CN114578740B - A software driver control method based on improved active disturbance rejection control - Google Patents

Abstract

The invention relates to the field of automatic control technology, and specifically relates to a software driver control method based on improved active disturbance rejection control. The total disturbance term is determined based on the control input gain estimate and the dynamic model of the software driver, and the total disturbance term is decomposed into known The disturbance information part and the unknown disturbance information part compensate the system control law according to the control input gain estimate, time delay and known disturbance information part, and obtain an input quantity of the extended state observer ESO. The output quantity of ESO is controlled by PD. Output the state feedback control law. The system control law is obtained based on the state feedback control law, the known disturbance information part, the output of the ESO, and the control input gain estimate. According to the known disturbance information part and the delay time, one input terminal of the ESO is Compensation can reduce the estimation burden of ESO and improve the system's immunity, tracking and robustness without increasing the system's adjustable parameters.

Description

A software driver control method based on improved active disturbance rejection control

Technical field

The invention relates to the field of automatic control technology, and in particular to a software driver control method based on improved active disturbance rejection control.

Background technique

Software actuators have attracted more and more attention from scholars due to their inherent flexibility, flexibility, safety and other characteristics. They can better adapt to unstructured environments and more reliably grasp various irregularly shaped objects.

However, due to the characteristics of highly nonlinear, strong coupling, time-varying and strong elastic effects of the software driver itself, it is very difficult to establish an accurate model of the software driver, which leads to the need to consider the actual existence of the software driver system when designing the controller. Various uncertainties, such as unmodeled system dynamics, external disturbances, and system internal parameter disturbances.

Active Disturbance Rejection Control (ADRC), as a controller that does not rely on an accurate model of the system, is very effective in solving control problems of nonlinear control systems with uncertainties such as disturbances. In order to further improve the control performance of the active disturbance rejection controller, in related technologies, the system disturbance is first estimated through one or more extended state observers ESO, and the estimated value is used as a compensation term to compensate for the ADRC. This can reduce the estimation burden of ESO in ADRC and improve the system's anti-interference capability. However, the problem with related technologies is that increasing the number of ESOs will lead to an increase in the system's adjustable parameters, which will make parameter adjustment more difficult. At the same time, it will also Increase the difficulty of system theoretical analysis.

Contents of the invention

In order to solve the above technical problems, the present invention provides a software driver control method based on improved active disturbance rejection control.

The present invention adopts the following technical solutions:

A software driver control method based on improved active disturbance rejection control, including:

Construct a dynamic model of a software driver;

Obtain the control input gain estimate and the delay time, determine the total disturbance term of the system based on the control input gain estimate and the dynamic model, and decompose the total disturbance term into two parts, which are known disturbance information. Part and unknown disturbance information part;

According to the control input gain estimate, the delay time and the known disturbance information part, the system control law is compensated to obtain the first input quantity, and the first input quantity and the system output quantity serve as the expansion state observer Two input quantities are used to obtain three output quantities of the expanded state observer. The first output quantity and the second output quantity of the expanded state observer and the system reference input signal are input to the PD controller and the state feedback control law is output;

The system control law is obtained according to the state feedback control law, the known disturbance information part, the third output quantity of the expanded state observer and the control input gain estimate.

Further, the system control law is compensated according to the control input gain estimate, the delay time and the known disturbance information part, and the first input quantity is obtained, including:

Delay compensation is performed on the system control law according to the delay time to obtain the delay compensation amount, and then the delay compensation amount is multiplied by the control input gain estimate to obtain the first intermediate compensation amount, and finally the all The first intermediate compensation amount is summed with the known disturbance information part to obtain the first input amount.

Further, the expanded state observer is designed as follows:

Among them, z ₁ , z ₂ and z ₃ are the three output quantities of the expansion state observer respectively, where z ₁ and z ₂ respectively represent the observation values of the bending angle and angular velocity of the software actuator by the expansion state observer, z ₃ represents the estimated value of the unknown disturbance information part by the extended state observer, b ₀ is the estimated value of the control input gain, t _d is the delay time, q is the bending angle of the software driver, τ is the system control law, β ₁ , β ₂ and β ₃ are the gains of the extended state observer respectively.

Further, the first output quantity and the second output quantity of the expanded state observer and the system reference input signal are input to the PD controller, and the output state feedback control law includes:

The state feedback control law u ₀ is as follows:

Among them, r represents the system reference input signal, k _p and k _d are the controller gains.

Further, obtaining the system control law based on the state feedback control law, the known disturbance information part, the third output quantity of the expanded state observer and the control input gain estimate includes:

The system control law τ is as follows:

u ₀ is the state feedback control law.

Further, the dynamic model of the software driver is:

Among them, M(q) is the system inertia term, is the Croriot term, G(q) is the gravity term, τ is the system control law, τ _d includes the unmodeled dynamics of the system as well as internal and external disturbances of the system, q is the bending angle of the soft-body actuator,/> is the angular velocity,/> is the angular acceleration.

First, establish a dynamic model of the software driver, and then obtain the estimated value of the system control input gain and delay time. Based on the estimated value of the control input gain and the dynamic model of the software driver, the total disturbance term of the system is determined, and the total disturbance term of the system is It is decomposed into two parts: known model information and unknown disturbance information. According to the control input gain estimate, time delay and known disturbance information, the system control law is compensated to obtain the first input quantity, the first input quantity and the system output. As the two input quantities of the expanded state observer, three output quantities of the expanded state observer are obtained. The first and second output quantities of the expanded state observer and the system reference input signal are input to the PD controller, and the output state Feedback control law, and the system control law is obtained based on the state feedback control law, the known disturbance information part, the third output quantity of the expanded state observer and the control input gain estimate. This method compensates one input end of the expanded state observer based on the known disturbance information part and time delay, which can reduce the estimation burden of the expanded state observer without increasing the adjustable parameters of the system, thereby improving the system's resistance. interference, tracking and robustness.

Description of the drawings

In order to explain the technical solutions of the embodiments of the present invention more clearly, the drawings required to be used in the embodiments will be briefly introduced below:

Figure 1 is a control diagram corresponding to a software driver control method based on improved active disturbance rejection control provided by an embodiment of the present application;

Figure 2 is a schematic structural diagram of a software driver based on a software driver control method based on improved active disturbance rejection control provided by an embodiment of the present application;

Figure 3 is an output schematic diagram of a software driver control method based on improved active disturbance rejection control provided by an embodiment of the present application;

Figure 4 is a schematic diagram of Monte Carlo experiment output of a software driver control method based on improved active disturbance rejection control provided by an embodiment of the present application.

Detailed ways

In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures and technologies are provided to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It will be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and/or components but does not exclude one or more other The presence or addition of features, integers, steps, operations, elements, components and/or collections thereof.

It will also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

As used in this specification and the appended claims, the term "if" may be interpreted as "when" or "once" or "in response to determining" or "in response to detecting" depending on the context. ". Similarly, the phrase "if determined" or "if [the described condition or event] is detected" may be interpreted, depending on the context, to mean "once determined" or "in response to a determination" or "once the [described condition or event] is detected ]" or "in response to detection of [the described condition or event]".

In addition, in the description of this application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

Reference in this specification to "one embodiment" or "some embodiments" or the like means that a particular feature, structure or characteristic described in connection with the embodiment is included in one or more embodiments of the application. Therefore, the phrases "in one embodiment", "in some embodiments", "in other embodiments", "in other embodiments", etc. appearing in different places in this specification are not necessarily References are made to the same embodiment, but rather to "one or more but not all embodiments" unless specifically stated otherwise. The terms “including,” “includes,” “having,” and variations thereof all mean “including but not limited to,” unless otherwise specifically emphasized.

In order to illustrate the technical solutions described in this application, specific implementations will be described below.

Figure 1 is a control diagram corresponding to a software driver control method based on improved active disturbance rejection control provided by an embodiment of the present application.

The software driver control method based on improved active disturbance rejection control includes the following steps:

Construct a dynamic model of the software driver as:

Among them, M(q) is the system inertia term, is the Croriot term, G(q) is the gravity term, τ is the system control law, that is, the generalized force of the system, τ _d includes the unmodeled dynamics of the system as well as internal and external disturbances of the system, q is the bending angle of the soft actuator, /> is the angular velocity,/> is the angular acceleration.

It should be noted that the soft actuator in the embodiment of the present invention is a pneumatic soft actuator. Compared with a rigid robot, the G(q) term in the soft actuator model includes not only the gravitational potential energy part, but also the strain energy part.

Specifically, M(q), The expression of G(q) is as follows:

Among them, W _p , L _p , H _p1 , H _p2 , and L _pq are to simplify the expressions in formula (2). Their expressions are as follows:

As shown in Figure 2, W, L, and H represent the width, length, and height of the pneumatic software actuator respectively. W ₁ and W ₂ represent the width of the front wall and side wall respectively. L ₁ and L ₂₀ represent the length and cavity of the outer wall respectively. The initial length between the two side walls of the chamber, H ₁ , H ₂ , H ₃ represent the height of the outer wall, upper wall and bottom layer respectively, N is the number of chambers, G is the shear modulus, m is the mass of the software driver, g is the gravitational constant. The values of each of the above parameters can be identified through parameter identification.

Obtain the control input gain estimate b ₀ and the delay time t _d , and determine the total disturbance term of the system based on the control input gain estimate b ₀ and the dynamics model. And add the total disturbance term/> It is decomposed into two parts, namely the known disturbance information part/> and unknown disturbance information section/>

It is understandable that for the convenience of analysis, the established software driver dynamics model is transformed into the following form:

It can be seen from this that the total disturbance term of the system In addition, b ₀ is the estimated value of the system control input gain, that is, the system inertia term, that is, the estimated value of the system's real control input gain M ^-1 (q), which is an adjustable parameter.

Furthermore, in order to better estimate the unknown disturbance information part Will/> Expand to a new system state x ₃ and assume/> The first derivative of exists and is equal to h, that is/> In this way, the software driver system can be written in the form of the following state equation:

In addition, according to an embodiment of the present invention, the estimated value b ₀ of the system control input gain, the delay time t _d and the known disturbance information part can be obtained respectively through the method of model identification.

According to a specific embodiment of the present invention, the software driver model can be identified through the system identification toolbox in matlab, and the identified object is:

From this, the estimated value of the system control input gain b ₀ =10330, the delay time t _d =0.002 and the known disturbance information part can be determined based on the identified objects.

According to the control input gain estimate b ₀ , the delay time t _d and the known disturbance information part A second-order linear active disturbance rejection controller is designed to control the software driver to improve the control performance of the system.

details as follows:

According to the control input gain estimate b ₀ , the delay time t _d and the known disturbance information part Compensate the system control law τ to obtain the first input quantity of the extended state observer ESO. The specific compensation process is as follows: perform delay compensation on the system control law τ according to the delay time t _d to obtain the delay compensation amount τ (tt _d ), then multiply the time delay compensation amount τ(tt _d ) and the control input gain estimate b ₀ to obtain the first intermediate compensation amount, and finally combine the first intermediate compensation amount and the known disturbance information part/> Do the sum to obtain the first input quantity of the extended state observer ESO.

Obtain the system output quantity q as the second input quantity of the expanded state observer ESO, that is, the first input quantity and the system output quantity q mentioned above are used as the two input quantities of the expanded state observer ESO, and after processing by the expanded state observer ESO , three output quantities of the extended state observer ESO are obtained, which are z ₁ , z ₂ and z ₃ respectively.

The expanded state observer ESO is designed as follows:

Among them, z ₁ and z ₂ respectively represent the bending angle q and angular velocity of the software actuator by the expansion state observer ESO. Observation value, z ₃ represents the unknown disturbance information part of the extended state observer ESO/> The estimated values of β ₁ , β ₂ and β ₃ are the gains of the expansion state observer ESO respectively. Since z ₁ and z ₂ respectively represent the bending angle q and angular velocity of the software actuator by the expansion state observer ESO/> Observed values, then f ₁ (z ₁ , z ₂ ) is also expressed as the known disturbance information part.

When the observer gain is properly adjusted, the observer outputs z ₁ , z ₂ and z ₃ are closer to the bending angle q and angular velocity of the software actuator. Observed values and unknown disturbance information/> estimated value. Among them, parameters can be adjusted through methods such as bandwidth parameterization method or multi-objective optimization method.

It can be understood that the conventional ADRC (self-disturbance controller) directly uses the control variable τ, that is, the product of the system control law τ and the estimated value b ₀ of the control input gain, as an input quantity of the extended state observer ESO. However, if As shown in Figure 1, the improved active disturbance rejection controller (MADRC) of the embodiment of the present invention performs two parts of compensation on the control variable τ and then inputs it to the ESO. The first part of the compensation is based on the delay time t _d . Compensation is made with the variable τ, that is, the control variable τ output by the controller is delayed by the delay time t _d . This can avoid the out-of-synchronization of the two input quantities of ESO, namely the control variable τ and the output quantity q, caused by the actual system delay. problems, thereby improving the control performance of the control system. The second part of compensation is to convert the known model information into Add to the product of the control variable τ(tt _d ) after the first part of compensation and the estimated value b ₀ of the control input gain. As a result, the estimation burden of ESO can be reduced, the observation accuracy of ESO can be improved, and the anti-interference ability, rapidity and robustness of the system can be improved.

The first output quantity z ₁ and the second output quantity z ₂ of the extended state observer and the system reference input signal r are input to the PD controller, and the output state feedback control law u ₀ is as follows:

Among them, k _p and k _d are the controller gains.

According to the state feedback control law u ₀ , the known disturbance information part The third output quantity z ₃ of the extended state observer and the control input gain estimate b ₀ are used to obtain the system control law τ, as follows:

Therefore, the system control law τ is designed to compensate for the estimation of unknown disturbances, thereby improving system tracking and anti-interference capabilities.

It can be understood that, as shown in Figure 1, the improved active disturbance rejection controller MADRC according to the embodiment of the present invention also adds a known disturbance information part when compensating for unknown disturbances.

The designed control law Substituting into the software driver model we get:

That is, by compensating for the estimated unknown disturbance and the known disturbance information part, the software actuator system can be converted into an integral series object.

Therefore, for this compensated integral series object, the control method of the PD controller mentioned above is designed, that is, the state feedback control law u ₀ .

Therefore, according to a specific embodiment of the present invention, the improved active disturbance rejection controller MADRC of the embodiment of the present invention is used to control the software driver shown in the aforementioned step S1, and a disturbance is added at the simulation time of 25s to use the traditional ADRC and PID are used as comparison controllers to obtain the simulation output diagram as shown in Figure 3. It can be seen from Figure 3 that the improved active disturbance rejection controller MADRC according to the embodiment of the present invention has good performance in terms of tracking speed and disturbance immunity. All are better than traditional ADRC and PID. At the same time, when the model parameters L, L1, L20 and m of the software driver change within the range of ±20% of their values, and when W2 and H1 change within the range of ±10% of their values, the improvements of the embodiments of the present invention are respectively made. The anti-interference controller MADRC, traditional ADRC and PID control software driver were subjected to Monte Carlo experiments to obtain the experimental diagram shown in Figure 4. It can be seen from the figure that compared with the traditional ADRC and PID controller, the embodiment of the present invention has The value of the time multiplied by the integrated absolute value of the error (ITAE) index output by the improved active disturbance rejection controller MADRC is smaller, where ITAE-sp represents the ITAE index in the tracking stage before the disturbance is added, ITAE-id represents the ITAE index after adding disturbance, so the improved active disturbance rejection controller in the embodiment of the present invention is more robust.

In summary, according to the software driver control method based on improved active disturbance rejection control according to the embodiment of the present invention, the dynamic model of the software driver is first established, and then the estimated value and delay time of the system control input gain are obtained. According to the estimated control input gain value and the dynamic model of the software actuator to determine the total disturbance term of the system, and decompose the total disturbance term of the system into the known disturbance information part and the unknown disturbance information part. Finally, based on the known disturbance information part and the delay time, design the second An order linear active disturbance rejection controller controls the software driver. Therefore, the software driver control method based on improved active disturbance rejection control according to the embodiment of the present invention can reduce the estimation burden of ESO without increasing the system's adjustable parameters, thereby improving the system's immunity, tracking and robustness. Great sex.

The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still implement the above-mentioned implementations. The technical solutions described in the examples are modified, or some of the technical features are equivalently replaced; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions in the embodiments of this application, and should be included in within the protection scope of this application.

Claims (1)

1. A software driver control method based on improved active disturbance rejection control, which is characterized by including: Construct a dynamic model of a software driver; Obtain the control input gain estimate and the delay time, determine the total disturbance term of the system based on the control input gain estimate and the dynamic model, and decompose the total disturbance term into two parts, which are known disturbance information. Part and unknown disturbance information part; According to the control input gain estimate, the delay time and the known disturbance information part, the system control law is compensated to obtain the first input quantity, and the first input quantity and the system output quantity serve as the expansion state observer Two input quantities are used to obtain three output quantities of the expanded state observer. The first output quantity and the second output quantity of the expanded state observer and the system reference input signal are input to the PD controller and the state feedback control law is output; The system control law is obtained according to the state feedback control law, the known disturbance information part, the third output quantity of the expanded state observer and the control input gain estimate; Compensating the system control law according to the control input gain estimate, the delay time and the known disturbance information part to obtain the first input quantity includes: Delay compensation is performed on the system control law according to the delay time to obtain the delay compensation amount, and then the delay compensation amount is multiplied by the control input gain estimate to obtain the first intermediate compensation amount, and finally the all The first intermediate compensation amount is summed with the known disturbance information part to obtain the first input amount; The expanded state observer is designed as follows: Among them, z ₁ , z ₂ and z ₃ are the three output quantities of the expansion state observer respectively, where z ₁ and z ₂ respectively represent the observation values of the bending angle and angular velocity of the software actuator by the expansion state observer, z ₃ represents the estimated value of the unknown disturbance information part by the extended state observer, b ₀ is the estimated value of the control input gain, t _d is the delay time, q is the bending angle of the software driver, τ is the system control law, β ₁ , β ₂ and β ₃ are the gains of the expanded state observer respectively; The first output quantity and the second output quantity of the expanded state observer and the system reference input signal are input to the PD controller, and the output state feedback control law includes: The state feedback control law u ₀ is as follows: Among them, r represents the system reference input signal, k _p and k _d are the controller gains; Obtaining the system control law based on the state feedback control law, the known disturbance information part, the third output quantity of the expanded state observer and the control input gain estimate includes: The system control law τ is as follows: u ₀ is the state feedback control law; The dynamic model of the software driver is: Among them, M(q) is the system inertia term, is the Croriot term, G(q) is the gravity term, τ is the system control law, τ _d includes the unmodeled dynamics of the system as well as internal and external disturbances of the system, q is the bending angle of the soft-body actuator,/> is the angular velocity,/> is the angular acceleration.