Network connectivity self-healing method and apparatus

Abstract

The application discloses a network connection self-healing method and device, wherein the method comprises the following steps: an intelligent agent in a multi-agent system determines a position state vector of itself relative to a formation reference origin at each time, and reports the position state vector to a control station to trigger the control station to calculate an average value of the position state vectors of all intelligent agents at the corresponding time, and feed back the average value to the intelligent agent; in response to receiving the average value, the intelligent agent determines a global position state vector difference value of itself based on the average value, determines a local position state vector difference value of itself based on the position state vectors of other intelligent agents in direct connection with itself at the current time, generates a new control speed for the intelligent agent based on the global position state vector difference value, the local position state vector difference value and an expected speed, and controls the intelligent agent to fly at the control speed. The self-healing of network splitting can be realized by adopting the application.

Description

Technical Field

[0001] This invention relates to mobile communication technology, and in particular to a network connectivity self-healing method and apparatus. Background Technology

[0002] In multi-agent networks such as drones, unmanned ships, and mobile robot swarms, autonomous agents can interact with each other through communication networks to achieve global objectives. Therefore, maintaining communication connectivity between agents is crucial for the coordinated completion of system goals by multi-agent systems. Multi-agent network connectivity is a key indicator for measuring the communication connectivity of multi-agent systems and has received widespread attention in research on multi-agent communication networking, formation control, and cooperative strategies.

[0003] Current research on multi-agent network connectivity largely relies on the assumption of constant connectivity, meaning that agents always maintain communication connections. However, in reality, multi-agent systems cannot guarantee communication connectivity throughout network evolution. The high mobility of agents makes multi-agent networks prone to fragmentation, where some agents lose connection with others. However, current research rarely considers network fragmentation, where communication connections between agents are interrupted, leading to network fragmentation. This is an unavoidable problem in the implementation of multi-agent systems, especially during network initialization and when network failures cause node failures. Currently, no mature self-healing scheme for network connectivity has been proposed to address network fragmentation caused by high agent mobility, restore disconnected connections between agents, and achieve overall network connectivity self-healing. Summary of the Invention

[0004] In view of this, the main objective of the present invention is to provide a network connectivity self-healing method and apparatus that can restore disconnected connections between agents when a network splits, thereby achieving self-healing of the overall network connectivity.

[0005] To achieve the above objectives, the technical solution proposed in this embodiment of the invention is as follows:

[0006] A network connectivity self-healing method, comprising:

[0007] In a multi-agent system, each agent determines its own position state vector relative to the formation reference origin at each time step and reports it to the control station. This triggers the control station to calculate the average value of the position state vectors of all agents at the corresponding time step and feeds it back to the agent.

[0008] In response to receiving the average value, the agent determines its global position state vector difference based on the average value and its own position state vector at the current moment, and determines its local position state vector difference based on the position state vectors of other agents directly connected to it at the current moment. Based on the global position state vector difference, the local position state vector difference, and a preset desired speed, a new control speed is generated for the agent to make the agent's position offset vector in the formation converge to the preset desired position offset vector, and the agent is controlled to fly at the current control speed.

[0009] This invention also proposes a network connectivity self-healing device, including a processor and a memory;

[0010] The memory stores an application program that can be executed by the processor, which enables the processor to execute the network connectivity self-healing method as described above.

[0011] This invention also proposes a computer-readable storage medium storing computer-readable instructions for execution by a processor of any of the network connectivity self-healing methods described above.

[0012] In summary, the network connectivity self-healing scheme proposed in this invention introduces a federated architecture, which combines distributed control with centralized control. The control station calculates the average position state vector of all agents at each time step and feeds it back to each agent. This allows each agent to determine its own global position state vector difference based on this average. Furthermore, based on this global position state vector difference and the local position state vector difference generated from the position state vectors of neighboring agents, the current control rate is generated. This ensures that the position offset vector of an agent in the formation converges to a preset desired position offset vector. Thus, in the event of network splitting, simple global coordination and restoration of connections between agents can be achieved with low communication and computational overhead, realizing self-healing of overall network connectivity. This overcomes the impact of network splitting on multi-agent systems in practical applications, fully leveraging the advantages of multi-agent systems to efficiently and effectively complete the target task. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of a multi-agent system architecture according to an embodiment of the present invention;

[0014] Figure 2 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0015] Figure 3 This is a simulation result diagram of an embodiment of the present invention;

[0016] Figure 4 for Figure 3 Example of a step-by-step connected snapshot. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] To achieve network connectivity self-healing in network fragmentation scenarios, this invention introduces a federated architecture. In this architecture, each agent achieves distributed autonomous control through direct connections with surrounding agents, while also communicating with a control station and operating under its monitoring. The control station can be a ground control center or a manned aircraft. In an integrated air-space-ground network, cellular mobile communication, satellite communication links, or other feasible communication links can be used to connect the control station with the multi-agent system.

[0019] like Figure 1 As shown, the federated control architecture comprises N agents and a control station. The control station receives data from the agents and, through simple global coordination such as global averaging, minimization, and maximization, sends control information to the agents, supervising and regulating the multi-agent system, thus acting as a server. The agents also communicate with each other, exchanging their state information and collaboratively completing tasks. In this way, when a network split occurs, each agent can obtain global state information through communication with the control station. Based on this global state information and the state information of directly connected agents, it can adjust its control input (i.e., flight rate) to converge its position in the formation to a predetermined desired position, thereby achieving the desired formation of the agent system and enabling network connectivity self-healing.

[0020] Figure 2 This is a schematic diagram of the network connectivity self-healing method according to an embodiment of the present invention, as shown below. Figure 2 As shown, this embodiment mainly includes the following:

[0021] Step 201: At each moment, the agents in the multi-agent system determine their own position state vector relative to the formation reference origin and report it to the control station, so as to trigger the control station to calculate the average value of the position state vectors of all agents at the corresponding moment and feed it back to the agents.

[0022] In this step, after the control station receives the position state vectors reported by all agents in the system, it calculates the corresponding average value and provides it to each agent. This allows each agent to know its own global position state vector difference and then adjust its flight speed based on the global position state vector difference, so that its position offset vector in the formation can converge to the expected value, thereby achieving self-healing of network communication.

[0023] In one implementation, the control station can broadcast the average value of the position state vector to each agent.

[0024] In one implementation, each agent can specifically be configured according to x i (t)=p i (t)-(δ i +v*t), determine its own position and state vector relative to the formation reference origin.

[0025] Where t represents the time when the agent's control law is used, indicating that at time t, t≥0, x i (t) is the position and state vector of agent i relative to the formation reference origin, p i (t) is the position vector of agent i at time t; δ i The desired position offset vector of agent i in the formation, whose value is determined by the preset desired formation, v * This is the preset expected speed, a value shared by all agents, used to characterize the expected operating speed of all agents.

[0026] Step 202: In response to receiving the average value, the agent determines its global position state vector difference based on the average value and its own position state vector at the current moment, and determines its local position state vector difference based on the position state vectors of other agents directly connected to it at the current moment. Based on the global position state vector difference, the local position state vector difference, and a preset desired speed, a new control speed is generated for the agent so that the agent's position offset vector in the formation converges to the preset desired position offset vector, and the agent is controlled to fly at the current control speed.

[0027] In this step, each agent will generate the current control rate based on the average of the position and state vectors of all agents sent by the control station, combined with the position and state vectors of other agents that are directly connected to it at the current moment. This will enable the formation of the agent system to converge to the desired formation even in the event of network splitting, thereby achieving network communication self-healing.

[0028] In one implementation, each agent can specifically follow the... Determine the difference Δx between your own global position and state vectors. i (t).

[0029] Where, x i (t) is the position and state vector of agent i at time t relative to the formation reference origin. Let τ be the average value of the position state vector received by agent i at time t, and τ be the total transmission and processing delay required to obtain the average value.

[0030] Specifically, x i (t) can be calculated according to the formula x i (t)=p i (t)-(δ i +v*t), calculated as follows.

[0031] in, It is the position offset vector of agent i in the desired formation (i.e., the desired position offset vector). It is the expected speed, p i (t) represents the actual position of agent i at time t.

[0032] In one implementation, each agent can specifically follow the... Determine the local position state vector difference Δp i (t).

[0033] Among them, g ij Let g be the connection state between agent i and agent j. If they are connected, then g... ij =1, if not connected, then g ij =0; Δp ij (t)=x i (t)-x j (t); x i (t) is the position and state vector of agent i at time t relative to the formation reference origin; x j (t) is the position and state vector of agent j at time t relative to the formation reference origin.

[0034] In practical applications, the connection state between agent i and agent j can be specifically represented by the following formula:

[0035]

[0036] In this scenario, each agent is assumed to have a fixed communication radius R. If the distance between two agents is less than R, they are connected; otherwise, they are not connected.

[0037] In one implementation, each agent can follow u i (t)=v *-θΔp i (t)-γΔx i (t), generating a new control rate u i (t).

[0038] Where θ is a preset local compensation coefficient; γ is a preset global compensation coefficient; both θ and γ are greater than zero, and can be set to appropriate values ​​by those skilled in the art according to actual application needs; v * The preset desired speed; Δp i (t) represents the difference in the local position state vector of agent i at time t; Δx i (t) represents the difference in the global position and state vector of agent i at time t.

[0039] In one implementation, before performing the above steps to achieve network connectivity self-healing, the following method can be used to further determine the time required to complete the current network connectivity self-healing:

[0040] Each agent pre-calculates its initial position state vector relative to the formation reference origin and reports the initial position state vector and its expected position offset vector in the formation to the control station. This triggers the control station to calculate the time required to establish a corresponding connection for each unconnected agent pair based on the initial position state vector and the expected position offset vector reported by all agents. The expected relationship between the agent pairs in the formation is a direct connection, meaning that they can communicate directly without the need for forwarding from other nodes.

[0041] In one implementation method, it can be specifically according to Calculate the duration.

[0042] in, The time required to establish a connection between agent i and agent i. x i (0) is the initial position state vector of agent i. The initial position state vector of all agents is the average value; R is the preset communication radius of a single agent; N is the number of agents in the multi-agent system; δ i The desired position offset vector of agent i in the formation; δ j The desired position offset vector of agent j in the formation; γ is the preset global compensation coefficient, which is greater than zero.

[0043] Using the above method, it is possible to perform calculations for each pair of agents (j, i) ∈ ε * The time required to connect unconnected agents i and j is calculated, where ε * This indicates that the relationship between agents i and j in the desired formation is a direct connection.

[0044] As can be seen from the above technical solutions, the above method embodiments introduce a federated architecture, retaining the distributed execution capability of distributed control and the fast convergence capability of centralized control. Through global coordination of agents by the control station, interrupted agents can restore connections, achieving overall network connectivity self-healing. This makes the above technical solutions applicable to scenarios where traditional distributed control cannot work, including situations where network initialization or failures lead to network fragmentation. This network connectivity self-healing method based on federated control can achieve simple global coordination and restore connections between agents with low communication and computational overhead. Therefore, the above method embodiments can overcome the impact of network fragmentation on the performance of multi-agent systems in practical applications, fully leveraging the advantages of multi-agent systems and ensuring efficient and high-quality completion of the target task.

[0045] The following uses a simulation example and combines... Figure 3 and Figure 4 This further illustrates the feasibility of the above-mentioned plan.

[0046] Figure 3 The effectiveness of the above scheme in achieving network communication self-healing using federal control is described. The figure shows the position trajectory of the multi-agent system, the formation shape, and the communication connection diagram during the control process. Figure 3 The diagram shows the process by which multiple agents, initially far apart and with communication interrupted, gradually form a formation and establish communication connections. The squares represent the initial positions of the agents, the dots represent the node positions forming the multi-agent system, and the lines connecting the dots represent communication connections.

[0047] Figure 4 Four snapshots depict the gradual interconnection of a multi-agent network under federal control.

[0048] In the diagram, dots represent agents in an interrupted state, while dots connected by straight lines represent agents in a state of communication connection. It can be seen that in the first three diagrams, agents have not achieved complete communication connection, and some agents remain isolated. In the last diagram, all agents can connect to form a multi-agent network.

[0049] Based on the above method embodiments, this invention also proposes a network connectivity self-healing device, including a processor and a memory; the memory stores an application program executable by the processor, used to cause the processor to execute the network connectivity self-healing method as described above. Specifically, a system or device equipped with a storage medium can be provided, on which software program code implementing the functions of any of the above embodiments is stored, and the computer (or CPU or MPU) of the system or device can read and execute the program code stored in the storage medium. Furthermore, the operating system or other devices operating on the computer can perform some or all of the actual operations through instructions based on the program code. The program code read from the storage medium can also be written to a memory located in an expansion board inserted into the computer or to a memory located in an expansion unit connected to the computer, and then, based on the instructions of the program code, the CPU or other devices installed on the expansion board or expansion unit can execute some or all of the actual operations, thereby realizing the functions of any of the above embodiments of the network connectivity self-healing method.

[0050] Specifically, the memory can be implemented as various storage media such as electrically erasable programmable read-only memory (EEPROM), flash memory, and programmable programmable read-only memory (PROM). The processor can be implemented as one or more central processing units (CPUs) or one or more field-programmable gate arrays (FPGAs), wherein the FPGA integrates one or more CPU cores. Specifically, the CPU or CPU core can be implemented as a CPU or an MCU.

[0051] This application also implements a computer program product, including a computer program / instruction that, when executed by a processor, implements the steps of the network connectivity self-healing method described above.

[0052] It should be noted that not all steps and modules in the above processes and structural diagrams are mandatory; some steps or modules can be omitted as needed. The execution order of the steps is not fixed and can be adjusted as required. The division of modules is merely for the convenience of description and functional division. In actual implementation, a module can be implemented by multiple modules, and the functions of multiple modules can also be implemented by the same module. These modules can be located in the same device or in different devices.

[0053] The hardware modules in each embodiment can be implemented mechanically or electronically. For example, a hardware module may include specially designed permanent circuitry or logic devices (such as dedicated processors, such as FPGAs or ASICs) to perform specific operations. A hardware module may also include programmable logic devices or circuitry (such as general-purpose processors or other programmable processors) temporarily configured by software to perform specific operations. The choice between mechanical implementation, dedicated permanent circuitry, or temporarily configured circuitry (such as software-configured circuitry) can be made based on cost and time considerations.

[0054] In this document, "illustrative" means "serving as an example, illustration, or description," and any illustration or embodiment described herein as "illustrative" should not be construed as a preferred or more advantageous technical solution. For the sake of brevity, the figures only schematically represent the parts relevant to the invention and do not represent their actual structure as a product. Furthermore, for the sake of clarity and ease of understanding, in some figures, components with the same structure or function are only schematically depicted, or only one is labeled. In this document, "a" does not mean that the number of relevant parts of the invention is limited to "only one," and "a" does not exclude the possibility that the number of relevant parts of the invention is "more than one." In this document, terms such as "upper," "lower," "front," "rear," "left," "right," "inner," and "outer" are used only to indicate the relative positional relationship between relevant parts, and not to limit the absolute position of these relevant parts.

[0055] The solutions described in this specification and embodiments, if involving the processing of personal information, will be processed only under the premise of having a legal basis (such as obtaining the consent of the personal information subject, or being necessary for the performance of a contract), and will only be processed within the scope stipulated or agreed upon. A user's refusal to process personal information beyond what is necessary for basic functions will not affect the user's use of basic functions.

[0056] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A network connectivity self-healing method, characterized in that, include: In a multi-agent system, each agent determines its own position state vector relative to the formation reference origin at each time step and reports it to the control station. This triggers the control station to calculate the average value of the position state vectors of all agents at the corresponding time step and feeds it back to the agent. In response to receiving the average value, the agent determines its global position state vector difference based on the average value and its own position state vector at the current moment, and determines its local position state vector difference based on the position state vectors of other agents directly connected to it at the current moment. Based on the global position state vector difference, the local position state vector difference, and a preset desired speed, a new control speed is generated for the agent to make the agent's position offset vector in the formation converge to the preset desired position offset vector, and the agent is controlled to fly at the current control speed.

2. The method according to claim 1, characterized in that, The intelligent agent according to x i (t)=p i (t)-(δ i +v * t), determine its own position and state vector relative to the formation reference origin; Where t is the time t-th moment, t is greater than or equal to zero, and x i (t) is the position and state vector of agent i relative to the formation reference origin, p i (t) is the position vector of agent i at time t; δ i The desired position offset vector of agent i in the formation, v * This is the preset desired speed.

3. The method according to claim 1, characterized in that, The intelligent agent according to Determine the difference Δx between your own global position and state vectors. i (t), where x i (t) is the position and state vector of agent i at time t relative to the formation reference origin. Let τ be the average value of the position state vector received by agent i at time t, and τ be the total transmission and processing delay required to obtain the average value.

4. The method according to claim 1, characterized in that, The intelligent agent according to Determine the local position state vector difference Δp i (t); where g ij Let g be the connection state between agent i and agent j. If they are connected, then g... ij =1, if not connected, then g ij =0; Δp ij (t)=x i (t)-x j (t); x i (t) is the position and state vector of agent i at time t relative to the formation reference origin; x j (t) is the position and state vector of agent j at time t relative to the formation reference origin.

5. The method according to claim 1, characterized in that, The intelligent agent follows u i (t)=v * -θΔp i (t)-γΔx i (t), generating a new control rate u i (t); where θ is the preset local compensation coefficient; γ is the preset global compensation coefficient; both θ and γ are greater than zero; v * The preset desired speed; Δp i (t) represents the difference in the local position state vector of agent i at time t; Δx i (t) represents the difference in the global position and state vector of agent i at time t.

6. The method according to claim 1, characterized in that, The method further includes: The agent pre-calculates its initial position state vector relative to the formation reference origin, and reports the initial position state vector and its expected position offset vector in the formation to the control station. This triggers the control station to calculate the time required to establish a corresponding connection for each unconnected agent pair based on the initial position state vector and the expected position offset vector. The expected relationship between the agent pairs in the formation is a direct connection.

7. The method according to claim 6, characterized in that, according to Calculate the duration, where, The time required to establish a connection between agent i and agent i. x i (0) is the initial position state vector of agent i. The initial position state vector of all agents is the average value; R is the preset communication radius of a single agent; N is the number of agents in the multi-agent system; δ i The desired position offset vector of agent i in the formation; δ j The desired position offset vector of agent j in the formation; γ is the preset global compensation coefficient, which is greater than zero.

8. A network connectivity self-healing device, characterized in that, Including processor and memory; The memory stores an application program that can be executed by the processor, which causes the processor to execute any of the network connectivity self-healing methods as described in claims 1 to 7.

9. A computer-readable storage medium, characterized in that, It stores computer-readable instructions that are executed by a processor using any of the network connectivity self-healing methods as described in claims 1 to 7.

Images