A communication link-aware multi-agent dynamic event-triggered consensus method

Abstract

The application discloses a kind of communication link perception multi-agent dynamic event trigger consistency methods, comprising S1: the communication network topology graph of information exchange between multi-agent is established;S2: determine the dynamics equation of multi-agent system;S3: introduce adaptive control, design complete distributed consistency control input strategy based on event trigger;S4: the solution of algebraic riccati equation obtains feedback gain matrix;S5: based on Lyapunov theorem and channel capacity parameter, design communication link perception dynamic event trigger control strategy;S6: reach multi-agent consistency by complete distributed control input strategy and communication link perception dynamic event trigger control strategy;The communication link perception dynamic event trigger mechanism of the application is completely distributed, through adaptive coupling gain, control input adaptive adjustment, without relying on global information related to network topology, realize and apply in large networked multi-agent system.

Description

Technical Field

[0001] This invention relates to the field of information interaction technology, and specifically to a method for coherent dynamic event triggering of multiple agents with communication link awareness. Background Technology

[0002] In recent years, cooperative control of multi-agent systems (MASs) has gained widespread attention in industry and academia, with applications including drone swarm flight, mobile robots for express delivery sorting, satellite attitude synchronization, wireless sensor network estimation, and power allocation in smart grids. Consistency, as the foundation of cooperative control, has become a hot research topic in the field of artificial intelligence control.

[0003] Multi-agent system consensus refers to the ability of a group of agents to interact with each other through a designed consensus protocol, ultimately achieving state convergence and task completion.

[0004] Typically, multi-agent systems consist of multiple tiny embedded systems. The agents themselves have limited resources such as computing power, communication bandwidth, and battery capacity. To reduce the amount of communication in the network and the update frequency of the controller, event-triggered control technology has been proposed.

[0005] Under the event-triggered mechanism, each agent sends local information to its neighboring nodes only when needed. That is, communication and control input updates between agents only occur when the triggering condition is met. Obviously, event-triggered control avoids real-time updates of the controller, which can effectively save communication and computing resources, thereby extending the lifespan of the agents.

[0006] To increase the triggering time interval, reduce the frequency of event triggering, reduce the amount of unnecessary information transmission, and further reduce the communication burden, dynamic event triggering mechanisms have gained attention.

[0007] The triggering conditions of the dynamic event-triggered control strategies proposed in the prior art are fixed, the triggering frequency cannot be intelligently adjusted, and the influence of the actual network state is not considered. The triggering mechanism cannot intelligently adjust the information interaction frequency between agents according to the network state, which leads to the system's inefficient and unreasonable use of communication resources. At the same time, the existing technologies require knowledge of global information related to the network topology when designing control strategies, which makes them difficult to implement and apply in large-scale networked multi-agent systems. Summary of the Invention

[0008] The purpose of this invention is to provide a communication link-aware multi-agent dynamic event triggering consistency method. This method considers channel capacity in the triggering conditions, enabling the event triggering mechanism to intelligently adjust according to changes in communication resources. When communication network performance is poor, the frequency of information interaction between agents is reduced, further alleviating the communication burden and saving communication resources. When communication network performance is good, the frequency of information interaction between agents is appropriately increased, achieving better control performance and enabling the multi-agent system to converge faster. Furthermore, the system control scheme design does not depend on network-related global information, achieving true full distribution.

[0009] The objective of this invention can be achieved through the following technical solutions:

[0010] A communication link-aware multi-agent dynamic event-triggered consensus method includes the following steps:

[0011] S1: Establish a communication network topology diagram for information exchange among multiple agents;

[0012] S2: Determine the dynamic equations of the multi-agent system;

[0013] S3: Introduce adaptive control and design a fully distributed consistency control input strategy based on event triggering;

[0014] S4: Solve the algebraic Riccati equation to obtain the feedback gain matrix;

[0015] S5: Based on Lyapunov's theorem and channel capacity parameters, design a communication link-aware dynamic event triggering control strategy;

[0016] S6: Achieve multi-agent consensus through a fully distributed control input strategy and a communication link-aware dynamic event-triggered control strategy.

[0017] As a further aspect of the present invention: In S1, the network topology for communication and exchange between multiple agents is undirected, represented by a graph. express;

[0018] The node set consists of N intelligent agents. All communication links that enable information exchange between intelligent agents constitute the edge set.

[0019] If node i can obtain information from node j, then node j is a neighbor node of node i, and all neighbors of node i constitute a set. A = [a ij ] N×N It is an adjacency matrix;

[0020] When (i,j)∈ε, then a ij =a ji=1, otherwise a ij =0.

[0021] As a further aspect of the present invention: In S2, based on a system composed of N agents, the kinematic equation of agent i is expressed as follows:

[0022]

[0023] Where, x i (t)∈R n u represents the state of the agent. i (t)∈R m w represents the control input of the intelligent agent. i (t)∈R n Given bounded external disturbances, (A,B) is stable, and for any external disturbance, there exists a matrix F∈R. m×n The condition D = BF is satisfied.

[0024] As a further aspect of the present invention: In S3, the process of setting the fully distributed consensus control input strategy is as follows:

[0025] Define a state estimate; when an update is triggered, the agent's state estimate is...

[0026]

[0027] Define the combined variable as

[0028]

[0029]

[0030] At this time, the state error is

[0031]

[0032] For each agent, a fully distributed adaptive control input is obtained.

[0033]

[0034]

[0035] Where K is the control gain to be designed, and F is the constant matrix to be determined;

[0036] For agent i, This is called the trigger time, h i (t) represents the adaptive coupling gain corresponding to agent i, h i (0)>0.

[0037] As a further aspect of the present invention: in S4, the solution P of the algebraic Riccati equation is greater than 0;

[0038] PA+A T P-PBB T P+I n =0;

[0039] Let the feedback gain matrix be K = B. T P, Γ = PBB T P.

[0040] As a further aspect of the present invention: In S5, the dynamic event triggering conditions for communication link awareness are obtained based on Lyapunov's theorem and channel capacity parameters, thus obtaining the next triggering time for each agent.

[0041]

[0042]

[0043] In the formula, η i (0)>0, β i >0, θ≥2, α i =δ i / γ i ,0<δ i <1, Γ=PBB T P, ρ and v are positive constants. This refers to the channel capacity.

[0044] The beneficial effects of this invention are as follows: By considering channel capacity in the triggering conditions, the frequency of information interaction between intelligent agents can be intelligently adjusted according to the actual network state. From the perspective of communication resources, based on event triggering and combined with real-time network status, the triggering frequency is increased more frequently when the bandwidth is idle to obtain better control performance; when the bandwidth is busy, the triggering frequency is reduced to alleviate the communication burden. From the perspective of system convergence, appropriately increasing the triggering frequency helps to quickly stabilize the system, that is, the system convergence speed is faster; reducing the triggering frequency can effectively save communication resources.

[0045] Furthermore, this communication link-aware dynamic event triggering mechanism is completely distributed. Through adaptive coupling gain, the control input can be adaptively adjusted without relying on global information related to the network topology, enabling the control scheme to be implemented and applied in large-scale networked multi-agent systems. Attached Figure Description

[0046] The invention will now be further described with reference to the accompanying drawings.

[0047] Figure 1 This is a flowchart of the present invention;

[0048] Figure 2 This is a communication topology diagram of the multi-agent system of the present invention;

[0049] Figure 3 This is the velocity state evolution curve of the agent under the fully distributed communication link sensing dynamic event triggering mechanism of this invention;

[0050] Figure 4 This is a schematic diagram of the adaptive coupling gain of the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figures 1-2 As shown, this invention provides a communication link-aware multi-agent dynamic event triggering consensus method, comprising the following steps:

[0053] S1: Establish a communication network topology diagram for information exchange among multiple agents;

[0054] S2: Determine the dynamic equations of the multi-agent system;

[0055] S3: Introduce adaptive control and design a fully distributed consistency control input strategy based on event triggering;

[0056] S4: Solve the algebraic Riccati equation to obtain the feedback gain matrix;

[0057] S5: Based on Lyapunov's theorem and channel capacity parameters, design a communication link-aware dynamic event triggering control strategy;

[0058] S6: Achieve multi-agent consensus through a fully distributed control input strategy and a communication link-aware dynamic event-triggered control strategy.

[0059] In S1, the network topology for communication and exchange among multiple agents is undirected, represented by a graph. express;

[0060] The node set consists of N intelligent agents. All communication links that enable information exchange between intelligent agents constitute the edge set. This means that node i can obtain information from node j. At this time, node j is a neighbor node of node i, and all the neighbors of node i constitute a set. A = [a ij ] N×N It is an adjacency matrix;

[0061] When (i,j)∈ε, then a ij =a ji =1, otherwise a ij =0.

[0062] In S2, the process of determining the dynamic equations of the multi-agent system is as follows:

[0063] Design a system consisting of N agents, where the kinematic equations of agent i are expressed as follows:

[0064]

[0065] Where, x i (t)∈R n u represents the state of the agent. i (t)∈R m w represents the control input of the intelligent agent. i (t)∈R n Given bounded external disturbances, (A,B) is stable, and for any external disturbance, there exists a matrix F∈R. m×n The condition D = BF is satisfied.

[0066] In S3, the process of designing an event-triggered fully distributed consistency control input strategy is as follows:

[0067] Define the state estimate, that is, the state estimate of the agent when an update is triggered.

[0068]

[0069] Define the combined variable as

[0070]

[0071]

[0072] At this time, the state error is

[0073]

[0074] For each agent, a fully distributed adaptive control input was designed.

[0075]

[0076]

[0077] K is the control gain to be designed, and F is the constant matrix to be determined;

[0078] For agent i, This is called the trigger time, h i (t) represents the adaptive coupling gain corresponding to agent i, h i (0)>0.

[0079] In S4, the feedback gain matrix is ​​the solution to the algebraic Riccati equation (ARE) P>0;

[0080] PA+A T P-PBB T P+I n =0 (7)

[0081] Let the feedback gain matrix be K = B. T P, Γ = PBB T P.

[0082] In S5, the dynamic event triggering conditions for communication link awareness are obtained based on Lyapunov's theorem and channel capacity parameters, thus obtaining the next triggering time for each agent.

[0083]

[0084]

[0085] In the formula, η i (0)>0, β i >0, θ≥2, α i =δ i / γ i ,0<δ i <1, Γ=PBB T P, ρ and v are positive constants. The channel capacity is used as the triggering condition for the dynamic event triggering mechanism, which is related to the channel capacity.

[0086] In S6, the stability and consistency of the system are proved based on Lyapunov's theorem. The proof process is as follows:

[0087] Define x(t) = col(x1(t), ..., x N (t)), and e(t)=col(e1(t),…,e N (t)), q(t)=col(q1(t),…,q N (t)), and A multi-agent system achieves consistency if and only if q(t)→0 and t→∞;

[0088]

[0089] Lyapunov function

[0090]

[0091] in, h0 is a positive constant;

[0092] From (8), we can know that η i Since W(t)≥0, W(t)>0 holds true.

[0093] According to q(t) k )=q(t)+e(t), differentiating with respect to V(t) gives

[0094]

[0095] According to Young's inequality theorem, we have

[0096]

[0097] From the properties of the Laplacian matrix, we know that

[0098]

[0099] Substituting (12) and (13) into (11) yields

[0100]

[0101] According to Young's inequality theorem

[0102]

[0103] Substituting (15) into (14) yields

[0104]

[0105] definition have to

[0106]

[0107] Since the communication graph is an undirected topological graph, then

[0108]

[0109]

[0110] Final scaling to

[0111]

[0112] Here, 2≤λ2h0≤σ, σ≥2;

[0113] Substituting (8) and (20) into (10) yields

[0114]

[0115] in

[0116] According to the comparison theorem

[0117]

[0118] prove,

[0119] W(t) asymptotically converges to the following set

[0120]

[0121] Therefore, W(t)≥V(t)>0, hence q i (t) and h i (t) are all eventually consistent and bounded, indicating that the multi-agent system eventually achieves consistency.

[0122] To verify the effectiveness of the fully distributed communication link-aware dynamic event-triggered control strategy in this application, a multi-agent system with five agents was designed, and its network communication topology is shown in the figure below. Figure 1 As shown, the initial state of each agent is random, where

[0123]

[0124]

[0125] In the formula, ω0 = 0.001.

[0126] Let μ = 0.5, and obtain the feedback gain matrix K by solving the algebraic Riccati equation (7):

[0127]

[0128] Other parameters are defined as β. i =0.004, δ i =0.999, γ i =||PBB T P||=22.6329,π i =0.003, γ i =(2d i hi (t)+θ)||Γ||=44.6199, σ=2, where i=1,...,5, ρ=2, v=0.5.

[0129] like Figure 3 As shown, the velocity states of all agents eventually reached a consensus, as... Figure 4 As shown, the adaptive coupling gain corresponding to each agent converges to an effective steady-state value. The multi-agent system can adaptively adjust the control input, effectively avoiding the use of global information to trigger the control strategy.

[0130]

[0131]

[0132] Table 1

[0133] As shown in Table 1, the proposed communication link-aware dynamic event triggering mechanism can balance communication resources and convergence speed. When the communication network performance is good, the information interaction frequency between agents is increased to obtain a faster convergence speed. When the communication network performance is poor, the information interaction between agents is reduced to further save communication resources.

[0134] One of the core points of this invention is the proposal of a dynamic event-triggered consistency control scheme for a fully distributed communication link-aware multi-agent system. The proposed control strategy allows the frequency of information interaction between agents to be intelligently adjusted according to changes in the actual network state. When the network state is poor, the information interaction frequency is reduced, thereby alleviating the communication burden and further saving communication resources. When the network state is good, the state is updated more frequently, increasing the triggering frequency to obtain better control performance and accelerate the convergence speed of the system. The results show that the triggering frequency can be balanced between communication resources and convergence speed, demonstrating practical feasibility.

[0135] The second key point of this invention is that the communication link-aware dynamic event triggering mechanism is completely distributed. Through adaptive coupling gain, the control input can be adaptively adjusted without relying on global information related to the network topology, which enables the control scheme to be implemented and applied in large-scale networked multi-agent systems.

[0136] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A communication link-aware multi-agent dynamic event triggering consistency method, characterized in that, Includes the following steps: S1: Establish a communication network topology diagram for information exchange among multiple agents; S2: Determine the dynamic equations of the multi-agent system; S3: Introduce adaptive control and design a fully distributed consistency control input strategy based on event triggering; S4: Solve the algebraic Riccati equation to obtain the feedback gain matrix; S5: Based on Lyapunov's theorem and channel capacity parameters, design a communication link-aware dynamic event triggering control strategy; S6: Achieve multi-agent consensus through a fully distributed control input strategy and a communication link-aware dynamic event-triggered control strategy; In S1, the network topology for communication and exchange among multiple agents is undirected, represented by a graph. express; The node set consists of N intelligent agents. All communication links that enable information exchange between intelligent agents constitute the edge set. ; node Able to start from nodes To obtain information, the node For nodes Neighboring nodes, nodes All neighbors constitute a set It is an adjacency matrix; when ,but ,otherwise ; In S2, based on a system composed of N intelligent agents, the intelligent agents... The kinematic equations of are expressed as follows: in, Indicates the state of the agent. Indicates the control input of the intelligent agent. It is a bounded external interference. It is stable, and there exists a matrix that is resistant to external disturbances. satisfy ; In S3, the process of setting the fully distributed consensus control input strategy is as follows: Define a state estimate; when an update is triggered, the agent's state estimate is... Define the combined variable as At this time, the state error is For each agent, a fully distributed adaptive control input is obtained. = in, For the control gain that needs to be designed, Let be the constant matrix to be determined. , This is the next trigger moment for the agent; For intelligent agents , This is called the trigger moment. For the corresponding intelligent agent Adaptive coupling gain, .

2. The communication link-aware multi-agent dynamic event triggering consistency method according to claim 1, characterized in that, In S4, the solution to the algebraic Riccati equation ; Let the feedback gain matrix be... .

3. The communication link-aware multi-agent dynamic event triggering consistency method according to claim 1, characterized in that, In S5, the dynamic event triggering conditions for communication link awareness are obtained based on Lyapunov's theorem and channel capacity parameters, thus obtaining the next triggering time for each agent. ; ； ； In the formula, , For positive integers, This refers to the channel capacity.

Images