Network switching control method and system based on weighted edge dependence

Abstract

The invention relates to the technical field of networked control systems, in particular to a network switching control method and system based on weighted edge dependency, and the method comprises the steps: firstly building a networked control system model containing uncertain items, and carrying out the optimization of the uncertain items based on a set packet loss number threshold value of the networked control system; switching signals are introduced into the networked control system model containing the uncertain items, and a continuous time linear switching system is established; designing a sampling controller based on the time-varying time-lag of each subsystem to be switched, obtaining the control input of the continuous-time linear switching system, and establishing a time-varying time-lag uncertain switching system; and carrying out weighting processing based on edge dependence on a plurality of subsystems to be switched to obtain weighted allowable edge dependence residence time of all other subsystems switched to a certain subsystem, switching the subsystems based on the weighted allowable edge dependence residence time and a set residence time threshold value, and switching all the subsystems to be switched to the certain subsystem according to the weighted allowable edge dependence residence time and the set residence time threshold value. When part of control input information of the networked control system is lost, normal operation of the system can still be ensured, and the network service quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of networked control systems, and in particular to a network switching control method and system based on weighted edge dependency. Background Art

[0002] With the rapid development of information communication technology and computer networks, the traditional point-to-point feedback control system has been replaced by a more complex remote control or distributed control management system called a networked control system. Compared with traditional control systems, networked control systems can greatly reduce system size, integration and maintenance costs, and improve system flexibility and reliability.

[0003] However, due to the aging of important components such as controllers, unknown dynamic changes in the external environment, and the presence of the network in the control loop, some unfavorable factors such as random delays and packet loss, the loss of control input is inevitable. When the control input is lost, the information of the controller cannot be transmitted normally, which reduces the system performance and may even cause the networked control system to fail to operate normally. Therefore, it is very important and necessary to study networked control systems.

[0004] Generally speaking, a switching system is composed of a finite number of switching subsystems and a switching strategy used to regulate the operation of each subsystem. How to introduce the idea of ​​switching into the study of networked control systems and design the corresponding controller based on whether there is a loss phenomenon in the control input has always been a research hotspot. Summary of the invention

[0005] The present invention provides a network switching control method and system based on weighted edge dependency.

[0006] The technical solution of the present invention is as follows:

[0007] A network switching control method based on weighted edge dependency comprises the following steps:

[0008] S1. Based on the state, control input and uncertainty factors of the networked control system, a networked control system model including uncertainties is established;

[0009] S2. Based on a set packet loss threshold of the networked control system, a switching signal is introduced into a networked control system model containing uncertain terms to establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched;

[0010] S3. Based on the time-varying delay of each subsystem to be switched, a sampling controller is designed to obtain the control input of the continuous-time linear switching system and establish an uncertain switching system with time-varying delay;

[0011] S4. Perform edge dependency-based weighted processing on several subsystems to be switched to obtain weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and a set residence time threshold.

[0012] Specifically, in S4, based on the weighted admissible edge dependency residence time and the set residence time threshold, the subsystem is switched, specifically:

[0013] ,

[0014] in, is the dwell time threshold, is the weighting coefficient, i.e., weight, And meet ; is the edge dependency residence time of switching to any other subsystem i of a certain subsystem j.

[0015] Furthermore, the Lyapunov-Krasovskii functional applied to subsystem switching is constructed, and the conditions for ensuring the stability of the time-varying and time-delay uncertain switching system when the control input is lost are obtained, which is expressed as:

[0016] ,

[0017] in, is the Lyapunov-Krasovskii functional corresponding to the switching signal σ(t), P, Q, and R are all positive definite matrices, , , are all integral variables.

[0018] Specifically, the networked control system model in S1 for:

[0019] ,

[0020] in, is the state of the networked control system, For control input, A and B are matrices of known dimensions; and is an uncertain item, which is expressed as follows:

[0021] ,

[0022] in, , , All are known matrices; is a time-varying matrix function about t, satisfying the following conditions: , is the identity matrix.

[0023] Specifically, the continuous-time linear switching system in S1 is:

[0024] ,

[0025] Where σ(t) is the switching signal, and , N is the number of subsystems to be switched, N is a positive integer; when , indicating that the jth subsystem is running at time t, and the uncertainty satisfy: ; is a known matrix.

[0026] Furthermore, the control input of the continuous-time linear switching system in S1 is for:

[0027] ,

[0028] in, is the kth sampling moment, is the state of the continuous-time linear switching system, is the k+1th sampling moment, and t is the time interval At any time, is the gain matrix of the sampling controller;

[0029] Define the time interval as , the time-varying delay is , the established time-varying and time-delay uncertain switching system is expressed as follows:

[0030] ,

[0031] ,

[0032] in, is a bounded continuous function, For time interval The maximum value of .

[0033] Specifically, in S4, several subsystems to be switched are weightedly processed based on edge dependency, specifically: each subsystem is taken as a node, an edge is constructed between any two nodes with a switching relationship, the weight of each edge is set based on the residence time of switching between the two nodes, and the residence time is weighted and summed according to the weight to obtain the weighted allowable edge-dependent residence time of all other subsystems switched to a certain subsystem.

[0034] Specifically, the uncertainty factors in S1 include environmental change factors, system parameter fluctuation factors, and network communication instability factors. The network communication instability factors include: network delay and network packet loss.

[0035] Furthermore, the number of subsystems to be switched in S2 is It is associated with the packet loss threshold m of the networked control system, specifically:

[0036] ,

[0037] Among them, m is the packet loss threshold, that is, the maximum number of input losses.

[0038] The present invention also provides a network switching control system based on weighted edge dependency, comprising:

[0039] Networked control system model building module: used to build a networked control system model containing uncertainties based on the state, control input, and uncertainty factors of the networked control system;

[0040] A continuous-time linear switching system establishment module: used to introduce a switching signal into a networked control system model containing uncertain terms based on a set packet loss threshold of the networked control system, and establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched;

[0041] Time-varying and time-delay uncertain switching system establishment module: used to design a sampling controller based on the time-varying time delay of each subsystem to be switched, obtain the control input of the continuous-time linear switching system, and establish a time-varying and time-delay uncertain switching system;

[0042] Subsystem switching module: It is used to perform weighted processing based on edge dependency on several subsystems to be switched, obtain the weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and the set residence time threshold.

[0043] In addition, the present invention provides a network switching control device based on weighted edge dependency, comprising a processor and a memory, wherein the processor implements the network switching control method based on weighted edge dependency as described above when executing a computer program stored in the memory.

[0044] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps of the network switching control method based on weighted edge dependency as described above.

[0045] The beneficial effects of the present invention are:

[0046] 1. The present invention proposes a weighted allowable edge-dependent average residence time switching method, that is, a weighted processing based on edge dependency is performed on several subsystems to be switched to obtain the weighted allowable edge-dependent residence time of all other subsystems switched to a certain subsystem. Based on the weighted allowable edge-dependent residence time and the set residence time threshold, the subsystem is switched, which takes into account both the difference between the residence times of subsystems switched to the same subsystem and the connection between them, so as to obtain a larger residence time selection range.

[0047] 2. The present invention designs a sampling controller and models the networked control system as a time-varying, time-delay uncertain switching system based on whether the controller information is lost. When the control input is detected to be lost, the controller of the corresponding subsystem is applied. When part of the control input information of the networked control system is lost, the normal operation of the system can be ensured, thereby improving the quality of network services.

[0048] 3. The present invention applies Lyapunov-Krasovskii functional theory to the time-varying and time-delay uncertain switching system, obtains relevant robust stability conclusions, and obtains the gain matrix of the controller, effectively controls the networked control system, improves network performance, and improves network service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In the attached picture:

[0050] Figure 1 Schematic diagram of the flow of a network switching control method based on weighted edge dependency in an embodiment. DETAILED DESCRIPTION

[0051] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings.

[0052] Example

[0053] This embodiment provides a network switching control method based on weighted edge dependency. Figure 1 , including the following steps:

[0054] S1. Based on the state, control input and uncertainty factors of the networked control system, a networked control system model including uncertainties is established;

[0055] Networked Control System Model for:

[0056] ,

[0057] in, is the state of the networked control system, For control input, A and B are matrices of known dimensions; and is an uncertain item, which is expressed as follows:

[0058] ,

[0059] in, , , All are known matrices; is a time-varying matrix function about t, satisfying the following conditions: , is the identity matrix.

[0060] In this embodiment, the matrix parameters of the system are set as follows:

[0061] .

[0062] Specifically, the uncertainty factors include environmental change factors, system parameter fluctuation factors, and network communication instability factors. The network communication instability factors include: network delay and network packet loss.

[0063] S2. Based on the set packet loss threshold of the networked control system, a switching signal is introduced into the networked control system model containing uncertain terms to establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched.

[0064] The continuous-time linear switching system is:

[0065] ,

[0066] Where σ(t) is the switching signal, and , N is the number of subsystems to be switched, N is a positive integer; when , indicating that the jth subsystem is running at time t, and the uncertainty satisfy: ; is a known matrix.

[0067] In this embodiment, the number of subsystems to be switched is It is associated with the packet loss threshold m of the networked control system, specifically:

[0068] ,

[0069] Among them, m is the packet loss threshold, that is, the maximum number of input losses.

[0070] Specifically, the maximum number of packet losses allowed by the networked control system is set to m, that is, the system needs to switch when the number of packet losses is less than or equal to m. When the number of packet losses is 0, the system is in an ideal working state; when the number of packet losses is between 1 and m, the system needs to switch to different subsystems to cope with different degrees of packet loss.

[0071] For each control input, there are two possibilities: lost or not lost. The corresponding packet loss number ranges from 1 to m, with a total of 2 m There are 2 possible subsystems, excluding the case where the number of packet losses is 0. m -1 subsystem to be switched.

[0072] set up , according to the number of control input losses, the continuous-time linear switching system consists of three subsystems, and the specific matrix values ​​are as follows:

[0073] .

[0074] Assume that the switching sequence is 1→2→3→1→. Set the system initial state to , the control result of the system under the present invention can be obtained, which can prove that the stable operation of the system is guaranteed under the design of the switching scheme. That is, when part of the control input information of the networked control system is lost, it can ensure that the system can normally receive and transmit information, thereby improving the quality of network services.

[0075] At the same time, in this embodiment, by solving a set of matrix inequalities, the gain matrix of the controller can be obtained as follows:

[0076] ,

[0077] in, , , They are the gain matrices of the controller when subsystem j=1, 2, and 3 respectively.

[0078] S3. Based on the time-varying delay of each subsystem to be switched, a sampling controller is designed to obtain the control input of the continuous-time linear switching system and establish a time-varying delay uncertain switching system.

[0079] The control input of the continuous-time linear switching system for:

[0080] ,

[0081] in, is the kth sampling moment, is the state of the continuous-time linear switching system, is the k+1th sampling moment, and t is the time interval At any time, is the gain matrix of the sampling controller;

[0082] Define the time interval as , the time-varying delay is , the established time-varying and time-delay uncertain switching system is expressed as follows:

[0083] ,

[0084] ,

[0085] in, is a bounded continuous function, For time interval The maximum value of ;

[0086] S4. Perform edge dependency-based weighted processing on several subsystems to be switched to obtain weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and a set residence time threshold.

[0087] A weighted processing based on edge dependency is performed on several subsystems to be switched. Specifically, each subsystem is taken as a node, and edges are constructed between any two nodes with a switching relationship. The weight of each edge is set based on the residence time of switching between the two nodes. The residence time is weighted and summed according to the weight to obtain the weighted allowable edge-dependent residence time of all other subsystems switched to a certain subsystem.

[0088] Based on the weighted admissible edge dependency dwell time and the set dwell time threshold, the subsystem is switched, specifically:

[0089] ,

[0090] in, is the dwell time threshold, is the weighting coefficient, i.e., weight, And meet ; is the edge dependency residence time of switching to any other subsystem i of a certain subsystem j.

[0091] As long as the weighted allowable edge dependency residence time is greater than or equal to the set residence time threshold, you can switch to any subsystem that meets this condition.

[0092] The Lyapunov-Krasovskii functional applied to subsystem switching is constructed, and the conditions for ensuring the stability of the time-varying and time-delay uncertain switching system when the control input is lost are obtained. The formula is expressed as:

[0093] ,

[0094] in, is the Lyapunov-Krasovskii functional corresponding to the switching signal σ(t), P, Q, and R are all positive definite matrices, , , are all integral variables.

[0095] The controller gain matrix is ​​obtained as .

[0096] The present invention also provides a network switching control system based on weighted edge dependency, comprising:

[0097] Networked control system model building module: used to build a networked control system model containing uncertainties based on the state, control input, and uncertainty factors of the networked control system;

[0098] A continuous-time linear switching system establishment module: used to introduce a switching signal into a networked control system model containing uncertain terms based on a set packet loss threshold of the networked control system, and establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched;

[0099] Time-varying and time-delay uncertain switching system establishment module: used to design a sampling controller based on the time-varying time delay of each subsystem to be switched, obtain the control input of the continuous-time linear switching system, and establish a time-varying and time-delay uncertain switching system;

[0100] Subsystem switching module: used to perform weighted processing based on edge dependency on several subsystems to be switched, obtain the weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and the set residence time threshold.

[0101] The present invention proposes a weighted allowable edge-dependent average dwell time switching method, that is, a weighted processing based on edge dependence is performed on a number of subsystems to be switched, and the weighted allowable edge-dependent dwell time of all other subsystems switched to a certain subsystem is obtained. Based on the weighted allowable edge-dependent dwell time and the set dwell time threshold, the subsystem is switched, which not only considers the difference between the dwell times of subsystems switched to the same subsystem, but also considers the connection between them, and can obtain a larger dwell time selection range. The present invention designs a sampling controller, and models the networked control system into a time-varying and time-delay uncertain switching system based on whether the controller information is lost. When the control input is lost, the controller of the corresponding subsystem is applied, which can ensure the normal operation of the system when part of the control input information of the networked control system is lost, thereby improving the network service quality. The present invention applies the Lyapunov-Krasovskii functional theory to the time-varying and time-delay uncertain switching system, obtains the relevant robust stability conclusion, and obtains the gain matrix of the controller, effectively controls the networked control system, improves the network performance, and improves the network service quality.

Claims

1. A network switching control method based on weighted edge dependency, characterized in that: The following steps are involved: S1. Based on the state, control input and uncertainty factors of the networked control system, a networked control system model including uncertainties is established; S2. Based on a set packet loss threshold of the networked control system, a switching signal is introduced into a networked control system model containing uncertain terms to establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched; S3. Based on the time-varying delay of each subsystem to be switched, a sampling controller is designed to obtain the control input of the continuous-time linear switching system and establish an uncertain switching system with time-varying delay; S4. Perform edge dependency-based weighted processing on several subsystems to be switched to obtain weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and a set residence time threshold.

2. The network switching control method based on weighted edge dependency according to claim 1 is characterized in that: In S4, the subsystem is switched based on the weighted admissible edge dependency residence time and the set residence time threshold, specifically: ， in, is the dwell time threshold, is the weighting coefficient, i.e., weight, And meet ; is the edge dependency residence time of switching to any other subsystem i of a certain subsystem j.

3. The network switching control method based on weighted edge dependency according to claim 2 is characterized in that: The Lyapunov functional applied to subsystem switching is constructed, and the conditions for ensuring the stability of the time-varying and time-delay uncertain switching system when the control input is lost are obtained. The formula is expressed as: ， in, is the Lyapunov functional corresponding to the switching signal σ(t), P, Q, and R are all positive definite matrices, , , are all integral variables.

4. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: The networked control system model in S1 is: ， in, is the state of the networked control system, For control input, A and B are matrices of known dimensions; and is an uncertain item, which is expressed as follows: ， in, , , All are known matrices; is a time-varying matrix function about t, satisfying the following conditions: , is the identity matrix.

5. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: The continuous-time linear switching system in S1 for: ， Where σ(t) is the switching signal, and , N is the number of subsystems to be switched, N is a positive integer; when , indicating that at time t the jth subsystem is running, and the uncertainty satisfy: ; is a known matrix.

6. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: The control input of the continuous-time linear switching system in S1 for: ， in, is the kth sampling moment, is the state of the continuous-time linear switching system, is the k+1th sampling moment, and t is the time interval At any time, is the gain matrix of the sampling controller; Define the time interval as , the time-varying delay is , the established time-varying and time-delay uncertain switching system is expressed as follows: ， ， in, is a bounded continuous function, For time interval The maximum value of .

7. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: In S4, several subsystems to be switched are weightedly processed based on edge dependency, specifically: each subsystem is taken as a node, an edge is constructed between any two nodes with a switching relationship, the weight of each edge is set based on the residence time of switching between the two nodes, and the residence time is weighted and summed according to the weight to obtain the weighted allowable edge-dependent residence time of all other subsystems switched to a certain subsystem.

8. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: The uncertain factors in S1 include environmental change factors, system parameter fluctuation factors, and network communication instability factors. The network communication instability factors include: network delay and network packet loss.

9. The network switching control method based on weighted edge dependency according to claim 1, characterized in that: The number of subsystems to be switched in S2 It is associated with the packet loss threshold m of the networked control system, specifically: ， Among them, m is the packet loss threshold, that is, the maximum number of input losses.

10. A network switching control system based on weighted edge dependency, characterized in that: include: Networked control system model building module: used to build a networked control system model containing uncertainties based on the state, control input, and uncertainty factors of the networked control system; A continuous-time linear switching system establishment module: used to introduce a switching signal into a networked control system model containing uncertain terms based on a set packet loss threshold of the networked control system, and establish a continuous-time linear switching system, wherein the continuous-time linear switching system includes a plurality of subsystems to be switched; Time-varying and time-delay uncertain switching system establishment module: used to design a sampling controller based on the time-varying time delay of each subsystem to be switched, obtain the control input of the continuous-time linear switching system, and establish a time-varying and time-delay uncertain switching system; Subsystem switching module: used to perform weighted processing based on edge dependency on several subsystems to be switched, obtain the weighted allowable edge dependency residence time of all other subsystems switched to a certain subsystem, and switch the subsystem based on the weighted allowable edge dependency residence time and the set residence time threshold.

Images