Multi-agent self-triggering control method and system

Abstract

The invention provides a multi-agent self-triggering control method and system, each agent calculates a next triggering moment according to the information of the current moment and the information of an upstream neighbor agent, the calculated triggering moment is taken as a pre-triggering moment, and if other upstream neighbors are triggered before the pre-triggering moment, the other upstream neighbors are triggered before the pre-triggering moment. If yes, updating and calculating the pre-triggering moment, and otherwise, performing triggering control on the intelligent agent by taking the current pre-triggering moment as the triggering time; an event trigger parameter and a time trigger parameter are fused in the calculation process of the trigger moment. According to the method, the advantages of event triggering and time triggering are combined, communication and calculation in a specified mode only need to be carried out at the triggering moment, continuous communication and calculation are not needed, the communication efficiency is improved, the system stability is improved, and the energy consumption ratio is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-agent collaborative output regulation, and particularly relates to a multi-agent self-triggered control method and system. Background Art

[0002] The statements in this part merely provide background technical information related to the present invention and do not necessarily constitute prior art.

[0003] In the field of multi-agent control, time-triggered methods and event-triggered methods have been widely applied. The time-triggered method essentially calculates the trigger points by presetting a time interval length. Therefore, in order to ensure performance, small interval design is often required, otherwise the system may have poor effects or even be unstable. The event-triggered method is triggered when certain conditions are met and has the characteristic of more flexible trigger intervals compared with the time-triggered method.

[0004] In practical applications, improper design of the event-triggered mode will cause the Zeno effect. Although the time-triggered method will not cause the Zeno phenomenon no matter how it is designed, it often requires more trigger times to achieve a control effect similar to a reasonably designed event-triggered mode. Therefore, the current two technologies both have the characteristics of low efficiency and high energy consumption ratio.

[0005] Meanwhile, for a reasonably designed event-triggered method, its hidden defect is that even if a minimum time interval can be theoretically found, it may still be limited by physical entity devices, resulting in continuous triggering in the real scenario.

[0006] In summary, whether it is the time-triggered method or the event-triggered method, continuous communication and calculation are currently required, with low efficiency and high energy consumption ratio, which easily leads to poor system effects and instability. Summary of the Invention

[0007] To solve the above problems, the present invention proposes a multi-agent self-triggered control method and system. The present invention proposes a self-triggered control mechanism, which combines the advantages of event-triggered and time-triggered methods, and only needs to perform communication and calculation in a specified manner at the trigger moment, without continuous communication and calculation, improving the communication efficiency, helping to improve the system stability, and reducing the energy consumption ratio.

[0008] According to some embodiments, the present invention adopts the following technical solutions:

[0009] A multi-agent self-triggered control method includes the following steps:

[0010] Each agent calculates the next triggering time based on its own current information and the information of its upstream neighbor agents, and uses the calculated triggering time as the pre-triggering time. If other upstream neighbors are triggered before the pre-triggering time, the pre-triggering time is updated and calculated. Otherwise, the current pre-triggering time is used as the triggering time to control the triggering of its own agent.

[0011] The trigger moment calculation process incorporates event trigger parameters and time trigger parameters.

[0012] As an optional implementation, the process of each agent calculating the distance to the next triggering moment based on its own current moment information and the information of its upstream neighboring agents includes:

[0013] The next triggering moment of agent i is:

[0014] Among them, the is the next triggering moment of the current iteration number p, is the last triggering moment of agent i, is the time trigger parameter, The time interval parameter for event triggering.

[0015] As a further limited embodiment, the event triggering time interval parameter It is calculated based on the event trigger parameters, the information of the upstream neighbor agent, and the information of agent i.

[0016] As a further limited embodiment, the event triggering time interval parameter for:

[0017] When p is zero,

[0018] When p is not zero,

[0019]

[0020] in, is the external system matrix, which does not have a negative real part and is known in advance by each agent, σ i Indicates event trigger parameters,

[0021] is the number of agent i in the p-iteration process. The combined sequence of all upstream neighbor triggering moments in the time interval, μ represents the compensator gain. In order to model the flow of large amounts of data into another network topology, it is recorded as Assume that it is time-invariant and contains a directed spanning tree with node 0 as the root, and its edge weights are denoted by Here is the connection weight between agent i and node 0 in the graph ; is to merge the triggering times of all upstream neighbors of agent i in into a sequence, and denote the merged sequence as and define and Here is the previous triggering time of agent i, represents the next triggering time.

[0022] As a further implementation, the event triggering parameter σ i is:

[0023]

[0024] where 0 < w0 < 1, 0 < β < 1, h = max 1≤i≤N h i , h i is the i-th element of, where

[0025] is the Laplacian matrix of the graph, N i represents the set of upstream neighbors of agent i, Θ′ is the positive definite solution of the following matrix equation:

[0026]

[0027] represents the maximum value of the number of neighbors of all agents, I l represents the l×l identity matrix, Λ = diag(h1, h2,..., h N ).

[0028] As a further implementation, the time triggering parameter is:

[0029]

[0030] where

[0031] where 0 < w0 < 1, 0 < β < 1, N is the total number of agents, and N i is the number of neighbors of agent i.

[0032] As an alternative embodiment, each agent broadcasts its own trigger time to downstream neighbor agents.

[0033] A multi-agent self-trigger control system, comprising:

[0034] A calculation module, configured to calculate, for each agent, the distance to the next trigger time based on the information of its own current moment and the information of upstream neighbor agents, and use the calculated trigger time as the pre-trigger time;

[0035] A judgment module, configured to update the calculation of the pre-trigger time if another upstream neighbor triggers before the pre-trigger time, otherwise, perform trigger control on its own agent using the current pre-trigger time as the trigger time;

[0036] The calculation process of the trigger time incorporates event trigger parameters and time trigger parameters.

[0037] A computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the steps in the above method.

[0038] An electronic device, comprising a memory, a processor, and computer instructions stored on the memory and running on the processor, which, when run by the processor, complete the steps in the above method.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] The present invention innovatively proposes a self-trigger mechanism / mode by combining time and events, calculates the next trigger time of each agent based on the information of itself and neighbor agents, uses the calculated value as the pre-trigger time, and when the time reaches the pre-trigger time and no other upstream neighbor has triggered, this pre-trigger time becomes the real trigger time for triggering. This self-trigger mechanism is based on a logical framework that combines event triggering and time triggering, and takes the longer time interval of the two as the final time interval, reducing the trigger frequency and avoiding the Zeno phenomenon.

[0041] The self-trigger mechanism / mode provided by the present invention avoids continuous communication and calculation comparison. The communication mode adopted is the "broadcast - receive" mode, broadcasting to downstream agents after triggering, without the need for downstream agents to continuously communicate, saving a large amount of communication resources and calculation resources, enabling the system to converge faster under the premise of saving resources, and considering the asymmetric communication mode, thus having a wider applicability and being more in line with engineering practice.

[0042] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following specifically enumerates preferred embodiments and, in conjunction with the accompanying drawings, makes a detailed description as follows. Description of the Drawings

[0043] The accompanying drawings forming a part of this invention are used to provide a further understanding of the invention. The schematic embodiments of the invention and their descriptions are used to explain the invention and do not unduly limit the invention.

[0044] Figure 1 is a flowchart of a multi-agent self-triggered control method for an embodiment;

[0045] Figure 2 is a structural diagram of a multi-agent self-triggered control system for an embodiment;

[0046] Figure 3 is a schematic diagram of neighbor agents for an embodiment;

[0047] Figure 4 is a structural diagram of an electronic device for an embodiment. Detailed implementation manners

[0048] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0049] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further descriptions of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0050] It should be noted that the terms used herein are only for describing specific implementation manners and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0051] In the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0052] Embodiment 1

[0053] As mentioned in the background art, using event triggering alone cannot avoid high energy consumption or even falling into the Zeno phenomenon. If time triggering is used alone to achieve similar performance, the time interval needs to be set small enough to cause a relatively large number of triggers.

[0054] To solve this problem, this embodiment provides a multi-agent self-triggered control method, as Figure 1 shown, including the following steps:

[0055] S1. Each agent calculates the distance to the next trigger moment based on its own current information and the information of its upstream neighbor agents, and uses the calculated trigger moment as the pre-trigger moment;

[0056] S2. If a trigger occurs in another upstream neighbor before the pre-trigger time, the pre-trigger time is updated and calculated; otherwise, the trigger control of the own agent is performed using the current pre-trigger time as the trigger time;

[0057] The trigger moment calculation process incorporates event trigger parameters and time trigger parameters.

[0058] The method provided in this embodiment, hereinafter referred to as the self-triggering mechanism, combines event triggering and time triggering, and takes the longer time interval of the two as the final time interval.

[0059] The above method can be applied to the asymmetric communication mode of a generalized heterogeneous multi-agent system. In this communication mode, although two-way communication is still maintained between the agents, the system status and other information with a large amount of data obtained by differential equations are transmitted in one direction. The channel occupied by the transmission of large-scale data information is reduced to half of the original, and the intermittent small data trigger signal is still transmitted in two directions. At this time, the neighbor agent j of agent i does not necessarily have a neighbor agent. In the description of this embodiment, upstream and downstream neighbors are used to make a special distinction.

[0060] like Figure 3 As shown, for agent 0, there is no upstream neighbor agent, but there are two downstream neighbor agents, namely agent 2 and agent 4;

[0061] Agent 2 has an upstream neighbor, agent 0, and a downstream neighbor, agent 1. Agent 4 has an upstream neighbor, agent 0, and a downstream neighbor, agent 3. Agent 3 has no downstream neighbor agents.

[0062] Based on this, the flow of large amounts of data is modeled as another network topology diagram, recorded as Assume that it is time-invariant and contains a directed spanning tree with node 0 as the root, and its edge weights are denoted by

[0063] A generalized heterogeneous multi-agent system is formed between the agents, which is equipped with an external system compensator. The consistency error obtained by intermittent adjustment is used through a trigger mechanism to compensate for the external system state.

[0064] The generalized heterogeneous multi-agent system is described as follows:

[0065]

[0066] in, Represents the state, controller and error output of the i-th agent respectively; A i , B i , C i , D i , F i , P i , is a matrix of appropriate dimension, It is assumed that the system is strongly stabilizable and the regulator equation of the following form has a solution to (Π i ,Γ i ):

[0067]

[0068] v is generated by an external system: in does not have a negative real part and is known in advance by each agent.

[0069] The purpose of this invention is to realize a multi-agent system with a communication topology G(t) as shown in (1) and design a controller u i The error output of the system converges to 0 for any initial state, and when the external system state v≡0, the closed-loop system is asymptotically stable.

[0070] Among them, generalized, that is, the expression on the left side of the system (1) contains a generalized matrix E. When E is not a generalized matrix, this expression can be converted to the general case through matrix inversion. Second, heterogeneous, that is, each agent has a different system form. Specifically, ABCD in (1) all contain subscripts, which poses a greater challenge to controller design. The controller provided in this embodiment can handle both generalized and heterogeneous cases in the presence of discontinuous communication and discontinuous triggered computation.

[0071] A detailed introduction is given below.

[0072] First, each agent has a controller that can additionally ensure that the system is regular and pulse-free, and the achievement of this goal is reflected in the selection of controller parameters.

[0073] The controller of agent i is as follows:

[0074]

[0075] in, represents the controller dynamics of agent i, represents the kth triggering moment of agent i. i is the consistency error of the external system reference signal. Controller gain K 1i ,K 2i Designed for K 2i =Γ i -K1i Π i , K 1i To ensure that the closed-loop system is regular, stable, and pulse-free, the selection method can refer to the content given in Algorithm 8.1.1 of Duan Guangren's book "Generalized System Theory".

[0076] in, Λ=diag(h1,h2,…,h N ), h=max 1≤i≤N h i ,h i for Here z is the compensator dynamics, parameter μ represents the compensator gain, and λ min (·) represents the minimum eigenvalue of the matrix, is the Laplacian matrix of the graph.

[0077] The compensator is used to estimate the external system dynamics and its structure is It can be seen that the compensator dynamics are mainly determined by its structure and triggering mechanism. The self-triggering mechanism provides a detailed model for each intelligent controller to calculate the distance to the next trigger time based on the current time information and neighbor information. The subsequent calculation formula provides the design scheme of the variables designed in the algorithm.

[0078] Before presenting the algorithm, relevant symbols and definitions are first given.

[0079] New letters and expressions involved in the algorithm and subsequent calculations: and Merge agent i in The triggering moments of all upstream neighbors in the sequence are a sequence, and the combined sequence is recorded as and define ":=" means defined as, and j∈{1,2,...,p}, p is the total number of upstream neighbor agents triggered, where is the last triggering moment of agent i, Indicates the next trigger moment.

[0080] The algorithm involves the pre-trigger time, which means that it is currently impossible to determine whether the value is the next trigger time of agent i, because its next trigger time is affected by whether its neighbors trigger. For any agent, it is impossible to predict whether its upstream neighbors will trigger in the future. It is triggered before the moment, so the calculation result here cannot be used as the trigger moment. It can only be named the "possible trigger moment", that is, the pre-trigger moment. The real trigger moment is when no other upstream neighbors are triggered when the time reaches the pre-departure moment, then this pre-trigger moment becomes the real trigger moment.

[0081] Taking the agent i as the execution object, the self-triggering algorithm process is given as follows:

[0082] Specifically, the algorithm includes the following steps

[0083] Step (1): Calculate [[ID=,10]] The current iteration number p = 0;

[0084] Step (2): Read the received from the upstream neighbors

[0085] Step (3): Calculate Obtain the pre-trigger moment

[0086] Step (4): If the current moment If an upstream neighbor is triggered, the current iteration number is incremented by one and recalculated Obtain the pre-trigger moment

[0087] If no upstream neighbor is triggered, then Trigger and broadcast to downstream neighbors

[0088] Here

[0089] When p ≠ 0:

[0090] Among them

[0091]

[0092] b i (t) does not change continuously with time. For agent i, its change moment depends on the trigger moment of the upstream neighbor agent in .

[0093] Among them, σ i represents the event trigger parameter, represents the time trigger parameter.

[0094] The event trigger parameter σ i is designed as follows:

[0095] 0 < w0 < 1, 0 < β < 1, N irepresents the upstream neighbor set of agent i, Θ′ is a positive definite solution to the following matrix equation:

[0096]

[0097] Time trigger parameters The design is as follows:

[0098] in Here λ max (·) represents the maximum eigenvalue of the matrix, and α represents the time trigger parameter.

[0099] The event trigger parameter σ is used to measure the trigger difficulty. The larger the value, the easier it is to trigger. It represents the time interval that is directly obtained, that is, the timing starts at the kth trigger moment and passes The next trigger moment is obtained after the time period.

[0100] Note that the above design is based on This condition is evolved, and the specific steps are shown after the validity demonstration. There are: (σ i +1)||e i (t)||≤σ i ||q i (t)+e i (t)||≤σ i ||q i (t)||+σ i ||e i (t)||i.e. Established is || e i (t)||≤σ i ||q i (t)|| is a necessary condition. Therefore, if ||e i (t)||≤σ i ||q i (t)|| is a more conservative method to design the controller as a trigger condition. i (t)||≤σ i ||q i (t)|| condition is used to demonstrate the effectiveness. If the problem can be solved with this condition, then the algorithm flow given in this embodiment is also valid.

[0101] First, according to the strong stabilization of the system, we know that there exists a matrix pair (Q i ,R i ) and K1i Satisfies two equations: Among them A i1 Located in the left half plane.

[0102] Here A i1 and B′ i2 No actual meaning, B′ i2 for The lower left corner block of The lower right block dimension is

[0103] make η i =z i -v, the closed-loop system can be rewritten as follows:

[0104]

[0105] in, Rewrite the above form into a compact form, let γ1 = col(γ 11 ,…,γ N1 ),γ2=col(γ 12 ,…,γ N2 ),η=col(η1,…,η N ), e=col(e1,…,e N ), have:

[0106]

[0107] Obviously, when the slow subsystem converges, the fast subsystem converges synchronously, so only the slow subsystem needs to be studied.

[0108] Let x c =col(γ1,η), C i R i +D i K 1i R i =[C i1 ,C i2 ], The following augmented form is obtained:

[0109]

[0110] in

[0111] Note that if Right now at the same time Thus there is Then

[0112] In particular, when \(v\equiv0\), the asymptotic stability of the original system still holds. According to the definition of the output regulation problem, the original problem is solved at this time. Therefore, the asymptotic stability of system (6) is a sufficient condition for the solution of the original problem.

[0113] Next, the stability analysis of system (6) is carried out.

[0114] First, prove that where \(q = col(q_1,q_2,\cdots,q N );

[0115] Calculate its derivative

[0116] Next, use the proof by contradiction. Assume that there exists such that

[0117] Let Then There is where \(i = 1,2,\cdots,N\).

[0118] According to There is It can be seen that Furthermore, it can be obtained that:

[0119]

[0120] Next, according to whether the agent \(i'\) has satisfied the time-triggering condition at time \(t'\), it is divided into the following two cases for discussion.

[0121] (1) If the time-triggering condition has been satisfied but the event-triggering mechanism has not been satisfied yet, then From the event-triggering condition, it can be obtained that:

[0122]

[0123] Contradicting So this case does not hold.

[0124] (2) If the time-triggering condition has not been reached, then Then according to the comparison lemma, it can be known that:

[0125] always holds.

[0126] Substitute into There is

[0127] Note that At monotonically increasing, and Depend on The definition of This is obviously contradictory, so this situation does not hold.

[0128] In summary, the original hypothesis is not established, that is, there is no make

[0129] The stability of (6) is proved below.

[0130] Consider the Lyapunov function Here Ξ=diag(Ξ1,Ξ2), Ξ1 is a positive definite solution to the following equation: in Being in the left half plane ensures that the equation has a solution. Note that this equation is equivalent to So the time derivative of V is: and Λ=diag(h1,h2,…,h N ) and diag(ε1,…,ε N ) is a diagonal matrix and is therefore commutative, so we have

[0131] Consider the sets S1(t) and S2(t), where

[0132] Here S1(t) represents the set of agents that have met the time trigger condition but have not yet been triggered. Therefore, the event trigger condition of these agents is not met. Therefore, for i∈S1(t), ‖e i ‖≤σ i ‖q i ‖. According to the inequality It can be concluded that:

[0133]

[0134] For i∈S2(t), And the comparison lemma shows that:

[0135] then thereby Substitution Available in

[0136] Note that the upper left block is in the left half plane, which means that the first condition of negative definiteness is satisfied according to Schur's complement lemma. Now we prove the second condition. Scaling Φ and substituting Ξ2,Ω, we get:

[0137]

[0138] Record According to the Schur complement lemma, Ψ < 0. According to the Lyapunov stability theorem, V > 0, and There is Combined with the foregoing analysis, it is known that the output regulation problem of the generalized heterogeneous linear multi-agent system is solved. At the same time, from It is known that the proposed triggering mechanism will not bring about the Zeno phenomenon. Thus, it can be seen that based on ||e i (t)|| ≤ σ i ||q i (t)|| condition, the obtained system meets the conditions, so its more stringent condition The obtained self-triggering mechanism is also effective.

[0139] The following presents a self-triggering design method based on .

[0140] First, note that Thus, there is:

[0141]

[0142] It can be seen that the above formula gives an upper bound of ‖e i ‖ that monotonically increases with time. Therefore, when , the next self-triggering moment can be conservatively estimated by the following formula because it essentially uses an expression greater than ‖e i ‖ to establish an equation to solve for time t, so as to ensure that the formula always holds.

[0143] Where

[0144] Regarding the above formula as an equation about t and denoting its solution as We can obtain:

[0145]

[0146] In particular:

[0147] This method combines time and events and is improved to a self-triggering mode, avoiding continuous communication and computational comparison, saving a large amount of communication resources and computational resources. It can make the system converge faster under the premise of saving resources, and considers the asymmetric communication mode, so it has a wider applicability and is more in line with engineering practice.

[0148] This method has a wide range of applications and can be applied to multi-agent systems, such as swarm robot systems composed of multiple robots, fleet systems composed of multiple self-driving vehicles, logistics systems composed of multiple delivery robots, etc. The followers of each multi-agent system can adjust their own status in real time through the multi-agent collaborative controller.

[0149] Example 2

[0150] A multi-agent self-triggering control system, such as Figure 2 As shown, including:

[0151] The calculation module is configured so that each agent calculates the distance to the next trigger moment based on its own current moment information and the information of the upstream neighboring agent, and uses the calculated trigger moment as the pre-trigger moment;

[0152] The judgment module is configured to update and calculate the pre-trigger time if another upstream neighbor is triggered before the pre-trigger time; otherwise, the current pre-trigger time is used as the trigger time to perform trigger control of the own agent;

[0153] The trigger moment calculation process incorporates event trigger parameters and time trigger parameters.

[0154] Embodiment 3

[0155] A computer-readable storage medium (Memory) is a memory device in an electronic device used to store programs and data. It is understood that the computer-readable storage medium herein can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The computer-readable storage medium provides storage space that stores the processing system of the electronic device.

[0156] Furthermore, the storage space also stores one or more instructions suitable for being loaded and executed by the processor. These instructions may be one or more computer programs (including program code). It should be noted that the computer-readable storage medium herein may be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage; optionally, it may be at least one computer-readable storage medium located remotely from the aforementioned processor.

[0157] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer-readable storage medium to implement the following process:

[0158] Each agent calculates the next triggering time based on its own current information and the information of its upstream neighbor agents, and uses the calculated triggering time as the pre-triggering time. If other upstream neighbors are triggered before the pre-triggering time, the pre-triggering time is updated and calculated. Otherwise, the current pre-triggering time is used as the triggering time to control the triggering of its own agent.

[0159] The trigger moment calculation process incorporates event trigger parameters and time trigger parameters.

[0160] Embodiment 4

[0161] An electronic device, such as Figure 4 As shown, the electronic device includes a processor 1001, a communication interface 1002, and a computer-readable storage medium 1003. The processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 may be connected via a bus or other means.

[0162] Among them, the communication interface 1002 is used to receive and send data, the computer-readable storage medium 1003 can be stored in the memory of the electronic device, the computer-readable storage medium 1003 is used to store a computer program, the computer program includes program instructions, and the processor 1001 is used to execute the program instructions stored in the computer-readable storage medium 1003.

[0163] The processor 1001 (or CPU (Central Processing Unit)) is the computing core and control core of the electronic device, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to implement corresponding method processes or corresponding functions.

[0164] The processor 1001 is configured to execute the following process:

[0165] Each agent calculates the next triggering time based on its own current information and the information of its upstream neighbor agents, and uses the calculated triggering time as the pre-triggering time. If other upstream neighbors are triggered before the pre-triggering time, the pre-triggering time is updated and calculated. Otherwise, the current pre-triggering time is used as the triggering time to control the triggering of its own agent.

[0166] The trigger moment calculation process incorporates event trigger parameters and time trigger parameters.

[0167] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0169] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0170] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0171] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made by those skilled in the art that fall within the spirit and principles of the present invention and do not require creative effort are intended to be within the scope of protection of the present invention.

Claims

1. A multi-agent self-triggered control method, characterized in that, It includes the following steps: Each agent calculates the distance to the next trigger moment based on its own information at the current moment and the information of the upstream neighbor agents. The calculated trigger moment is used as the pre-trigger moment. If another upstream neighbor triggers before the pre-trigger moment, the calculation of the pre-trigger moment is updated. Otherwise, the current pre-trigger moment is used as the trigger time for the trigger control of its own agent; The calculation process of the trigger moment incorporates an event trigger parameter and a time trigger parameter.

2. The multi-agent self-triggered control method according to claim 1, characterized in that The process by which each agent calculates the distance to the next trigger moment based on its own information at the current moment and the information of the upstream neighbor agents includes: The next trigger time of agent i is: Among them, the is the next triggering moment of the current iteration number p, is the previous triggering moment of agent i, is the time triggering parameter, is the event triggering time interval parameter.

3. The multi-agent self-triggered control method according to claim 2, characterized in that, The event trigger time interval parameter It is calculated based on the event trigger parameter, the information of the upstream neighbor agent, and the information of agent i.

4. The multi-agent self-triggered control method according to claim 3, characterized in that, The event trigger time interval parameter is as follows: When p is zero, When p is not zero, Among them, is the external system matrix, which has no negative real part and is known to each agent in advance. σ i represents the event-triggering parameter. For agent i during the p-th iteration, in the sequence obtained by merging all the triggering moments of its upstream neighbors in the time interval. μ represents the compensator gain. To model the flow of a large amount of data as another network topology graph, it is denoted as Assume it is time-invariant and contains a directed spanning tree with node 0 as the root, and its edge weights are denoted as Here is the connection weight between agent i and node 0 in the graph below; is to merge all the triggering moments of agent i's upstream neighbors in into a sequence, and the merged sequence is denoted as and define and Here is the previous triggering moment of agent i, represents the next triggering moment.

5. The multi-agent self-triggered control method according to claim 3, characterized in that, Event trigger parameter σ i is as follows: where \(0 < w0 < 1\), \(0 < β < 1\), \(h=\max\) 1≤i≤N h i , h i is the \(i\)-th element of, where is the Laplacian matrix of the graph, N i represents the set of upstream neighbors of agent i, Θ′ is the positive definite solution of the following matrix equation: Denotes the maximum number of all intelligent agent neighbors, I l Denotes the l×l identity matrix, Λ = diag(h1, h2, …, h N ).

6. The multi-agent self-triggered control method according to claim 2, characterized in that, Time-triggered parameter is as follows: Among them where \(0 < w_0 < 1\), \(0 < \beta < 1\), \(N\) is the total number of agents, and \(N\) i is the number of neighbors of agent \(i\).

7. The multi-agent self-triggered control method according to claim 1, characterized in that Each agent broadcasts its own trigger time to the downstream neighbor agents.

8. A multi-agent self-triggered control system, characterized in that It includes: A calculation module, configured to calculate the distance to the next trigger moment by each agent according to its own information at the current moment and the information of the upstream neighbor agents, and use the calculated trigger moment as the pre-trigger moment; A judgment module, configured to update the calculation of the pre-trigger moment if another upstream neighbor triggers before the pre-trigger moment, otherwise, use the current pre-trigger moment as the trigger time for the trigger control of its own agent; The calculation process of the trigger moment incorporates an event trigger parameter and a time trigger parameter.

9. A computer-readable storage medium, characterized in that, For storing computer instructions, when the computer instructions are executed by a processor, the steps in the above method are completed.

10. An electronic device, characterized in that, It includes a memory, a processor, and computer instructions stored on the memory and running on the processor. When the computer instructions are run by the processor, the steps in the above method are completed.

Images