Energy-saving event-triggered agent formation control method and system

Abstract

This invention discloses an energy-saving event-triggered agent formation control method and system. The method includes a pre-defined agent formation comprising multiple agents, establishing a dynamic model considering external disturbances, setting desired states for the agents and acquiring their real-time states, initializing the latest trigger times and states for each agent, calculating the cooperative error and measurement error of each agent at the corresponding latest trigger time based on the desired state, real-time state, and latest trigger state, setting a first trigger condition, updating the latest trigger times and states of agents satisfying the first trigger condition and calculating the control input, sending the updated latest trigger states to other agents, setting a second trigger condition, and switching the control input satisfying the second trigger condition to zero. This method can significantly reduce energy consumption when the system is stable and avoids continuous communication, showing greater potential in reducing communication burden.

Description

Technical Field

[0001] This invention relates to the field of intelligent agent formation control technology, and in particular to an energy-saving event-triggered intelligent agent formation control method and system. Background Technology

[0002] As modern industrial tasks place increasingly higher demands on control systems, the challenge of accomplishing complex industrial tasks using a single agent is growing. Compared to single agents, multi-agent systems offer advantages such as high scalability, strong system performance, and high reliability, attracting widespread attention from researchers. In recent years, agent swarm control technology has developed rapidly and is one of the key technologies driving the realization of intelligent manufacturing, as well as a research hotspot in control science.

[0003] Most existing agent formation control algorithms do not consider resource constraints. Furthermore, formation control algorithms rely on communication between formation members. Therefore, continuous updates to the control algorithm mean continuous communication between formation members, which consumes a significant amount of the agent's communication bandwidth.

[0004] Currently, the two commonly used methods for saving communication resources are time-triggered and event-triggered methods. Time-triggered methods, also known as sampling, involve periodic communication between agents; while event-triggered methods determine when to communicate by designing event functions, which makes them highly promising in terms of resource saving.

[0005] However, existing event-triggered control algorithms maintain a constant state between two consecutive triggers, which leads to significant energy consumption. Furthermore, it's important to note that event-triggered algorithms consist of trigger states, meaning they are only effective for a short period after the trigger. If a significant amount of time has passed since the trigger and the next trigger has not yet arrived, the trigger state will differ from the actual state, meaning that control inputs at this point may compromise system stability. Summary of the Invention

[0006] To address the technical problem that existing agent formation control schemes require continuous communication and that agent event-triggered formation control schemes maintain a constant level between adjacent triggers, resulting in high energy consumption in the formation system, this invention application proposes an energy-saving event-triggered agent formation control method and system.

[0007] This invention provides an energy-saving event-triggered intelligent agent formation control method, which includes the following steps:

[0008] S1. Preset agent formation, which includes multiple agents and communication connections between them;

[0009] S2. Establish a dynamic model of multiple agents under unknown external disturbances;

[0010] S3. Set the desired state of multiple agents and obtain the real-time state of multiple agents. Initialize the latest trigger time and latest trigger state of multiple agents. Calculate the cooperative error of each agent in the agent formation at the corresponding latest trigger time and the measurement error of each agent based on the desired state, real-time state and latest trigger state.

[0011] S4. Set the first trigger condition based on the coordination error of each agent at the latest trigger time and the measurement error of each agent. Determine whether each agent in the agent formation meets the first trigger condition. Update the latest trigger time and latest trigger state of the agent using the time and state corresponding to the agent when the first trigger condition is met. Transmit the updated latest trigger state to other agents. Calculate the control input of the dynamic model corresponding to the agent that meets the first trigger condition. Update the coordination error of the agent that meets the first trigger condition at the latest trigger time.

[0012] S5. Set a second trigger condition based on the cooperative error of each agent at the latest trigger time, determine whether each agent in the agent formation meets the second trigger condition, and switch the control input of the dynamic model corresponding to the agent that meets the second trigger condition to zero.

[0013] Preferably, the dynamic model of the agent in S2 is as follows:

[0014]

[0015] Where, x i Let represent the real-time state of the i-th agent. For x i The first derivative, u i d is the control input for the i-th agent. i Let be the unknown external disturbance of the i-th agent.

[0016] Preferably, S3 specifically includes the following:

[0017] S31. Set the desired state of multiple agents, obtain the real-time state of multiple agents, and calculate the state error of each agent based on the real-time state and desired state of each agent.

[0018] S32. Initialize the latest trigger time and latest trigger state of multiple agents, and calculate the state error of each agent at the corresponding latest trigger time based on the latest trigger state and expected state of each agent.

[0019] S33. Set the cooperative error of each agent based on the state error of each agent and the state errors of other agents in the agent formation at their respective latest trigger times.

[0020] S34. Set the coordination error of each agent at the latest triggering time based on the state error of each agent at the latest triggering time and the state errors of other agents in the agent formation at their respective latest triggering times.

[0021] S35. Set the measurement error for each agent based on the collaborative error of each agent at the latest triggering time and the collaborative error of each agent.

[0022] Preferably, S33 can be specifically expressed by the formula:

[0023]

[0024] In the formula, x ei =x i -x di

[0025] Among them, e i Let x be the cooperative error of the i-th agent. ei Let be the state error of the i-th agent. Let j be the latest trigger time corresponding to the j-th agent. For the j-th agent at the latest trigger time State error, i,j=1,2,...,n, where n is the number of agents in the agent formation, a ij Let x represent the communication correlation coefficient between the i-th agent and the j-th agent. i Let x be the real-time state of the i-th agent. di Let be the desired state of the i-th agent.

[0026] Preferably, S34 can be specifically expressed by the formula:

[0027]

[0028] in, For the i-th agent at the corresponding latest trigger time The cooperative error, For the i-th agent at the corresponding latest trigger time The state error, For the j-th agent at the latest trigger time The state error is given by i,j = 1, 2, ..., n, where n is the number of agents in the agent formation.

[0029] Preferably, S35 can be specifically expressed by the formula:

[0030]

[0031] in, Let be the measurement error of the i-th agent.

[0032] Preferably, in S4, a first triggering condition is set based on the collaborative error of each agent at the latest triggering moment and the measurement error of each agent. The first triggering condition is specifically as follows:

[0033]

[0034] Where δ is a positive constant, ||.|| 2 It is the square of the 2-norm of the vector.

[0035] Preferably, in S4, the control input of the dynamic model corresponding to the agent that satisfies the first triggering condition is calculated, and the specific formula for calculating the control input is as follows:

[0036]

[0037] Among them, u i This is the control input for the i-th agent. For the i-th agent at the corresponding latest trigger time The cooperative error is given by k, where k is a positive constant.

[0038] Preferably, in S5, a second triggering condition is set based on the collaborative error of each agent at the latest triggering time. The second triggering condition is specifically as follows:

[0039]

[0040] Where γ are all positive constants, and ||·|| represents the 2-norm of the vector.

[0041] Another aspect of the present invention provides an intelligent agent formation control system, which includes multiple intelligent agents that are interconnected. Each of the multiple intelligent agents is provided with a receiving module, a control module, and a transmitting module connected in sequence, wherein:

[0042] Each agent's receiving module is used to receive the latest trigger state of each of the other agents in the agent formation and send it to the control module of the current agent;

[0043] The control module of each agent is used to obtain the real-time state and latest trigger state of the current agent, and also to receive the latest trigger state of each other agent in the agent formation, and to use the above-mentioned energy-saving event-triggered agent formation control method to control the state of the current agent, and send the latest trigger state of the current agent after updating to the transmission module.

[0044] Each agent's transmitter module is used to send the latest triggered state updated by the current agent to other agents in the agent formation.

[0045] The aforementioned energy-saving event-triggered intelligent agent swarm control method and system includes a pre-set intelligent agent swarm comprising multiple intelligent agents connected to each other. A dynamic model of the multiple intelligent agents under unknown external disturbances is established. The desired states of the multiple intelligent agents are set, and their real-time states are acquired. The latest trigger time and latest trigger state of the multiple intelligent agents are initialized. Based on the desired state, real-time state, and latest trigger state, the cooperative error and measurement error of each intelligent agent in the swarm at the corresponding latest trigger time are calculated. A first trigger condition is set based on the cooperative error and measurement error of each intelligent agent at the corresponding latest trigger time. The latest trigger time and latest trigger state of the intelligent agent are updated using the time and state corresponding to the first trigger condition, and the updated latest trigger state is transmitted to other intelligent agents. The control input of the dynamic model corresponding to the intelligent agent satisfying the first trigger condition is calculated. The cooperative error of the intelligent agent satisfying the first trigger condition at the corresponding latest trigger time is updated. A second trigger condition is set based on the cooperative error of each intelligent agent at the corresponding latest trigger time. The control input of the intelligent agent in the swarm satisfying the second trigger condition is switched to zero to reduce energy consumption. This method introduces two triggering mechanisms. One mechanism updates the latest trigger time and state, calculates the control input, and sends the latest trigger state to other agents. The calculated control input updates the real-time state of the agents to approximate the desired state. The other mechanism switches the control input to zero. Compared with traditional event-triggered control algorithms, this method can significantly reduce energy consumption when the system is stable and avoid continuous communication, thus having greater potential in reducing communication burden. Attached Figure Description

[0046] Figure 1 This is a flowchart of an energy-saving event-triggered intelligent agent formation control method according to an embodiment of the present invention;

[0047] Figure 2 This is a communication topology diagram of an intelligent agent formation in one embodiment of the present invention;

[0048] Figure 3This refers to the state error of each agent in the agent formation in one embodiment of the present invention;

[0049] Figure 4 It is the control input for each intelligent agent in the intelligent agent formation in one embodiment of the present invention;

[0050] Figure 5 It is the trigger time of each intelligent agent in the intelligent agent formation in one embodiment of the present invention. Detailed Implementation

[0051] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0052] In one embodiment, see Figure 1 and Figure 2 , Figure 1 This is a flowchart of an energy-saving event-triggered intelligent agent formation control method according to an embodiment of the present invention. Figure 2 This is a communication topology diagram of an agent formation in one embodiment of the present invention. An energy-saving event-triggered agent formation control method includes the following steps:

[0053] S1. Preset agent formation, which includes multiple agents and communication connections between them.

[0054] See Figure 2 , Figure 2 The communication topology of an agent formation consisting of four agents is shown in the figure. According to... Figure 2 The communication correlation coefficient 'a' between intelligent agents can be used to... ij (a ij Satisfy the following condition: If the i-th agent can obtain the state of the j-th agent, then a ij >0, otherwise a ij =0, and furthermore, the agent lacks self-loop communication, i.e., a ii =0) is represented as:

[0055] a 12 =a 21 =a 34 =a 41 =a 42 =1, a 13 =a 14 =a 23 =a 24 =a 31 =a 32 =a 43 =0,

[0056] For example, a 42 =1 indicates that the fourth agent can obtain the state information of the second agent, a24 =0 indicates that the second agent cannot obtain the state information of the fourth agent.

[0057] S2. Establish a dynamic model of multiple agents under unknown external disturbances.

[0058] In one embodiment, the dynamic model of the agent in S2 is specifically as follows:

[0059]

[0060] Where, x i Let represent the real-time state of the i-th agent. For x i The first derivative, u i d is the control input for the i-th agent. i Let be the unknown external disturbance of the i-th agent.

[0061] S3. Set the desired state of multiple agents and obtain the real-time state of multiple agents. Initialize the latest trigger time and latest trigger state of multiple agents. Calculate the cooperative error of each agent in the agent formation at the corresponding latest trigger time and the measurement error of each agent based on the desired state, real-time state and latest trigger state.

[0062] Specifically, during the entire process of controlling the agent formation using an energy-saving event-triggered agent formation control method, each agent will have multiple corresponding trigger moments and trigger states. Among these multiple trigger moments, the trigger moment closest to the "current moment" is the "latest trigger moment," and the state corresponding to the "latest trigger moment" is the "latest trigger state." Therefore, the "latest trigger moment" for each agent changes with the "current moment" throughout the control process, and the "latest trigger moments" for different agents may be the same or different. The latest trigger moment and latest trigger state of each agent can be initialized based on its initial moment and initial state.

[0063] In one embodiment, S3 specifically includes the following:

[0064] S31. Set the desired state of multiple agents, obtain the real-time state of multiple agents, and calculate the state error of each agent based on the real-time state and desired state of each agent.

[0065] Furthermore, the state error of each agent can be expressed as:

[0066] x ei =x i -x di (2)

[0067] Where, xi Let x be the state of the i-th agent. di Let x be the desired state of the i-th agent. ei Let be the state error of the i-th agent.

[0068] S32. Initialize the latest trigger time and latest trigger state of multiple agents, and calculate the state error of each agent at the corresponding latest trigger time based on the latest trigger state and expected state of each agent.

[0069] Furthermore, the state error of each agent at the corresponding latest trigger moment can be expressed as:

[0070]

[0071] in, Let represent the state of the i-th agent at the latest trigger moment, which is the latest trigger state of the i-th agent. Let be the desired state of the i-th agent at the latest triggering time. Let be the state error of the i-th agent at the latest trigger time. This represents the latest trigger time for the i-th agent.

[0072] It should be noted that the desired state of each agent can be a constant value or a value that changes over time. In this embodiment, to more accurately describe the desired state of each agent, the desired state of each agent is preset to a value that changes over time.

[0073] S33. Set the cooperative error of each agent based on the state error of each agent and the state errors of other agents in the agent formation at their respective latest trigger times.

[0074] Furthermore, the cooperative error of each agent can be expressed as:

[0075]

[0076] Among them, e i Let x be the cooperative error of the i-th agent. ei Let be the state error of the i-th agent. Let j be the latest trigger time corresponding to the j-th agent. For the j-th agent at the latest trigger time State error, i,j=1,2,...,n, where n is the number of agents in the agent formation, a ij Let a represent the communication correlation coefficient between the i-th agent and the j-th agent. ij ≥0, where, when a ijWhen a > 0, it means that the i-th agent can obtain the state of the j-th agent. ij When = 0, it means that the i-th agent cannot obtain the state of the j-th agent.

[0077] S34. Set the coordination error of each agent at the latest trigger time based on the state error of each agent at the latest trigger time and the state errors of other agents in the agent formation at their respective latest trigger times.

[0078] Furthermore, the cooperative error of each agent at the latest triggering moment can be expressed as:

[0079]

[0080] in, For the i-th agent at the corresponding latest trigger time Cooperative error, For the i-th agent at the corresponding latest trigger time The state error, For the j-th agent at the latest trigger time The state error is given by i,j = 1, 2, ..., n, where n is the number of agents in the agent formation.

[0081] S35. Set the measurement error for each agent based on the collaborative error of each agent at the latest triggering time and the collaborative error of each agent.

[0082] Furthermore, the measurement error of each agent can be expressed as:

[0083]

[0084] in, Let be the measurement error of the i-th agent.

[0085] S4. Set a first trigger condition based on the coordination error of each agent at the latest trigger time and the measurement error of each agent. Determine whether each agent in the agent formation meets the first trigger condition. Update the latest trigger time and latest trigger state of the agent using the time and state corresponding to the agent when the first trigger condition is met. Transmit the updated latest trigger state to other agents. Calculate the control input of the dynamic model corresponding to the agent that meets the first trigger condition. Update the coordination error of the agent that meets the first trigger condition at the latest trigger time.

[0086] In one embodiment, the first triggering condition is specifically:

[0087]

[0088] Where δ is a positive constant, ||.|| 2 It is the square of the 2-norm of the vector.

[0089] In one embodiment, the control input calculation formula for the dynamic model corresponding to the agent that satisfies the first triggering condition is specifically as follows:

[0090]

[0091] Among them, u i This is the control input for the i-th agent. For the i-th agent at the corresponding latest trigger time The cooperative error is given by k, where k is a positive constant.

[0092] Specifically, a first triggering condition is set based on the coordination error of each agent at the latest triggering time and the measurement error of each agent. That is, the measurement error of each agent and the coordination error of each agent at the latest triggering time are compared: if the first triggering condition is not met, the latest triggering time and the latest triggering state are not updated; if the first triggering condition is met, the "latest triggering time" of the agent that meets the first triggering condition is updated to the "current time" (that is, the time corresponding to when the first triggering condition is met), the "latest triggering state" of the agent that meets the first triggering condition is updated to the "current state" (that is, the state corresponding to when the first triggering condition is met), and the updated latest triggering state is sent to other agents in the agent formation.

[0093] For an agent that meets the first triggering condition, a control input will be calculated according to formula (8). This control input is used to update the real-time state of the agent that meets the first triggering condition (specifically, the real-time state is obtained by solving the dynamic model in formula (1) according to the control input) so that it is close to the desired state.

[0094] For agents that meet the first triggering condition, their coordination error at the latest triggering time will also be updated. Specifically, the coordination error of agents that meet the first triggering condition at the latest triggering time will be updated based on the latest triggering state of the agents that meet the first triggering condition and the latest triggering states of other agents that the agents have received when the first triggering condition is met. For details, please refer to formula (5), which will not be elaborated here.

[0095] S5. Set a second trigger condition based on the cooperative error of each agent at the latest trigger time, determine whether each agent in the agent formation meets the second trigger condition, and switch the control input of the dynamic model corresponding to the agent that meets the second trigger condition to zero.

[0096] In one embodiment, the second triggering condition is specifically:

[0097]

[0098] Where γ are all positive constants, and ||·|| represents the 2-norm of the vector.

[0099] Specifically, a second triggering condition is set and judged based on the collaborative error of each agent at its latest triggering moment. The control input of the agent that meets the second triggering condition is directly switched to zero, which can reduce the energy consumption of the entire system.

[0100] It should be noted that if an agent simultaneously meets the first trigger condition and the second trigger condition, the latest trigger time and the latest trigger state corresponding to that agent will be updated, the updated latest trigger state will be sent to other agents in the agent formation, and the control input of that agent will be directly switched to zero.

[0101] In one embodiment, an intelligent agent formation control system includes multiple intelligent agents that are interconnected. Each intelligent agent is equipped with a receiving module, a control module, and a transmitting module connected sequentially.

[0102] Each agent's receiving module is used to receive the latest trigger state of each of the other agents in the agent formation and send it to the control module of the current agent;

[0103] The control module of each agent is used to obtain the real-time state and latest trigger state of the current agent, and also to receive the latest trigger state of each other agent in the agent formation, and to use the above-mentioned energy-saving event-triggered agent formation control method to control the state of the current agent, and send the latest trigger state of the current agent after updating to the transmission module.

[0104] Each agent's transmitter module is used to send the latest triggered state updated by the current agent to other agents in the agent formation.

[0105] For specific limitations regarding an intelligent agent formation control system, please refer to the limitations of an energy-saving event-triggered intelligent agent formation control method mentioned above, which will not be repeated here.

[0106] Furthermore, the energy-saving event-triggered intelligent agent formation control method of the present invention was verified through experiments.

[0107] In this embodiment, verification is performed through numerical simulation.

[0108] See Figure 2 , Figure 2The communication topology of an agent formation consisting of four agents is shown in the figure.

[0109] according to Figure 2 We can obtain: a 12 =a 21 =a 34 =a 41 =a 42 =1, a 13 =a 14 =a 23 =a 24 =a 31 =a 32 =a 43 =0, for example, a 42 =1 indicates that the fourth agent can obtain the state information of the second agent, a 24 =0 indicates that the second agent cannot obtain the state information of the fourth agent.

[0110] Set the initial state of the agent as: x1(0)=1, x2(0)=2, x3(0)=3, x4(0)=4;

[0111] Let the desired state of the agent be: x d1 =0.25, x d2 =0.5, x d3 =0.75, x d4 =1. In this embodiment, for the convenience of calculation and verification, the expected state of each agent is set to a constant value. Alternatively, the expected state of each agent can be set to a value that changes over time.

[0112] Control input satisfies: |u i |≤1, i=1,2,3,4;

[0113] The disturbance is chosen as: d1 = 2 × 10 -4 sin(0.1t), d2=3×10 -4 sin(0.2t), d3=1×10 -4 cos(0.2t), d4=2×10 -4 cos(0.1t);

[0114] The control parameters are set as follows: k = 0.5, δ = 1 × 10 -7 γ = 6 × 10 -3 .

[0115] Under the above simulation conditions, the simulation results are as follows: Figure 3-5 As shown, where, Figure 3 This refers to the state error of each agent in the agent formation in one embodiment of the present invention. Figure 4This is the control input for each agent in the agent formation in one embodiment of the present invention. Figure 5 This refers to the trigger time of each agent in the agent formation in one embodiment of the present invention.

[0116] The simulation results show that, under the action of the aforementioned energy-saving event-triggered formation control algorithm, Figure 2 The state errors of the four agents in the agent formation shown converge to near zero within 10 seconds. Once the entire system stabilizes, the control inputs of the four agents are mostly zero, which significantly reduces system energy consumption. Furthermore, the triggering frequency of the multi-agent system is very low, indicating that the algorithm is highly efficient in reducing communication burden.

[0117] The aforementioned energy-saving event-triggered intelligent agent swarm control method and system includes a pre-set intelligent agent swarm comprising multiple intelligent agents connected to each other. A dynamic model of the multiple intelligent agents under unknown external disturbances is established. The desired states of the multiple intelligent agents are set, and their real-time states are acquired. The latest trigger time and latest trigger state of the multiple intelligent agents are initialized. Based on the desired state, real-time state, and latest trigger state, the cooperative error and measurement error of each intelligent agent in the swarm at the corresponding latest trigger time are calculated. A first trigger condition is set based on the cooperative error and measurement error of each intelligent agent at the corresponding latest trigger time. The latest trigger time and latest trigger state of the intelligent agent are updated using the time and state corresponding to the first trigger condition, and the updated latest trigger state is transmitted to other intelligent agents. The control input of the dynamic model corresponding to the intelligent agent satisfying the first trigger condition is calculated. The cooperative error of the intelligent agent satisfying the first trigger condition at the corresponding latest trigger time is updated. A second trigger condition is set based on the cooperative error of each intelligent agent at the corresponding latest trigger time. The control input of the intelligent agent in the swarm satisfying the second trigger condition is switched to zero to reduce energy consumption. This method introduces two triggering mechanisms. One mechanism updates the latest trigger time and the latest trigger state (i.e., the state corresponding to the latest trigger time), calculates the control input, and sends the latest trigger state to other agents. The calculated control input updates the real-time state of the agents to make it close to the desired state. The other triggering mechanism switches the control input to zero. Compared with traditional event-triggered control algorithms, this method can not only significantly reduce energy consumption when the system is stable, but also avoid continuous communication, and has greater potential in reducing communication burden.

[0118] The present invention provides a detailed description of an energy-saving event-triggered intelligent agent formation control method and system. Specific examples have been used to illustrate the principles and implementation methods of the invention. The descriptions of these embodiments are merely illustrative of the core ideas of the invention. It should be noted that those skilled in the art can make various improvements and modifications to the invention without departing from its principles, and these improvements and modifications also fall within the scope of protection of the claims.

Claims

1. An energy saving event-triggered intelligent agent platoon control method, characterized in that, The method includes the following steps: S1. A preset intelligent agent formation, wherein the intelligent agent formation includes multiple intelligent agents and the multiple intelligent agents are connected to each other; S2. Establish a dynamic model of the multiple intelligent agents under unknown external disturbances; S3. Set the desired state of multiple agents and obtain the real-time state of multiple agents. Initialize the latest trigger time and latest trigger state of multiple agents. Calculate the coordination error of each agent in the agent formation at the corresponding latest trigger time and the measurement error of each agent based on the desired state, real-time state and latest trigger state. S4. Set a first trigger condition based on the coordination error of each agent at the latest trigger time and the measurement error of each agent. Determine whether each agent in the agent formation meets the first trigger condition. Update the latest trigger time and latest trigger state of the agent using the time and state corresponding to the agent when the first trigger condition is met. Transmit the updated latest trigger state to other agents. Calculate the control input of the dynamic model corresponding to the agent that meets the first trigger condition. Update the coordination error of the agent that meets the first trigger condition at the latest trigger time. S5. Set a second triggering condition based on the collaborative error of each agent at the latest triggering time, determine whether each agent in the agent formation meets the second triggering condition, and switch the control input of the dynamic model corresponding to the agent that meets the second triggering condition to zero. The dynamic model of the agent described in S2 is as follows: wherein, is the real-time state of the ith agent, is the first derivative of is the control input of the ith agent, is the unknown external disturbance of the ith agent;​ The first triggering condition in S4 is specifically: in, Let be the measurement error of the i-th agent. For the i-th agent at the corresponding latest trigger time The cooperative error, For positive integers, It is the square of the 2-norm of the vector; The second triggering condition in S5 is specifically as follows: wherein are all normal numbers, denotes the 2-norm of a vector.

2. The energy-efficient event-triggered swarm control method of claim 1, wherein, Specifically, S3 includes the following: S31. Set the desired state of multiple intelligent agents, obtain the real-time state of multiple intelligent agents, and calculate the state error of each intelligent agent based on the real-time state and desired state of each intelligent agent. S32. Initialize the latest trigger time and latest trigger state of the multiple agents, and calculate the state error of each agent at the corresponding latest trigger time based on the latest trigger state and expected state of each agent. S33. Set the collaborative error of each agent based on the state error of each agent and the state errors of other agents in the agent formation at their respective latest trigger times. S34. Set the coordination error of each agent at the latest triggering time according to the state error of each agent at the latest triggering time and the state errors of other agents in the agent formation at their respective latest triggering times; S35. Set the measurement error of each agent based on the coordination error of each agent at the latest triggering time and the coordination error of each agent.

3. The energy-saving event-triggered intelligent agent formation control method as described in claim 2, characterized in that, S33 can be specifically expressed by the formula: In the formula, in, Let be the cooperative error of the i-th agent. Let be the state error of the i-th agent. Let j be the latest trigger time corresponding to the j-th agent. For the j-th agent at the latest trigger time The state error, , The number of agents in the agent formation. This represents the communication correlation coefficient between the i-th agent and the j-th agent. Let represent the real-time state of the i-th agent. Let be the desired state of the i-th agent.

4. The energy-saving event-triggered intelligent agent formation control method as described in claim 3, characterized in that, S34 can be specifically expressed by the following formula: in, For the i-th agent at the corresponding latest trigger time Cooperative error, For the i-th agent at the corresponding latest trigger time The state error, For the j-th agent at the latest trigger time The state error, , The number of agents in the agent group.

5. The energy-saving event-triggered intelligent agent formation control method as described in claim 4, characterized in that, S35 can be specifically expressed by the following formula: in, Let be the measurement error of the i-th agent.

6. The energy-saving event-triggered intelligent agent formation control method as described in claim 5, characterized in that, In S4, the control input of the dynamic model corresponding to the agent that satisfies the first triggering condition is calculated. The specific formula for calculating the control input is as follows: in, This is the control input for the i-th agent. For the i-th agent at the corresponding latest trigger time Cooperative error, It is a positive number.

7. A smart agent formation control system, characterized in that, The system includes multiple intelligent agents that are interconnected. Each of the intelligent agents is equipped with a receiving module, a control module, and a transmitting module connected in sequence. Each agent's receiving module is used to receive the latest trigger state of each of the other agents in the agent formation and send it to the control module of the current agent; The control module of each agent is used to obtain the real-time state and latest trigger state of the current agent, and is also used to receive the latest trigger state of each other agent in the agent formation, and to use the energy-saving event-triggered agent formation control method as described in any one of claims 1 to 6 to control the state of the current agent, and to send the updated latest trigger state of the current agent to the transmission module. Each agent's transmission module is used to send the latest triggered state of the current agent to other agents in the agent formation.

Images