KR20240035098A - Apparatus for controlling clutch actuator of double clutch transmission and method therof - Google Patents

Abstract

The present invention discloses a clutch actuator control device and method for a double clutch transmission. The clutch actuator control device of the double clutch transmission of the present invention includes a clutch actuator; Memory; and a processor operatively coupled to the clutch actuator and memory; wherein the processor controls each gain for PID control of the artificial neural network structure for errors caused by the difference between the target position and the actual position of the clutch actuator. It is characterized by separating error components, setting parameters for each term, applying them to the error components of each PID term, and then outputting the control command calculated through the activation function to the clutch actuator.

Description

Clutch actuator control device for double clutch transmission and method thereof {APPARATUS FOR CONTROLLING CLUTCH ACTUATOR OF DOUBLE CLUTCH TRANSMISSION AND METHOD THEROF}

The present invention relates to a clutch actuator control device and method for a double clutch transmission, and more specifically, to control the position by reflecting the characteristics of the clutch actuator in a double clutch transmission, by controlling the error component due to the difference between the actual position and the target position. It relates to a clutch actuator control device and method for a double clutch transmission that automatically determines the gain ratio through intelligent PID control in the form of an artificial neural network.

Generally, vehicles are equipped with a transmission device to adjust the speed of the vehicle while driving. These transmission devices are divided into manual transmission devices operated by the driver and automatic transmission devices that automatically shift gears according to the vehicle's driving speed.

At this time, a double clutch transmission (DCT) equipped with two power transmission clutches has been developed as an automatic transmission.

Here, the double clutch transmission selectively transmits the rotational force input from the engine to the two input sides using two clutches, and outputs the transmission after shifting using the rotational force of gears disposed on the two input sides.

Such double-clutch transmissions are developed and applied in different formats depending on each automobile manufacturer, and currently, 5-speed automatic transmissions are the mainstream, improving power performance through improved fuel efficiency and efficient use of engine driving power. For this purpose, 6-speed and 7-speed automatic transmissions are being implemented.

That is, in a double clutch transmission, the 1st, 3rd, and 5th gears are connected to the first clutch, and the 2nd, 4th, and 6th gears are connected to the second clutch, and the double clutch transmission is shifted by the TCU (Transmission Control Unit). By controlling this, the first clutch and the second clutch alternately transmit power to the flywheel, thereby effecting shifting.

A brief look at the shifting process of this double clutch transmission is as follows.

First, shifting to first speed using the rotational force of the engine (ENT) involves synchronizing the first gear and the first output side through the sleeve of the first synchro mechanism, and then shifting to first speed by operating the first clutch. It will come true. Afterwards, when the vehicle speed increases in the first speed state and it is desired to shift to the second speed, the second speed gear and the first power transmission side are synchronized through the sleeve of the second synchro mechanism in the first speed state, and then the first clutch If you operate the second clutch while disengaging, the gear shift is performed to the second speed.

That is, like the first and second speed shifting processes, in other shifting processes, the synchro mechanism connects the gear of the corresponding shift stage to the corresponding power transmission side and alternately operates the first and second clutches to effect shifting. It will come true.

The background technology of the present invention is disclosed in Republic of Korea Patent No. 10-0716626 (registered on May 3, 2007, double clutch driving method for automatic transmission vehicle).

The clutch actuator, which operates the clutch above, controls how much the clutch and engine are engaged based on the absolute position.

Here, the reference point of the position is set to the 0 position by detecting the mechanical wall after applying current to the actuator motor.

Once the reference point of the position is determined in this way, the clutch actuator is controlled from the reference point to the target position through the PID controller.

At this time, the PID controller is configured to operate in the direction of reducing the error of the target control point and the current point by setting each gain that determines the degree to which the proportional, integral, and derivative terms for the error are reflected.

Since these PID controllers do not reflect the characteristics of the system to be controlled, the performance is better when the uncertainty of the system is not large. Since each gain uses a fixed value, it has the advantage of being easy to design, so it is used for simple control or high-level systems. This is widely adopted in existing lower level controllers.

However, since a fixed value is used for each gain, it is difficult to ensure even performance of the product if the hardware variation of the product is large. As the system becomes more complex, fixed gain values must be used in a table, so the table becomes large and calibration becomes more difficult. Not only does it require a lot of man-hours, but there is a problem that the performance of the product also depends on the person doing the calibration.

In addition, when producing a clutch actuator, product deviations occur depending on the tolerance of parts, the amount of lubricant applied, etc., and control using fixed values has a problem in that control performance deteriorates depending on the deviation.

The present invention was created to improve the problems described above. The purpose of the present invention according to one aspect is to control the position by reflecting the characteristics of the clutch actuator in a double clutch transmission by controlling the position by the difference between the actual position and the target position. To provide a clutch actuator control device and method for a double clutch transmission that automatically determines the gain ratio of the error component through intelligent PID control in the form of an artificial neural network.

A clutch actuator control device for a double clutch transmission according to one aspect of the present invention includes a clutch actuator; Memory; and a processor operatively coupled to the clutch actuator and memory; wherein the processor controls each gain for PID control of the artificial neural network structure for errors caused by the difference between the target position and the actual position of the clutch actuator. It is characterized by separating error components, setting parameters for each term, applying them to the error components of each PID term, and then outputting the control command calculated through the activation function to the clutch actuator.

In the present invention, the artificial neural network structure is characterized as a single-layer neural network structure.

In the present invention, the processor is characterized by learning according to Hebbian Rule and setting changes in parameters for each PID item.

In the present invention, the processor is characterized by calculating a control command as the sum of the ratios of the error components of each PID item through an activation function.

The method of controlling the clutch actuator of a double clutch transmission according to one aspect of the present invention involves the processor separating the error caused by the difference between the target position and the actual position of the clutch actuator into error components for each gain for PID control of the artificial neural network structure. step; A step of the processor setting parameters for each item of the PID; The processor applying the parameters for each term to the error component of each PID term; A step where the processor calculates a control command through an activation function for the result applied to the error component of each PID item; And a step of the processor outputting a control command to the clutch actuator.

In the present invention, the artificial neural network structure is characterized as a single-layer neural network structure.

In the present invention, the step of setting parameters for each item is characterized in that the processor learns according to Hebbian Rule and sets changes in parameters for each PID item.

In the present invention, the step of calculating a control command is characterized in that the processor calculates the sum of the ratios of the error components of each PID item through an activation function.

The clutch actuator control device and method of a double clutch transmission according to one aspect of the present invention is to control the position by reflecting the characteristics of the clutch actuator in the double clutch transmission by converting the error component due to the difference between the actual position and the target position into an artificial neural network. The gain ratio can be automatically determined through intelligent PID control, which not only maintains robust control performance despite hardware deviations and characteristic changes, but also reduces the man-hours required for calibration in the development stage.

1 is a block diagram showing a clutch actuator control device of a double clutch transmission according to an embodiment of the present invention.
Figure 2 is a schematic diagram showing the artificial neural network PID control structure in the clutch actuator control device of a double clutch transmission according to an embodiment of the present invention.
Figure 3 is an exemplary diagram showing simulation results by the clutch actuator control device of a double clutch transmission according to an embodiment of the present invention.
Figure 4 is an exemplary diagram showing the results of an actual vehicle test using the clutch actuator control device of a double clutch transmission according to an embodiment of the present invention.
Figure 5 is a flowchart for explaining a method of controlling a clutch actuator of a double clutch transmission according to an embodiment of the present invention.

Hereinafter, a clutch actuator control device and method of a double clutch transmission according to the present invention will be described with reference to the attached drawings. In this process, the thickness of lines or sizes of components shown in the drawing may be exaggerated for clarity and convenience of explanation. In addition, the terms described below are terms defined in consideration of functions in the present invention, and may vary depending on the intention or custom of the user or operator. Therefore, definitions of these terms should be made based on the content throughout this specification.

Figure 1 is a block diagram showing the clutch actuator control device of a double clutch transmission according to an embodiment of the present invention, and Figure 2 is an artificial neural network PID control in the clutch actuator control device of a double clutch transmission according to an embodiment of the present invention. It is a schematic diagram showing the structure, Figure 3 is an exemplary diagram showing the simulation results by the clutch actuator control device of the double clutch transmission according to an embodiment of the present invention, and Figure 4 is a diagram showing the simulation results of the double clutch transmission according to an embodiment of the present invention. This is an example diagram showing the actual vehicle test results using the clutch actuator control device.

1 shows a clutch actuator control device for a double clutch transmission according to an embodiment of the present invention, which may include a clutch actuator 30, a memory 10, and a processor 20.

The clutch actuator 30 can be engaged or disengaged by moving the clutch (not shown) through an actuator motor (not shown).

The processor 20 can control a plurality of hardware or software components connected to the processor 20 by running an operating system or application, and can perform various data processing and calculations. According to one embodiment, the processor 20 may be implemented as a system on chip (SoC). The processor 20 may load and process instructions or data received from at least one of the other components into memory and store various data in the memory 10 .

The processor 20 is operatively coupled to the clutch actuator 30 and the memory 10 to compensate for the error (e) caused by the difference between the target position (r) and the actual position (y) of the clutch actuator. Separate the error component for each gain for PID control of the artificial neural network structure, set the parameter (θ _ik ) for each term, apply it to the error component of each PID term, and then apply the control command (u) calculated through the activation function. It can be output to the clutch actuator 30.

_That is, based on the artificial neural network PID control structure, the processor 20 generates error components (x _p , _d ).

Here, the PID control structure can be configured as shown in Equation 1.

Here, the artificial neural network structure can be composed of a single layer neural network structure.

Additionally, as shown in FIG. 2, the processor 20 separates the error e into error components for each gain through the PID error calculator 210 and uses them as input to the artificial neural network.

At this time, the error components (x _p , x _i , x _d ) for each gain can be expressed as Equation 2.

Afterwards, the processor 20 sets the parameter (θ _i ) for each term as shown in Equation 3 and applies it to the error component of each PID term.

Here, the processor 20 can learn according to Hebbian Rule and set changes in the parameters (θ _p , θ _i , θ _d ) for each PID term.

The neuron's learning process for setting parameter changes can be expressed as Equation 3, where η is a gain value that can control the degree to which neuron learning is strengthened.

The final output by artificial neural network PID control with parameters applied in this way is summed and output as shown in Equation 4.

Afterwards, the processor 20 calculates a control command (u) as the sum of the ratios of the error components of each PID item through an activation function and outputs it to the clutch actuator 30.

Here, the activation function acts as a bridge connecting neurons, and the output value is used as the accuracy or probability of the path taken.

In this embodiment, one error was processed to create three error inputs, so the activation function can be expressed as the sum of the ratios for each error component. In other words, the output result of the activation function becomes the strength and percentage of the control amount to reduce the error in the latest state, which has the same meaning as how much duty to control the actuator motor.

In this embodiment, the activation function can be defined as Equation 5.

As shown in Figure 3, looking at the simulation results by the clutch actuator control device of the double clutch transmission according to this embodiment, it can be seen that when the red dotted line is the target value, it is almost identical to the blue dotted line showing the simulation result. .

In addition, when looking at the actual vehicle test results by the clutch actuator control device of the double clutch transmission as shown in FIG. 4, the dotted line is the target position and the solid line is the actual position, and it can be seen that position control is being performed normally.

As described above, according to the clutch actuator control device of the double clutch transmission according to the embodiment of the present invention, the error due to the difference between the actual position and the target position can be controlled to control the position by reflecting the characteristics of the clutch actuator in the double clutch transmission. The gain ratio can be automatically determined through intelligent PID control in the form of an artificial neural network, which not only maintains robust control performance despite changes in hardware deviations and characteristics, but also reduces the man-hours required for calibration in the development stage.

Figure 5 is a flowchart for explaining a method of controlling a clutch actuator of a double clutch transmission according to an embodiment of the present invention.

As shown in FIG. 5, in the clutch actuator control method of a double clutch transmission according to an embodiment of the present invention, first, the processor performs PID control of an artificial neural network structure for errors caused by the difference between the target position and the actual position of the clutch actuator. Separate the error components for each gain (S10).

_That is, based on the artificial neural network PID control structure, the processor 20 generates error components (x _p , _d ).

Here, the PID control structure can be configured as shown in Equation 6.

Here, the artificial neural network structure can be composed of a single layer neural network structure.

Therefore, as shown in FIG. 2, the processor 20 separates the error e into error components for each gain through the PID error calculator 210 and uses them as input to the artificial neural network.

At this time, the error components (x _p , x _i , x _d ) for each gain can be expressed as Equation 7.

After separating the air components for each gain in step S10, the processor 20 sets the parameter (θ _i ) for each term as shown in Equation 8 (S20).

Here, the processor 20 can learn according to Hebbian Rule and set changes in the parameters (θ _p , θ _i , θ _d ) for each PID term.

The learning process of neurons to set parameter changes can be expressed as Equation 8, where η is a gain value that can control the degree to which neuron learning is strengthened.

After setting the parameter changes for each PID term in step S20, the processor 20 applies the parameters to the error component of each term (S30).

After applying the parameters to the error component of each term in step S30, the processor 20 calculates a control command through an activation function for the result applied to the error component of each PID term (S40).

That is, the processor 20 sums the final output by the artificial neural network PID control to which the parameters are applied, as shown in Equation 9, and outputs it.

Afterwards, the processor 20 calculates the control command (u) as the sum of the ratios of the error components of each PID item through the activation function.

Here, the activation function acts as a bridge connecting neurons, and the output value is used as the accuracy or probability of the path taken.

In this embodiment, one error was processed to create three error inputs, so the activation function can be expressed as the sum of the ratios for each error component. In other words, the output result of the activation function becomes the strength and percentage of the control amount to reduce the error in the latest state, which has the same meaning as how much duty to control the actuator motor.

In this embodiment, the activation function can be defined as Equation 10.

After calculating the control command in step S40, the processor 20 outputs the control command (u) to the clutch actuator 30 to operate the clutch actuator 30 (S50).

As described above, according to the method of controlling the clutch actuator of a double clutch transmission according to an embodiment of the present invention, an error due to the difference between the actual position and the target position can be adjusted to control the position by reflecting the characteristics of the clutch actuator in the double clutch transmission. The gain ratio can be automatically determined through intelligent PID control in the form of an artificial neural network, which not only maintains robust control performance despite changes in hardware deviations and characteristics, but also reduces the man-hours required for calibration in the development stage.

Implementations described herein may be implemented, for example, as a method or process, device, software program, data stream, or signal. Although discussed only in the context of a single form of implementation (eg, only as a method), implementations of the features discussed may also be implemented in other forms (eg, devices or programs). The device may be implemented with appropriate hardware, software, firmware, etc. The method may be implemented in a device such as a processor, which generally refers to a processing device that includes a computer, microprocessor, integrated circuit, or programmable logic device. Processors also include communication devices such as computers, cell phones, portable/personal digital assistants (“PDAs”) and other devices that facilitate communication of information between end-users.

The present invention has been described with reference to the embodiments shown in the drawings, but these are merely exemplary, and those skilled in the art will recognize that various modifications and other equivalent embodiments are possible therefrom. You will understand.

Therefore, the true technical protection scope of the present invention should be determined by the claims below.

10: memory
20: processor
30: Clutch actuator

Claims (8)

Clutch actuator;
Memory; and
A processor operatively coupled to the clutch actuator and the memory,
The processor separates the error caused by the difference between the target position and the actual position of the clutch actuator into an error component for each gain for PID control of the artificial neural network structure, and sets parameters for each term to determine the error component of each PID term. A clutch actuator control device for a double clutch transmission, characterized in that it outputs a control command calculated through an activation function to the clutch actuator after applying it to the clutch actuator.
The clutch actuator control device of a double clutch transmission according to claim 1, wherein the artificial neural network structure is a single-layer neural network structure.
The clutch actuator control device of claim 1, wherein the processor sets changes in the parameters for each PID term by learning according to Hebbian Rule.
The clutch actuator control device of a double clutch transmission according to claim 1, wherein the processor calculates the control command as a sum of ratios of error components of each PID term through the activation function.
A step where the processor separates the error caused by the difference between the target position and the actual position of the clutch actuator into error components for each gain for PID control of the artificial neural network structure;
setting, by the processor, parameters for each PID item;
the processor applying the parameters for each term to an error component of each PID term;
The processor calculating a control command through an activation function for the result applied to the error component of each PID item; and
A method of controlling a clutch actuator of a double clutch transmission, comprising: the processor outputting the control command to the clutch actuator.
The clutch actuator control method of a double clutch transmission according to claim 5, wherein the artificial neural network structure is a single-layer neural network structure.
The method of claim 5, wherein the step of setting the parameters for each term is characterized in that the processor learns according to Hebbian Rule and sets changes in the parameters for each PID term. Clutch actuator control method for double clutch transmission.
The clutch actuator control method of a double clutch transmission according to claim 5, wherein in calculating the control command, the processor calculates the sum of ratios for error components of each PID term through the activation function.

Images