controlled by the PI control law, the output of the electric motor changes depending on the slope ratio and slope on the vehicle traveling path, so that optimum control gains P (k) and I (k) for the P term and I term This is because they are also different. The downhill is not included in the travel pattern classification conditions because it is not necessary to drive the motor.

  In this embodiment, the slope ratios shown in FIG. 12 are classified into 10 sections as shown in FIG. Further, the slopes shown in FIG. 14 are classified into three categories as shown in FIG. Therefore, the traveling pattern is classified into 30 sections by combining the slope ratio and the slope.

  As shown in FIG. 16, the database of the data center 100 is provided with a management table for managing information relating to the driving state of the electric motor including at least the control gain and the motor driving efficiency corresponding to the 30 sections. This management table is provided for each type of electric motor.

  Then, the server 110 of the data center 100 executes the process shown in the flowchart of FIG. First, in step S500, a PWM pattern, control gain, slope ratio, slope gradient, SOC, and motor drive efficiency are received from the vehicle as information regarding the drive state of the electric motor.

  In the subsequent step S510, the evaluation function Jn for comprehensively evaluating the motor drive efficiency and the responsiveness is calculated by setting the motor drive efficiency of the received data as un and the control gain of the P term as gn. The evaluation function Jn is calculated by, for example, Jn = Q · un + R · gn (Q and R are arbitrary weighting factors).

  In step S520, in the management table of the database 120, when data is already stored in the travel pattern classification corresponding to the slope ratio and slope of the received data, the motor drive efficiency and control of the P term in the data are stored. An evaluation function J is calculated from the gain.

  In step S530, the evaluation functions Jn and J calculated in steps S510 and S520 are compared in size. At this time, when the evaluation function Jn calculated from the received data is determined to be equal to or higher than the evaluation function J calculated from the stored data, the motor control state based on the newly received data is stored. It can be considered that it is generally better than the control state of the motor based on the data. Therefore, it progresses to step S540 and the classification of the applicable travel pattern of a management table is updated with the received data.

  On the other hand, if it is determined in step S530 that the evaluation function Jn is smaller than the evaluation function J, the motor control state based on the stored data is greater than the motor control state based on the newly received data. It can be regarded as excellent overall. Therefore, it progresses to step S550 and the control gain preserve | saved at the classification of the applicable driving | running | working pattern of a management table is transmitted with respect to the vehicle which transmitted data.

  Next, the process in the vehicle side control apparatus 200 is demonstrated, referring the flowchart of FIG.

  First, in step S600, a torque command based on the driver's driving operation or the like is received, and a command current corresponding to the torque command is derived from a predefined map. In the subsequent step S610, the difference between the command current and the detected current is input to the vector control unit to determine the switching mode of the inverter 340 and calculate the P and I terms in the PI control law.

  In step S620, when the vehicle has traveled for a certain time or a certain distance and the travel pattern is determined, the driving state of the electric motor such as the slope ratio, slope gradient, PWM pattern, control gain, SOC, and motor drive efficiency is related. Information is transmitted to the data center 100. Note that when the running pattern is not fixed, the process of step S620 is skipped.

In step S630, it is determined whether or not a recommended control gain has been obtained from the data center 100. When the recommended control gain is obtained, the process proceeds to step S640, and the recommended control gain for the P term and the I term is set to the control gain for calculating the motor voltage. Thereby, the control gain is learned. In step S660, the motor voltage Vm is calculated using the set control gain. The motor voltage Vm is calculated according to Equation 7 below.

  On the other hand, if it is determined in step S630 that the recommended control gain is not obtained, the process proceeds to step S650, and the motor voltage Vm is calculated using the P term and I term calculated in step S610.

  Finally, in step S670, the calculated motor voltage Vm is output to the PWM modulator.

  As mentioned above, although preferable embodiment of this invention was described, this invention is not restrict | limited to embodiment mentioned above, In the range which does not deviate from the main point of this invention, it can change and implement variously.

  For example, in the above-described embodiment, the example in which the present invention is applied as a vehicle control system that performs engine control and motor control has been described. However, the vehicle control system according to the present invention is not limited to an engine or a motor, It is applicable if it is installed and performs learning control.

  In the above-described second embodiment, the example in which the electric motor is PI-controlled has been described. However, the control law includes the P term (proportional term), the I term (integral term), and the D term (differential term). As long as it is executed using at least two terms.

100 data center 110 server 120 database 200 vehicle side control device 210 communication module 220 engine control ECU

Claims (9)

A vehicle control system for controlling a control target device mounted on a vehicle,
The vehicle control system includes a vehicle-side control device (200) and a data center (100) that can communicate with each other.
The vehicle-side control device is
A sensor unit (14, 23, 24, 28, 360) for detecting information related to the operation state of the control target device;
Control units (220, 330) for calculating a control amount using an arithmetic expression including a learning value determined by learning control based on information detected by the sensor unit, and controlling the device to be controlled according to the calculated control amount And comprising
The vehicle-side control device periodically transmits information necessary for calculating a learning value to the data center,
The data center executes learning control based on information periodically transmitted from the vehicle-side control device, and calculates a learning value suitable for a control target device mounted on the corresponding vehicle (110). With
When the learning control unit calculates the learning value, the data sensor transmits the learning value to the vehicle-side control device of the corresponding vehicle, and when the vehicle-side control device receives the learning value from the data center, the received learning A vehicle control system characterized in that the control amount is calculated using a value. The vehicle-side control device also transmits vehicle specifying information for specifying a vehicle to the data center,
Before the calculation of the learning value suitable for the control target device mounted on the relevant vehicle is completed, the data center has already been calculated for the control target device of the same type of other vehicle based on the vehicle specifying information. 2. The vehicle control system according to claim 1, wherein the learning value is selected and transmitted as a default value.   The learning control unit models the device to be controlled based on the information detected by the sensor unit, and uses the model to shift the control result to be obtained for the control amount from the actual control result. The vehicle control system according to claim 1, wherein a correction amount for compensating for is calculated as a learning value.   The vehicle control system according to any one of claims 1 to 3, wherein the vehicle-side control device uses an engine mounted on the vehicle as a control target device and controls an operating state of the engine.   The vehicle control system according to claim 4, wherein the vehicle-side control device controls a fuel injection amount by an injector in order to control an operation state of the engine. The vehicle-side control device uses an electric motor as a travel drive source as a control target device, and feedback-controls the energization state of the motor so that a command current corresponding to a command torque is energized to the electric motor. ,
The vehicle-side control device transmits information related to the traveling path of the vehicle and information related to the driving state of the electric motor when traveling on the traveling path to the data center,
The data center classifies information on the travel road, and for each classification, at least information on a driving state of an electric motor that is evaluated to be most desirable according to a predetermined evaluation standard from information obtained from a plurality of vehicles. It has a stored database, and when a new information on the traveling path of the vehicle and information on the driving state of the electric motor when traveling on the traveling path is transmitted from a certain vehicle, it is stored in the corresponding classification. It is determined which of the information related to the driving state of the electric motor that is transmitted and the information related to the transmitted driving state is superior in the above evaluation criteria. In addition, the feed for obtaining the driving state as the learning value based on the stored information on the driving state. Tsu sends click control gain, when the information is better regarding the driving state has been transmitted, updates the information about the driving state of storing,
2. The vehicle control system according to claim 1, wherein when the feedback control gain is received from the data center, the vehicle-side control device executes the feedback control using the feedback control gain.   The vehicle control system according to claim 6, wherein the vehicle-side control device performs control to change the feedback control gain before receiving the feedback control gain from the data center.   The feedback control is performed using at least two of a proportional term, an integral term, and a derivative term, and the feedback control gain is a combination of feedback control gains of each item used for the feedback control. The vehicle control system according to claim 6 or 7, characterized by comprising:   The vehicle control system according to claim 6, wherein the data center includes the database for each type of the electric motor.

Images