A multi-agent cluster event-triggered coordination method for non-deterministic interference environment
By using a distributed adaptive event-triggered protocol, the communication and controller update problems of multi-agent systems in nondeterministic disturbance environments are solved, achieving efficient collaborative control under bandwidth-constrained conditions, extending system lifespan and broadening the application scope.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TONGJI UNIV
- Filing Date
- 2023-04-10
- Publication Date
- 2026-05-19
AI Technical Summary
Existing multi-agent systems rely on global information or continuous communication and controller updates when facing uncertain interference environments, resulting in high communication frequency, limited bandwidth, and shortened system lifespan, making them difficult to apply effectively in complex scenarios.
By adopting a distributed adaptive event-triggered protocol, the dynamic equations of a multi-agent system are constructed by acquiring agent state and uncertainty information, determining the event-triggered protocol and input controller, and realizing collaborative control among agents, thus avoiding continuous communication and controller updates.
It improves the robustness of multi-agent systems in environments with limited communication bandwidth and nondeterminism, reduces communication requirements and controller update frequency, and extends system lifetime. It is applicable to both leaderless and leaderless multi-agent systems.
Smart Images
Figure CN116520691B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of multi-agent control technology, and in particular to a multi-agent cluster event-triggered collaborative method for non-deterministic interference environments. Background Technology
[0002] In recent years, multi-agent systems have attracted widespread attention due to their broad application areas, including industry, transportation, agriculture, and astronomy. Compared to individual systems, multi-agent systems can be applied to more complex scenarios.
[0003] In multi-agent practice, completing tasks under various complex and demanding conditions is a pressing problem, and reducing the communication frequency between neighboring agents is a crucial issue. To achieve this, designing a reasonable event triggering mechanism to determine necessary communication moments is an effective solution. Existing solutions can be categorized into three types based on threshold setting methods: state-based thresholds, time-based thresholds, and hybrid thresholds; and into two types based on triggering conditions: static triggering conditions and dynamic triggering conditions.
[0004] In real-world scenarios, uncertainty is prevalent in various models and systems. Therefore, considering the existence of uncertainty and disturbances is a necessary condition for the practical application of multi-agent systems.
[0005] Many existing solutions rely on providing global information to overcome uncertainty. However, since global information is generally difficult to compute and use by local agents, these algorithms are hard to apply to real-world multi-agent systems. Many other solutions rely on continuous communication and continuous controller updates to achieve control objectives. However, the former is extremely vulnerable in bandwidth-constrained scenarios, and high-frequency information exchange is impractical; the latter can cause severe mechanical fatigue and reduce lifespan. Summary of the Invention
[0006] The purpose of this invention is to overcome the shortcomings of existing technologies that rely on providing global information or continuous communication and continuous updates to the controller to overcome uncertainty. These methods are difficult to implement, have limited application scenarios, and reduce the lifespan of the system. Therefore, this invention provides a multi-agent cluster event-triggered collaborative method for non-deterministic interference environments.
[0007] The objective of this invention can be achieved through the following technical solutions:
[0008] A multi-agent cluster event-triggered coordination method for nondeterministic interference environments, the method comprising the following steps:
[0009] Acquire information from each agent, construct a multi-agent system, and determine the state information of each agent based on the state input and uncertainty of each agent through the dynamic equations of the multi-agent system.
[0010] The estimation error is obtained, the event triggering protocol is determined based on the state information of the agent and the estimation error, and the control objective of the multi-agent system is determined based on the state information of the agent, wherein the control objective is to make the states of the multiple agents tend to be consistent.
[0011] Obtain the adaptive coupling gain information of the agent, and determine the input controller based on the adaptive coupling gain information of the agent;
[0012] Determine whether the state of the agent satisfies the event triggering protocol. If it does, neighboring agents communicate and update the input controller corresponding to the agent to achieve the control objective.
[0013] Furthermore, the dynamic equations of the multi-agent system are described as follows:
[0014]
[0015] Where, x i ∈R n Let u represent the state of the i-th agent. i ∈R p f represents the state input of the i-th agent. i (t,x i )∈R p This represents the uncertainty of the i-th agent and satisfies the constraints. A and B are existing constants, and A and B are known constant matrices that are dimensionally compatible.
[0016] Furthermore, the multi-agent system includes a leaderless multi-agent system and a multi-agent system with a leader, wherein the leader in the multi-agent system only sends signals to other agents, and the model of the leader is described as follows: Where A is a dimension-compatible known constant matrix, and x0 represents the state of the leader agent.
[0017] Furthermore, the control objective of the leaderless multi-agent system is described as follows:
[0018]
[0019] Where, x i (t) represents the state of the i-th agent at time t, x j (t) represents the state of the j-th agent at time t.
[0020] Furthermore, the control objective of the multi-agent system with a leader is described as follows:
[0021]
[0022] Where, x i x0(t) represents the state of the i-th agent at time t, and x0(t) represents the state of the leader agent at time t.
[0023] Furthermore, the input controller of the leaderless multi-agent system is described as follows:
[0024]
[0025] Where K and Γ are the feedback gain matrices to be determined; d i (t), e i (t) is the initial value d i (0)≥1、e i (0)≥1 adaptive coupling gain, in This represents the instant when the k-th event is triggered by the i-th agent; For consensus error, where There are also in make in Represents the state estimate for agent j; g(s) is a nonlinear function, where s∈R p hour,
[0026] K and Γ are obtained by solving K = -B T Q and Γ = QBB T Q is obtained, where the inverse matrix P of Q is equal to Q. -1 It is a linear matrix inequality AP+PA T -BB T For solutions with a value less than 0, B is a known constant matrix that is dimension-compatible.
[0027] Furthermore, the input controller of the multi-agent system with a leader is described as follows:
[0028]
[0029] In the formula, K and Γ are the feedback gain matrices to be determined; d i (t), e i (t) is the initial value d i (0)≥1、e i (0)≥1 adaptive coupling gain, in This represents the instant when the k-th event is triggered by the i-th agent; To track errors, there are also in,
[0030] Furthermore, the event triggering protocol of the leaderless multi-agent system is described as follows:
[0031]
[0032] In the formula, This represents the instant when the (k+1)th event is triggered by the i-th agent. γ i i = 1, ..., N, θ i i = 1, ..., 4 are positive constants; For estimation error; ε i (t) is an internal variable, satisfying... Where ε i (0)>0, k i >0, σ i >0, δ i >0.
[0033] Furthermore, the event triggering protocol of the multi-agent system with a leader is described as follows:
[0034]
[0035]
[0036] In the formula, This represents the instant when the (k+1)th event is triggered by the i-th agent. γ i i = 1, ..., N, θ i ,i=1,…,4 are positive constants; ε i (t) is an internal variable, satisfying... Where ε i (0)>0, k i >0, σ i >0, δ i >0.
[0037] Furthermore, the communication topology between the multi-agent system is an undirected graph, and the agents include robots, drones, unmanned vehicles, and unmanned ships.
[0038] Compared with the prior art, the present invention has the following advantages:
[0039] 1. This solution proposes a novel fully distributed adaptive event-triggered protocol that can be used in environments with limited communication bandwidth and uncertain interference, thereby improving the robustness of multi-agent systems and broadening their application scope.
[0040] 2. This solution avoids continuous communication and controller updates between adjacent intelligent agents, reduces communication bandwidth requirements, lowers the update frequency of the system controller, saves system energy, and avoids wear and tear on hardware caused by high-frequency controller updates, reducing mechanical fatigue and effectively extending service life.
[0041] 3. This solution considers both the presence and absence of a leader in a multi-agent system and designs corresponding event-triggered control protocols for each, thus having a wider range of applicability;
[0042] 4. This scheme realizes distributed cooperative control of a multi-agent system under the constraints of control updates and interactions between neighboring agents, in the presence of matching uncertainty.
[0043] 5. This solution achieves discrete control updates for each agent, discrete-time communication between neighboring agents, and a fully distributed controller simultaneously without requiring any global information, thus overcoming the shortcomings of existing technologies in this regard. Attached Figure Description
[0044] Figure 1 This is a schematic diagram showing the connection between the multi-agent system and the input controller of the present invention;
[0045] Figure 2 This is the communication topology of a leaderless multi-agent system in an embodiment of the present invention;
[0046] Figure 3 This refers to the consensus error of the leaderless multi-agent system in this embodiment of the invention.
[0047] Figure 4 This refers to the event triggering time of the leaderless multi-agent system in this embodiment of the invention.
[0048] Figure 5 This is the communication topology of a multi-agent system with a leader and followers in an embodiment of the present invention;
[0049] Figure 6 The tracking error of the multi-agent system with leader and follower in the embodiments of the present invention.
[0050] Figure 7 This refers to the event triggering moment of a multi-agent system with a leader and followers in an embodiment of the present invention. Detailed Implementation
[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.
[0052] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.
[0053] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.
[0054] Example 1
[0055] This embodiment provides a multi-agent cluster event-triggered collaborative method for non-deterministic interference environments, such as... Figure 1 As shown, the method includes the following steps:
[0056] Acquire information from each agent, construct a multi-agent system, and determine the state information of each agent based on the state input and uncertainty of each agent through the dynamic equations of the multi-agent system.
[0057] The estimation error is obtained, the event triggering protocol is determined based on the state information of the agent and the estimation error, and the control objective of the multi-agent system is determined based on the state information of the agent, wherein the control objective is to make the states of the multiple agents tend to be consistent.
[0058] Obtain the adaptive coupling gain information of the agent, and determine the input controller based on the adaptive coupling gain information of the agent;
[0059] Determine whether the state of the agent satisfies the event triggering protocol. If it does, neighboring agents communicate and update the input controller corresponding to the agent to achieve the control objective.
[0060] The dynamic equations of a multi-agent system are described as follows:
[0061]
[0062] Where, x i ∈R n Let u represent the state of the i-th agent. i∈R p f represents the state input of the i-th agent. i (t,x i )∈R p This represents the uncertainty of the i-th agent and satisfies the constraints. A and B are existing constants, and A and B are known constant matrices that are dimensionally compatible.
[0063] Multi-agent systems include leaderless multi-agent systems and multi-agent systems with a leader. In a multi-agent system, the leader only sends signals to the other agents. The leader's model is described as follows: Where A is a dimension-compatible known constant matrix, and x0 represents the state of the leader agent.
[0064] The control objective of a leaderless multi-agent system is described as follows:
[0065]
[0066] Where, x i (t) represents the state of the i-th agent at time t, x j (t) represents the state of the j-th agent at time t.
[0067] The control objective of a multi-agent system with a leader is described as follows:
[0068]
[0069] Where, x i x0(t) represents the state of the i-th agent at time t, and x0(t) represents the state of the leader agent at time t.
[0070] The input controller of a leaderless multi-agent system is described as follows:
[0071]
[0072] Where K and Γ are the feedback gain matrices to be determined; d i (t), e i (t) is the initial value d i (0)≥1、e i (0)≥1 adaptive coupling gain, in This represents the instant when the k-th event is triggered by the i-th agent; For consensus error, where There are also in make in Represents the state estimate for agent j; g(s) is a nonlinear function, where s∈R p hour,
[0073] K and Γ are obtained by solving K = -B T Q and Γ = QBB T Q is obtained, where the inverse matrix P of Q is equal to Q. -1 It is a linear matrix inequality AP+PA T -BB T For solutions with a value less than 0, B is a known constant matrix that is dimension-compatible.
[0074] The input controller of a multi-agent system with a leader is described as follows:
[0075]
[0076] In the formula, K and Γ are the feedback gain matrices to be determined; d i (t), e i (t) is the initial value d i (0)≥1、e i (0)≥1 adaptive coupling gain, Among them ti k This represents the instant when the k-th event is triggered by the i-th agent; To track errors, there are also in,
[0077] The event-triggered protocol of a leaderless multi-agent system is described as follows:
[0078]
[0079] In the formula, This represents the instant when the (k+1)th event is triggered by the i-th agent. γ i i = 1, ..., N, θ i i = 1, ..., 4 are positive constants; For estimation error; ε i (t) is an internal variable, satisfying... Where ε i (0)>0, k i >0, σ i >0, δ i >0.
[0080] The event triggering protocol of a multi-agent system with a leader is described as follows:
[0081]
[0082]
[0083] In the formula, This represents the instant when the (k+1)th event is triggered by the i-th agent. γ i i = 1, ..., N, θ i ,i=1,…,4 are positive constants; ε i (t) is an internal variable, satisfying... Where ε i (0)>0, k i >0, σ i >0, δ i >0.
[0084] The communication topology between multi-agent systems is an undirected graph, and the agents include robots, drones, unmanned vehicles, and unmanned ships.
[0085] In a multi-agent system, agent i only acts when the adaptive event triggering condition is met. When an agent i is in a state, it sends its own state information to neighboring agents. The input controller updates the control input u of agent i only when agent i or its neighboring agents meet the adaptive event triggering condition. i (t).
[0086] To better illustrate the effectiveness of the multi-agent cluster collaboration technology and system for communication bandwidth-constrained and uncertain interference environments described in this invention, simulation verification is performed for two scenarios of the method proposed in this invention, as detailed below:
[0087] Scenario 1: Ten agents, with the following topology. Figure 2 As shown, the above dynamic equations are satisfied, and f i (t,x i ) = ie -t i = 4, 5, 6; f i (t,x i ) = i / (i+t), i = 7, 8; f i (t,x i )=sin(||x i (t)||) / i, i=9,10; the parameter can be chosen as θ i =1, i=1,…4; γ i =2,k i =0.25, σ i =0.25, δ i =0.1, i=1,…10, from which we can obtain K = [-1.000 -1.7321], Through simulation of the event triggering conditions, the following results were obtained: Figure 3 The consensus error shown also tends to stabilize, as the adaptive gain tends to stabilize. Figure 4 The trigger time is shown.
[0088] Scenario 2: A multi-agent system with one leader and eight followers, with the following topology. Figure 5 As shown, the above dynamic equations are satisfied, and f i (t,x i )=i / 10sin(t),i=1,…4;f i (t,x i ), i = 5, ..., 8 follow a normal distribution; the parameter can be chosen as θ i =1, i=1,…4; γ i =2,k i =0.25, σ i =1,δ i =0.2, i=0,1,…8, from which we can obtain K=[-1.4049 -0.4108 -0.0093 -0.0467], Through simulation of the event triggering conditions, the following results were obtained: Figure 6 As shown, the tracking error and adaptive coupling gain also tend to stabilize, such as Figure 7 The trigger time is shown.
[0089] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.
Claims
1. A multi-agent cluster event-triggered collaborative method for nondeterministic interference environments, characterized in that, The method includes the following steps: Acquire information from each agent, construct a multi-agent system, and determine the state information of each agent based on the state input and uncertainty of each agent through the dynamic equations of the multi-agent system. The estimation error is obtained, the event triggering protocol is determined based on the state information of the agent and the estimation error, and the control objective of the multi-agent system is determined based on the state information of the agent, wherein the control objective is to make the states of the multiple agents tend to be consistent. Obtain the adaptive coupling gain information of the agent, and determine the input controller based on the adaptive coupling gain information of the agent; Determine whether the state of the intelligent agent satisfies the event triggering protocol. If it does, neighboring intelligent agents communicate with each other and update the input controller corresponding to the intelligent agent to achieve the control objective. The multi-agent system includes a leaderless multi-agent system and a multi-agent system with a leader; The input controller of the leaderless multi-agent system is described as follows: Among them, K and It is the feedback gain matrix to be determined; , Initial value , Adaptive coupling gain, , , ;in Indicates the first The instant when the k-th event of an agent is triggered; For consensus error, where And there are ,in ;make ,in This indicates that for intelligent agents State estimation; It is a nonlinear function, in hour, ; K and By solving and We obtained, among which, inverse matrix It is a linear matrix inequality The solution, A known constant matrix that is dimensionally compatible; The input controller of the multi-agent system with a leader is described as follows: In the formula, To track errors, there are also ,in, , ; The event triggering protocol of the leaderless multi-agent system is described as follows: In the formula, Indicates the first The instant when the (k+1)th event of an agent is triggered, ; , It is a positive number; , , To estimate the error; It is an internal variable, satisfying ,in , , , ; The event triggering protocol of a multi-agent system with a leader is described as follows: In the formula, .
2. The multi-agent cluster event-triggered collaborative method for non-deterministic interference environments according to claim 1, characterized in that, The dynamic equations of the multi-agent system are described as follows: in, Indicates the first The state of an agent, Indicates the first The state input of each agent. Indicates the first The uncertainty of each agent and the constraints are satisfied. , It is an existing constant. and It is a known constant matrix that is dimension-compatible.
3. The multi-agent cluster event-triggered collaborative method for non-deterministic interference environments according to claim 1, characterized in that, The leader in a multi-agent system only sends signals to other agents, and the model of the leader is described as follows: Where A is a dimension-compatible known constant matrix. This indicates the state of the leader agent.
4. The multi-agent cluster event-triggered collaborative method for non-deterministic interference environments according to claim 3, characterized in that, The control objective of the leaderless multi-agent system is described as follows: in, Indicates the first The state of an agent at time t. Indicates the first The state of an agent at time t.
5. A multi-agent cluster event-triggered collaborative method for non-deterministic interference environments according to claim 3, characterized in that, The control objective of a multi-agent system with a leader is described as follows: in, Indicates the first The state of an agent at time t. This represents the state of the leader agent at time t.
6. The multi-agent cluster event-triggered collaborative method for nondeterministic interference environments according to claim 1, characterized in that, The communication topology between the multi-agent system is an undirected graph, and the agents include robots, drones, unmanned vehicles, and unmanned ships.