Ad hoc network dynamic slice architecture based on software defined radio and management and control method

Abstract

The invention relates to the field of wireless communication and intelligent networks, in particular to a software defined radio-based ad hoc network dynamic slicing architecture and a management and control method, comprising a basic communication layer, an intelligent control layer and a service execution layer. The basic communication layer adopts a multi-distribution multi-center clustering architecture; the intelligent control layer is responsible for intelligent management and resource scheduling of the self-organizing network, and the intelligent control layer comprises a plurality of command and control nodes; and the service execution layer adopts a micro-service-based architecture, and encapsulates different network functions and service logics into independent micro-service modules. According to the method, the technical problems of high plane coupling, resource scheduling rigidness, insufficient survivability and high hardware dependency in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communications and intelligent networks, and in particular to a dynamic slicing architecture and a management and control method for self-organizing networks based on software-defined radio. Background Art

[0002] Self-organizing networks and their management and control methods have important applications and value in many fields, including modern communications, intelligent transportation, and the Industrial Internet of Things. They are not only used for autonomous network construction, dynamic adjustment, and optimization, but also for addressing the changing connectivity requirements in different scenarios. This greatly improves the network's flexibility and adaptability, and is of great value and significance in improving the overall performance and operational efficiency of the system.

[0003] The common construction method of traditional self-organizing network technology is mainly based on a fixed topology structure and traditional routing algorithms, using a distributed control mechanism. Nodes maintain network connectivity through negotiation and collaboration, while the management and control method focuses on static allocation of network resources and simple traffic control to ensure the realization of basic communication functions.

[0004] However, existing self-organizing network technologies have exposed numerous flaws when dealing with complex scenarios such as multi-task collaboration, mixed transmission of heterogeneous services, and complex electromagnetic environments. Traditional self-organizing networks share transmission resources for service data, routing information, and control instructions, leading to increased latency jitter under high loads and even control signaling congestion. They lack differentiated service capabilities for different tasks and struggle to dynamically adapt to real-time changing service needs (such as low-latency communication for drone swarms and massive data backhaul from ground sensors). Furthermore, when nodes fail or the environment interferes, traditional self-organizing networks rely on static algorithms for path reconstruction, making it difficult to quickly generate the optimal topology to ensure normal network operation. Furthermore, fixed-function communication modules cannot flexibly support new protocols and waveforms, limiting network scalability and the speed of technological iteration. In summary, traditional self-organizing networks suffer from strong planar coupling, rigid resource scheduling, insufficient survivability, and high hardware dependence, making it difficult to meet the stringent requirements for network resilience and service quality driven by the convergence of 6G communications, edge computing, and artificial intelligence technologies. Summary of the Invention

[0005] The purpose of this application is to provide a dynamic slicing architecture and management method for self-organizing networks based on software-defined radio, which solves the technical problems existing in the existing technology, such as strong plane coupling, rigid resource scheduling, insufficient anti-destruction ability, and high hardware dependence.

[0006] In order to solve the above technical problems, the solution adopted by this application is as follows:

[0007] A dynamic slicing architecture for self-organizing networks based on software-defined radio is characterized by: including a hierarchical three-dimensional network plane system;

[0008] The layered three-dimensional network plane system includes a basic communication layer, an intelligent control layer, and a service execution layer;

[0009] The basic communication layer adopts a multi-distributed multi-center clustering architecture. The cluster nodes of the basic communication layer include cluster head nodes and ordinary nodes other than cluster head nodes. The cluster structure is generated based on the comprehensive performance of the cluster nodes. The cluster head nodes are elected using a multi-round voting mechanism based on the comprehensive performance of the cluster nodes. The comprehensive performance includes signal strength, computing power, storage resources, and residual energy.

[0010] The intelligent control layer includes multiple command nodes, each equipped with a phased array antenna and an AI decision engine. When the basic communication layer detects a change in the self-organizing network structure or configuration information, the cluster head node sends the change information to the command node of the intelligent control layer. Based on the change information, each command node of the intelligent control layer sends control instructions to each cluster head node of the basic communication layer to replan the self-organizing network structure, thereby enabling the intelligent control layer to manage the basic communication layer.

[0011] The business execution layer is a microservice-based architecture that encapsulates different network functions and business logic into independent microservice modules. The intelligent control layer issues control instructions, and the business execution layer receives and executes the control instructions. At the same time, the intelligent control layer monitors the execution process of the business execution layer and receives feedback signals from the business execution layer.

[0012] The microservice module is composed of subnet nodes, which work together to form a microservice module that performs specific task functions and businesses; the microservice module includes a reconnaissance task module, a relay task module, and an attack task module.

[0013] Preferably, each cluster node in the basic communication layer has a built-in GPS receiving module, and receives satellite signals in real time through the built-in GPS receiving module to obtain the location information of the cluster node; an Ethernet-based communication link is established between each cluster node to realize information exchange between the cluster nodes;

[0014] During information exchange, the information packet sent by the cluster node includes the time and location information of the cluster node. After receiving the information packet, the other cluster nodes synchronize the time through the adaptive filtering algorithm and multipath compensation algorithm to make the time base of the cluster nodes the same; ensuring that each cluster node communicates and works collaboratively under the same time base.

[0015] Preferably, the communication protocol of the self-organizing network adopts a hybrid routing protocol, and the hybrid routing protocol includes an optimized link state routing protocol and an on-demand distance vector routing protocol-enhanced;

[0016] During the operation of the self-organizing network, one of the above two communication protocols is dynamically selected as the dominant routing protocol; the dynamic selection rules are as follows: when the topology of the self-organizing network is stable, the optimized link state routing protocol is selected as the main protocol to establish the global optimal route. At this time, the on-demand distance vector routing protocol-enhanced auxiliary main communication protocol is used; when rapid changes in the local topology of the self-organizing network or sudden bursts of specific business traffic are detected, the on-demand distance vector routing protocol-enhanced rapid response is used as the dominant communication protocol.

[0017] Preferably, the communication signal transmission waveforms of the self-organizing network include OFDM and SC-FDE, and the two signal transmission waveforms are switched according to real-time channel status information; when the RMS delay spread is greater than 100ns and the available spectrum resources are less than 30%, OFDM is selected; when the carrier frequency offset is greater than 10% of the subcarrier spacing, it is switched to SC-FDE; the real-time channel status information includes multipath strength, signal-to-noise ratio, and frequency offset.

[0018] Preferably, the business execution layer adopts a microservice-based architecture, and the self-organizing network subnet nodes responsible for different network functions and business logics are encapsulated into independent microservice modules at the business execution layer. The microservice modules responsible for the same business can form a task subnet; the task subnet loads and unloads the corresponding microservice modules according to the creation and destruction requirements of new tasks; the microservice modules include a reconnaissance task module, a relay task module and an attack task module; the reconnaissance task module, the relay task module and the attack task module are respectively responsible for image acquisition and transmission, data relay and target attack.

[0019] Preferably, the control method comprises the following steps:

[0020] Step S1: Establish the architecture of basic communication layer, intelligent control layer, and business execution layer;

[0021] Step S2: Determine the spectrum range and access mode of different planes according to the requirements of different planes of the self-organizing network and the plane-level resource allocation rules;

[0022] Step S3: Monitor node status and link quality, and perform survivability enhancements upon node failure or link termination. Specifically, the self-organizing network continuously monitors node status and link quality. Upon detecting a node failure or link interruption, it immediately triggers a routing repair algorithm to recalculate the link path. After calculating the new link path, the routing table of the affected node is updated, and the updated information is broadcast to all relevant nodes. Simultaneously, the affected node caches data and, after a new path is established, retransmits the cached data to the target node.

[0023] Preferably, the step S1 of establishing the architecture of the basic communication layer, the intelligent control layer, and the service execution layer is implemented according to the following steps:

[0024] Step S1.1: Node initialization and topology construction of the basic communication layer;

[0025] Each node in the self-organizing network periodically broadcasts a broadcast frame containing its own identity, location, and energy status. Adjacent nodes receive the broadcast frames, extract, and record the information of the sending node. At the same time, each node sends an active detection frame to its neighboring nodes, and the neighboring nodes respond to the active detection frame. Nodes send and receive node information with each other, and based on the collected neighbor node information, the initial topology of the self-organizing network is constructed.

[0026] Step S1.2: Dynamic architecture and optimization of the intelligent control layer;

[0027] The command node uses a phased array antenna to scan the entire self-organizing network and collect the topology information of the self-organizing network; based on the topology information, it generates an initial command node-region controller-ordinary node control link; based on the initial control link, the command node identifies the cluster head node of each cluster in the self-organizing network, and the command node configures the regional controller served by the cluster head node according to the scale of each cluster, the distribution of nodes and business needs. The regional controller collects the status data of the nodes in the cluster in real time, and pre-processes the collected data, extracts key information and compresses the data. The control layer continuously monitors the status of the self-organizing network. When the status of the self-organizing network changes, the command node adjusts the control strategy and generates new control instructions. The control instructions are sent to the regional controller via the phased array antenna. The regional controller converts the control instructions into specific configuration instructions and distributes them to the ordinary nodes in the cluster. After receiving the configuration instructions, the ordinary nodes execute the instructions and feedback the execution results to the regional controller. The regional controller summarizes the feedback information and reports it to the command node to form a closed-loop control.

[0028] Step S1.3: Dynamic task deployment and management at the business execution layer;

[0029] The command node receives external commands and parses them to obtain task parameters, including task type, execution time, target area, latency, and bandwidth requirements. It then adds tasks to the scheduling queue based on their priority. It then selects qualified network nodes based on the task's resource requirements. Finally, it plans the subnet topology based on the task type and network node distribution, allocating resources such as communication frequency bands, time slots, and bandwidth to the network nodes. Once configured, the network node enters the activated state, the subnet is put into use, and the task begins executing.

[0030] Preferably, the node initialization and topology construction of the basic communication layer in step S1.1 are implemented according to the following steps:

[0031] Step S1.1.1: Self-discovery of self-organizing network nodes and construction of neighbor node lists;

[0032] Nodes in a self-organizing network periodically broadcast frames containing their own identification, location, and energy status. After receiving the broadcast frames, neighboring nodes extract and record the information of the sending node, including the identification, location, and energy status, and build a preliminary list of neighbor nodes.

[0033] Step S1.1.2: Initialize the cluster structure and elect the cluster head node;

[0034] All nodes randomly select an initial cluster and send a joining request to the selected cluster head node to initially form a cluster structure. After the initial cluster structure is formed, each cluster node in the cluster votes based on the comprehensive performance indicators of the cluster head candidate cluster nodes. The multiple cluster nodes with the highest number of votes jointly serve as cluster heads, forming a multi-center architecture with multiple cluster head nodes in the cluster.

[0035] Preferably, the dynamic architecture and optimization of the intelligent control layer in step S1.2 is implemented according to the following steps:

[0036] Step S1.2.1: Initialization and scanning of the accusation node;

[0037] During the construction of the intelligent control layer, the command node is activated. The command node uses the phased array antenna to scan the entire network and collect the self-organizing network topology information. The phased array antenna quickly obtains the spatial location information of all nodes in the self-organizing network, the node's remaining energy and other status information.

[0038] Step S1.2.2: Generate an initial control link based on the self-organizing network topology information in step S1.2.1;

[0039] After obtaining the full network topology information of the self-organizing network, the initial control link of the command node-region controller-ordinary node is generated;

[0040] Step S1.2.3: The command node identifies the cluster head node of each cluster in the self-organizing network, elects the regional controller and configures it;

[0041] After generating the initial control link, the command node identifies the cluster head node of each cluster in the self-organizing network. The cluster head node also serves as the regional controller. The command node configures the regional controller based on the size of each cluster, the distribution of nodes, and business needs. The configuration process includes assigning control tasks, setting communication parameters, and defining management policies.

[0042] Step S1.2.4: The regional controller collects the status data of the nodes in the cluster and performs preprocessing;

[0043] The regional controller collects the status data of the nodes in the cluster in real time and pre-processes the collected data to extract key information and compress the data to reduce the amount of data transmitted to the command node. The status data of the nodes in the cluster includes the node's energy consumption, communication load, and task execution progress. The pre-processing process includes data cleaning and feature extraction.

[0044] Step S1.2.5: Continuously monitor the status of the self-organizing network;

[0045] The intelligent control layer continuously monitors the state of the self-organizing network and, upon changes in the self-organizing network state, instructs the nodes to adjust their control strategies and generate new control instructions. Control strategy adjustments include reconfiguring regional controller tasks, adjusting link bandwidth, and optimizing data transmission paths. Self-organizing network state changes include node additions and departures, changes in link quality, and fluctuations in service demand.

[0046] Step S1.2.6: Distribute the control instructions in step S1.2.5 to ordinary nodes in the cluster and execute them;

[0047] The command node generates new control instructions based on the adjusted control strategy. The control instructions are sent to the regional controller through the phased array antenna. The regional controller converts the control instructions into specific configuration instructions and distributes them to ordinary nodes in the cluster. After receiving the configuration instructions, the ordinary nodes perform the corresponding operations and feed back the execution results to the regional controller. After summarizing the feedback information, the regional controller reports it to the command node, forming a closed-loop control.

[0048] Preferably, the step S1.3 is implemented by dynamically deploying and managing tasks on the business execution layer according to the following steps:

[0049] Step S1.3.1: receiving and parsing external instructions;

[0050] The command node receives instructions from entities outside the self-organizing network service execution layer to instruct the self-organizing network to perform specific tasks or operations. At the same time, the command node parses the task instructions and extracts key task parameters.

[0051] Step S1.3.2: Task scheduling and resource matching;

[0052] Add tasks to the scheduling queue based on their priority and urgency, and then analyze the network resource requirements of the tasks based on their parameters.

[0053] Step S1.3.3: Create and configure the task subnet;

[0054] Based on the task requirements in step S1.3.2 and the distribution of the selected subnet nodes, a task subnet is created and an independent space-time-frequency resource slice is allocated to the task subnet. At the same time, the command node sends configuration instructions to the task subnet. The subnet nodes are configured accordingly according to the configuration instructions and enter the task-ready state. The topology of the task subnet can be star or mesh. The configuration instructions include communication parameter settings, task execution program loading, and resource allocation.

[0055] Step S1.3.4: the task subnet executes the task;

[0056] After the configuration is completed, the task subnet executes the task, and the command node monitors the task execution status in real time through the regional controller. The task execution status includes data transmission progress, node energy consumption, and task completion status;

[0057] Step S1.3.5: Task completion and resource recovery;

[0058] When the task goal is achieved or the task execution time ends, the task is considered completed. After the task is completed, the command node issues a resource recovery instruction, and the subnet nodes in the task subnet release the allocated resources, including spectrum, time slots, and computing resources. After resource recovery, the subnet nodes return to the idle state and can be reused by other tasks.

[0059] The technical solution of this application has at least the following advantages and beneficial effects:

[0060] 1. In this invention, the self-organizing network is divided into three independent network plane systems: the basic communication layer, the intelligent control layer, and the service execution layer. The basic communication layer adopts a distributed multi-center clustering architecture, generating cluster structures based on an improved K-means algorithm. Each cluster node within a cluster votes based on the comprehensive performance indicators of the cluster head candidate. The cluster nodes with the highest number of votes jointly serve as cluster heads, forming a multi-center architecture with multiple cluster head nodes within the cluster. This architecture enhances the fault tolerance and load balancing of the self-organizing network and prevents a cluster node from being selected as the sole cluster head simply because it performs well on a single performance indicator.

[0061] The intelligent control layer's command and control nodes are equipped with phased array antennas, forming a hierarchical control chain from command and control nodes to regional controllers and then to common nodes. This hierarchical control chain architecture enables efficient management and flexible configuration of self-organizing network resources.

[0062] The business execution layer utilizes a microservices-based architecture, encapsulating different network functions and business logic into independent microservice modules that can communicate with each other. Furthermore, microservice modules can quickly create and destroy task subnets based on business needs. Microservice modules also support dynamic loading and unloading, dynamically loading required modules or unloading unnecessary modules at runtime based on changing task requirements, improving resource utilization and flexibility in the self-organizing network system.

[0063] In general, the basic communication layer, intelligent control layer, and business execution layer operate independently, avoiding cross-layer interference through resource isolation and enabling flexibly allocating resources of the self-organizing network.

[0064] 2. In the present invention, the self-organizing network continuously monitors the node status and link quality. When it detects that a node fails or the link is interrupted, it can recalculate the link path and generate a new link path. At the same time, the failed node caches data. After the new path is established, the cached data is retransmitted to the target node. Therefore, the self-organizing network has anti-destruction properties, and the operation of the self-organizing network is not affected when some nodes fail or the link is interrupted. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a block diagram of the self-organizing network dynamic slicing architecture of the present invention;

[0066] Figure 2 This is a flow chart of the control method of the present invention;

[0067] Figure 3 This is a flow chart of the architecture of the basic communication layer, intelligent control layer, and service execution layer of the present invention;

[0068] Figure 4 The dynamic architecture and optimization flow chart of the intelligent control layer of the present invention;

[0069] Figure 5 This is a flow chart of dynamic task deployment and management of the business execution layer of the present invention. DETAILED DESCRIPTION

[0070] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention will be briefly introduced below in conjunction with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0071] It should be understood that although the terms "first," "second," and so forth, may be used herein to describe various modules, these modules should not be limited by these terms. These terms are merely used to distinguish one module from another. For example, a first module may be referred to as a second module, and similarly, a second module may be referred to as a first module without departing from the scope of the exemplary embodiments of the present invention.

[0072] It should be understood that the term "and / or" that may appear in this document is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may indicate three situations: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" that may appear in this document describes another type of association object relationship, indicating that two relationships may exist. For example, A / and B may indicate two situations: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the previous and subsequent associated objects are in an "or" relationship.

[0073] Example:

[0074] Please refer to Figure 1-Figure 5 One aspect of the present invention provides a dynamic slicing architecture for self-organizing networks based on software-defined radio, including a layered three-dimensional network plane system.

[0075] Furthermore, the layered three-dimensional network plane system includes a basic communication layer, an intelligent control layer, and a business execution layer.

[0076] Specifically, the basic communication layer adopts a multi-distributed, multi-center clustering architecture, dynamically generating cluster structures based on an improved K-means algorithm. Cluster heads are elected using a multi-round voting mechanism based on the overall performance of cluster nodes. The basic communication layer is responsible for the network's fundamental communication functions, ensuring stable connections and data transmission between network nodes.

[0077] Specifically, the cluster nodes of the basic communication layer include the cluster head node and common cluster nodes except the cluster head node.

[0078] More specifically, cluster nodes are the basic building blocks of self-organizing networks. Specifically, they represent the various devices and terminals within the network. Cluster nodes collaborate with each other to achieve network communication and data transmission. In this solution, multi-dimensional information about cluster nodes is used as a feature vector and input into an improved K-means algorithm to determine the optimal cluster structure. This multi-dimensional information includes the geographic location, remaining energy, and communication capabilities of the coarse cluster nodes.

[0079] It should be noted that the traditional K-means algorithm is a partition-based clustering algorithm that iteratively optimizes cluster centers and cluster members to maximize the similarity of data points within cluster nodes and minimize the similarity between cluster nodes. The traditional K-means algorithm is well-known and readily understood by those skilled in the art, so it will not be described in detail here.

[0080] More specifically, in the present invention, the improvements to the traditional K-means algorithm lie in three aspects: energy perception, dynamic adjustment, and multi-center election.

[0081] More specifically, the improvement in energy perception refers to the introduction of a cluster node residual energy weight factor when calculating the similarity between cluster nodes, which can prevent the cluster head node from failing quickly due to excessive energy consumption and extend the stable operation period of the self-organizing network. The residual energy weight factor is the cluster node residual energy E p The ratio of the initial energy E0 of the cluster node is:

[0082]

[0083] It's important to note that with the introduction of the residual energy weighting factor, the algorithm considers not only the geographical location of cluster nodes but also their residual energy status when calculating similarity between cluster nodes. Specifically, when calculating the similarity between two cluster nodes, their residual energy weighting factors are compared, and the cluster node with more residual energy is given a greater weight. This gives the cluster node with more residual energy a relatively greater influence in cluster structure formation and cluster head election. This guides the algorithm to select cluster nodes with more sufficient energy as cluster heads, preventing cluster heads from rapidly failing due to excessive energy consumption.

[0084] More specifically, improvements in dynamic adjustment refer to real-time monitoring of cluster node status changes during the operation of the self-organizing network. When the degree of change in cluster nodes reaches a certain threshold, the K-means algorithm is triggered to re-cluster and dynamically adjust the cluster structure to ensure that the self-organizing network is always in good operating condition.

[0085] It should be noted that the thresholds include cluster node energy change thresholds and cluster node position change thresholds. When the remaining energy of a cluster node is less than 50% of its initial energy, the cluster node's energy state is determined to have changed significantly, triggering a re-clustering operation. For mobile cluster nodes, if their position changes by more than a certain distance (in a wireless communication environment, the moving distance exceeds 20% of the communication radius) or their moving speed exceeds a set value (5 meters / second), resulting in a decrease in the quality of their communication link with other nodes in the cluster, re-clustering is triggered to optimize the cluster structure.

[0086] More specifically, when performing the re-clustering operation, the improved K-means algorithm is run to update the latest status information of all cluster nodes. The latest status information includes the remaining energy of the cluster nodes and the spatial location of the cluster nodes. After the improved K-means algorithm data update is completed, the improved K-means algorithm is reinitialized, and the remaining energy and spatial location coordinates of the cluster nodes are normalized respectively. Then, based on the weight coefficients of the remaining energy and spatial location coordinates, the weighted average score of each cluster node is calculated. The cluster nodes are sorted according to the weighted average score, and the top K cluster nodes with the highest scores are selected as the initial cluster centers. Then, the improved K-means algorithm is run to calculate the distance between the cluster nodes and the cluster center based on the remaining energy and spatial location coordinates of the cluster nodes, and the cluster affiliation of the cluster nodes is updated. The distance measurement formula is:

[0087]

[0088] Among them, D i,j is the distance from node i to cluster center j, E r,i is the residual energy of node i, E 0,i is the initial energy of node i, d i,j is the distance between node i and cluster center j. α and β are weight coefficients, representing the residual energy and spatial coordinate weight of the cluster node, respectively. K is the number of cluster head nodes required and can be selected based on the actual needs of the cluster head nodes. After the algorithm iterative convergence, the cluster structure is updated based on the new cluster center and node affiliation relationships. Based on the new cluster structure, network resources are reallocated to adapt to dynamic network changes.

[0089] More specifically, improvements to multi-center elections involve the use of a multi-round voting mechanism to select cluster heads. This means that, based on the initial clustering, each cluster is assigned a temporary cluster head. During this multi-round voting process, each cluster node within the cluster votes based on the comprehensive performance metrics of the candidate cluster heads. The top cluster nodes share the role of cluster head, forming a multi-center architecture with multiple cluster heads within the cluster. This further enhances the fault tolerance and load balancing of the self-organizing network and prevents a cluster node from being selected as the sole cluster head simply because of its outstanding performance on a single performance metric. The comprehensive performance of candidate cluster heads includes signal strength, computing power, storage resources, and remaining energy.

[0090] More specifically, the basic communication layer integrates spatiotemporal reference synchronization technology. Specifically, cluster nodes in the basic communication layer have built-in GPS receiver modules that can receive satellite signals in real time and obtain their own geographic location information. Furthermore, Ethernet-based communication links are established between cluster nodes, enabling inter-cluster time synchronization by sending and receiving precise time protocol messages. Adaptive filtering and multipath compensation algorithms are used to optimize the time protocol synchronization process, achieving nanosecond-level time synchronization across the entire network.

[0091] It should be noted that the adaptive filtering algorithm and the multipath compensation algorithm are existing technologies, which can be understood by those skilled in the art and will not be described in detail here.

[0092] More specifically, the communication protocol of the self-organizing network adopts a hybrid routing protocol, which includes the optimized link state routing protocol (OLSR) and the on-demand distance vector routing protocol - enhanced (AODV-UU). During network operation, the optimized link state routing protocol (OLSR) or the on-demand distance vector routing protocol - enhanced (AODV-UU) is dynamically selected as the dominant routing protocol. Specifically, when the self-organizing network topology is relatively stable, the optimized link state routing protocol (OLSR) is used as the primary method to establish the global optimal route. When rapid changes in the local topology of the self-organizing network or sudden bursts of specific business traffic are detected, the on-demand distance vector routing protocol - enhanced (AODV-UU) responds quickly and establishes temporary on-demand routes to supplement and optimize the OLSR routing path.

[0093] More specifically, the communication signal transmission waveform of the self-organizing network adopts orthogonal frequency division multiplexing (OFDM) and single carrier frequency domain equalization (SC-FDE) adaptive switching, and can dynamically select OFDM or SC-FDE as the current transmission waveform based on real-time channel state information. Specifically, when the channel multipath effect is significant and spectrum resources are tight, OFDM is selected; when the channel frequency offset is large, SC-FDE is selected. Real-time channel state information includes multipath strength, signal-to-noise ratio, and frequency offset. Specifically, when the RMS delay spread is greater than 100ns and the available spectrum resources are less than 30%, OFDM is selected; when the carrier frequency offset is greater than 10% of the subcarrier spacing, it switches to SC-FDE.

[0094] Specifically, the basic communication layer is used to establish and maintain the physical connection of the network to ensure that data can be transmitted stably between nodes. On the other hand, it provides the basis for network connection and data transmission for the intelligent control layer. At the same time, the intelligent control layer manages and optimizes the basic communication layer.

[0095] Specifically, the intelligent control layer is responsible for the intelligent management and resource scheduling of the self-organizing network. The intelligent control layer contains multiple command and control nodes, which are equipped with phased array antennas and AI decision-making engines and have strong communication capabilities and intelligent decision-making capabilities.

[0096] More specifically, when the basic communication layer detects a change in the network structure or configuration, the cluster head node sends the change information to the intelligent control layer. The intelligent control layer replans the network structure based on the change and sends the updated configuration information back to the basic communication layer. A command node refers to a device with command, control, communication, and computing capabilities in a communication network. It is the core control point of a self-organizing network and is responsible for managing and coordinating other nodes in the network. The intelligent control layer sends control instructions to the cluster head node of the basic communication layer through the command node. These control instructions include instructions for adjusting communication parameters and optimizing routing. The network structure or configuration includes the addition, reduction, and movement of cluster nodes, the establishment and disconnection of links, changes in link quality, equipment failures, and link failures.

[0097] More specifically, the phased array antenna features high gain, wide beam coverage, and fast beam switching capabilities. Its operating frequency range covers major wireless communication bands, adapting to the communication needs of various business scenarios. The gain of the phased array antenna can reach over 30dBi, and the beam width is adjustable between 10° and 60°. The AI ​​decision engine boasts high concurrent processing capabilities and low-latency response, capable of processing the status information of thousands of command and control nodes in real time, making optimal resource allocation decisions and ensuring efficient network operation.

[0098] More specifically, the topology of the intelligent control layer is based on the idea of ​​software-defined networking (SDN), constructing an SDN hierarchical control chain of "command and control node-region controller-ordinary node", in which the region controller is concurrently served by the cluster head node of the basic communication layer.

[0099] More specifically, ordinary nodes represent devices with basic communication functions in the network, which can transmit data with other nodes. Ordinary nodes are controlled by regional controllers. The ordinary nodes here and the ordinary cluster nodes in the basic communication layer are essentially the same device or terminal, but the intelligent control layer emphasizes that ordinary nodes are controlled by regional controllers, and the basic communication layer focuses more on the identity of ordinary cluster nodes as belonging to a cluster rather than the cluster head node, so they are called differently.

[0100] More specifically, the command node is the core control node of the self-organizing network, responsible for issuing control commands to the regional controllers to control ordinary nodes. Furthermore, the command node is equipped with a phased array antenna and an AI decision engine, responsible for global network resource management and command distribution. The command node communicates with the regional controllers via the phased array antenna, obtaining real-time information about the self-organizing network topology and the status of each node, and coordinating the work of the regional controllers.

[0101] More specifically, the regional controller, which doubles as the cluster head node in the basic communication layer, manages communication and resource allocation among cluster nodes. The regional controller communicates with the command and control nodes via phased array antennas, receives control instructions from the command and control nodes, converts them into configuration commands, and issues them to the normal nodes in the cluster. The regional controller possesses a certain level of local decision-making capability, enabling rapid recovery and adjustment in the event of some network anomalies.

[0102] More specifically, ordinary nodes communicate and transmit data according to the configuration instructions of the regional controller, and have computing and storage capabilities, and can perform simple data processing and caching.

[0103] More specifically, the SDN controller runs on a standalone server within the self-organizing network and collaborates with regional controllers to achieve intelligent management and control of the self-organizing network. The SDN controller communicates with the regional controllers using the OpenFlow protocol, enabling flexible control of network traffic and dynamic resource allocation. Specifically, the SDN controller adjusts the link bandwidth and priority between regional controllers in real time based on network traffic changes, ensuring efficient network operation.

[0104] More specifically, the control instruction transmission mechanism of the intelligent control layer is time division duplex (TDD) combined with reserved access.

[0105] More specifically, control command transmission utilizes time division duplex (TDD) mode, dividing the wireless channel into uplink and downlink time slots. Time slot allocation enables bidirectional data transmission, dynamically adjusting the uplink / downlink time slot ratio based on network service requirements and the communication environment. During command transmission, a reserved access mechanism is employed to reserve dedicated time slots and spectrum resources for critical commands. This pre-allocation of specific time slots and spectrum resources ensures the reliable transmission of critical commands, preventing command loss or delays due to network congestion or other interference.

[0106] Specifically, the intelligent control layer is responsible for generating and issuing control instructions, supervising and managing the execution process of the business execution layer, receiving feedback from the business execution layer, and adjusting task instructions based on the feedback information; the business execution layer is responsible for accepting control instructions from the intelligent control layer and executing the control instructions, and at the same time feeding back the instruction task execution status and results to the intelligent control layer.

[0107] Specifically, the business execution layer adopts a microservices-based architecture, encapsulating different network functions and business logic into independent microservice modules. Specifically, these microservice modules include a reconnaissance task module, a relay task module, and an attack task module. The reconnaissance task module is responsible for image data acquisition and transmission; the relay task module is responsible for data relay; and the attack task module is responsible for executing attack operations against targets. Microservice modules interact with each other via a lightweight communication protocol (RESTful API). Each microservice module reserves a set of RESTful interfaces that other modules can call via HTTP requests, enabling inter-module communication and writing. Microservice modules can rapidly create and destroy task subnets based on business needs. A task subnet consists of multiple microservice modules responsible for different functions, used to complete specific tasks. Furthermore, microservice modules support dynamic loading and unloading, allowing them to dynamically load required modules or unload unneeded modules at runtime based on changing task requirements, improving resource utilization and flexibility of the self-organizing network system. Task subnets are dynamically created and destroyed based on business needs.

[0108] More specifically, a microservice module is a logical unit consisting of multiple subnet nodes that work together to implement specific network functions and business logic. A subnet node can be an ordinary node or a cluster head node.

[0109] More specifically, the service execution layer uses a multi-dimensional tag matching mechanism for subnet nodes. This mechanism defines multi-dimensional tags for each subnet node in the self-organizing network. This multi-dimensional tag matching enables cross-subnet node resource sharing. Furthermore, each subnet node is allocated an independent space-time-frequency resource slice to ensure resource isolation and efficient utilization between subnet nodes.

[0110] More specifically, a multi-dimensional label is defined for each subnet node, including computing power, communication radius, energy status, and task type. Using a label matching algorithm, the optimal subnet node is found based on business needs and the availability of self-organizing network resources. Each subnet node is then allocated an independent space-time-frequency resource slice. This enables resource sharing and collaboration across subnet nodes, improving resource utilization and task execution efficiency.

[0111] More specifically, space-time-frequency resource slicing includes space resource slicing, time slot resource slicing, and spectrum resource slicing. Space resource slicing utilizes spatial multiplexing technology to allocate independent space resources to each subnet node, improving spatial resource utilization. Time slot resource slicing slices resources in the time dimension, allocating specific time slots to each subnet node to avoid resource conflicts. Spectrum resource slicing allocates independent spectrum resources to each subnet node based on service needs and spectrum availability.

[0112] More specifically, when performing resource slicing in the time dimension, the system time of the self-organizing network is divided into several time slots. Each time slot has a fixed duration. The time slot is the basic time unit for resource allocation. The resources of the self-organizing network are allocated to different subnet nodes within the time slot to maximize spectrum efficiency and service isolation.

[0113] More specifically, the resource slices of each subnet node are dynamically adjusted according to the changes in business needs in the self-organizing network. Thus, resource isolation between subnet nodes is achieved through independent space-time-frequency resource slices, effectively avoiding resource interference and conflicts.

[0114] It should be noted that spatial multiplexing technology is a technology that utilizes the characteristics of multipath propagation to transmit multiple data streams through different spatial channels at the same frequency, which is an existing technology; in addition, spectrum resource slicing is a spectrum allocation technology commonly used in the field of wireless communications. The specific spectrum slicing process will not be described here in detail, and those skilled in the art can understand it.

[0115] It should be noted that the tag matching algorithm is an existing technology, and its operation and tag matching process will not be described in detail here, and those skilled in the art can understand it.

[0116] Furthermore, the layered three-dimensional network plane system also includes a hardware decoupling and software definition system, and the hardware decoupling and software definition system includes a reconfigurable hardware platform and a waveform library.

[0117] Specifically, the reconfigurable hardware platform provides the operating basis, and the waveform library dynamically loads the protocol stack to adapt to communication requirements.

[0118] Specifically, the reconfigurable hardware platform is a heterogeneous computing architecture based on FPGA, ARM, and RF front-end. Specifically, FPGA provides efficient parallel computing capabilities and flexible hardware programmability, which can accelerate specific signal processing tasks. The present invention uses Xilinx UltraScale+ series FPGAs; the ARM processor is responsible for running control software and executing complex algorithms, including AI decision engines and resource scheduling algorithms. The present invention uses a quad-core ARM Cortex-A72 processor with a main frequency of 1.5GHz, equipped with 4GB DDR4 memory and 64GB eMMC storage to ensure that the system can smoothly run multiple tasks; the RF front-end integrates a multi-mode RF module, which is responsible for signal transmission and reception. The multi-mode RF module supports full frequency band coverage from 2GHz to 60GHz.

[0119] More specifically, the reconfigurable hardware platform provides hardware support for the operation of the waveform library. The reconfigurable hardware platform supports dynamic loading technology of the waveform library, and switches between different communication protocol stacks through the collaborative work of FPGA and ARM processor.

[0120] Specifically, the waveform library includes 50 communication protocol stacks. Dynamic loading technology enables switching between different communication protocol stacks, including LTE-D2D, Wi-Fi 6, and TSN time-sensitive networking.

[0121] It should be noted that dynamic loading technology is a software engineering strategy that allows a program to load required functional modules or resources on demand during runtime, rather than loading all content at once when the program starts. It is a technology that defines existing software radio and communication technologies. Those skilled in the art can understand it and will not be elaborated here.

[0122] Another aspect of the present invention provides a method for intelligent management and control of a dynamic slicing architecture in an ad hoc network based on software-defined radio, which is implemented by following the steps below:

[0123] Step S1: Architecture of basic communication layer, intelligent control layer, and business execution layer;

[0124] Step S2: According to the requirements of different planes of the self-organizing network, the spectrum range and access method of different planes are determined.

[0125] Specifically, in step S2, the plane-level resource allocation rules are as follows:

[0126]

[0127]

[0128] Step S3: Monitor the node status and link quality, and perform survivability enhancement processing after node failure or link termination.

[0129] In this step, the self-organizing network continuously monitors node status and link quality. Upon detecting a node failure or link interruption, the routing repair algorithm is immediately triggered to recalculate the link path. After calculating the new link path, the routing table of the affected node is updated and the updated information is broadcast to all relevant nodes. Simultaneously, the affected node caches data and, once a new path is established, retransmits the cached data to the target node.

[0130] It should be noted that the route repair algorithm is an existing technology, and the detailed process of calculating a new path using the route repair algorithm will not be described here in detail, as it can be understood by those skilled in the art.

[0131] When a network node fails or a link is interrupted, route repair is triggered and can complete path recalculation within 100ms.

[0132] Furthermore, the architecture of the basic communication layer, intelligent control layer, and service execution layer in step S1 of the present invention is implemented according to the following steps:

[0133] Step S1.1: Node initialization and topology construction of the basic communication layer;

[0134] In this step, each node in the ad hoc network activates Bluetooth low energy scanning and periodically broadcasts frames containing its identity, location, and energy status. Upon receiving these frames, neighboring nodes extract and record the information about the sending node. Simultaneously, each node in the ad hoc network sends active probe frames to neighboring nodes, requesting responses to obtain more detailed neighbor information and confirm their reachability. Based on this collected neighbor node information, an initial topology map of the ad hoc network is constructed.

[0135] In the present invention, step S1.1 is implemented as follows:

[0136] Step S1.1.1: Self-discovery of self-organizing network nodes and construction of neighbor node lists;

[0137] Specifically, nodes in a self-organizing network periodically broadcast broadcast frames containing their own identification, location, and energy status. After receiving the broadcast frames, adjacent nodes extract and record the information of the sending node, which includes identification, location, and energy status, and build a preliminary list of neighbor nodes.

[0138] Step S1.1.2: Initialize the cluster structure and elect the cluster head node;

[0139] Specifically, all nodes randomly select an initial cluster and send a joining request to the selected cluster head node to initially form a cluster structure; after the initial cluster structure is formed, each cluster node in the cluster votes based on the comprehensive performance indicators of the cluster head candidate cluster nodes, and the multiple cluster nodes with the highest number of votes jointly serve as cluster heads, forming a multi-center architecture with multiple cluster head nodes in the cluster.

[0140] Step S1.2: Dynamic architecture and optimization of the intelligent control layer;

[0141] Step S1.2 is implemented as follows:

[0142] Step S1.2.1: Initialization and scanning of the accusation node;

[0143] Specifically, during the construction of the intelligent control layer, the command node is activated. Using a phased array antenna, the command node scans the entire network and collects information about the self-organizing network topology. The high-gain, wide-beam coverage of the phased array antenna allows it to quickly obtain information about the spatial location of all nodes in the self-organizing network, including their remaining energy.

[0144] Step S1.2.2: Generate an initial control link based on the self-organizing network topology information in step S1.2.1;

[0145] Specifically, after obtaining the entire network topology information of the self-organizing network, an initial control link of the command node-region controller-ordinary node is generated.

[0146] Step S1.2.3: The command node identifies the cluster head node of each cluster in the self-organizing network, elects the regional controller and configures it;

[0147] Specifically, after establishing the initial control link, the command node identifies the cluster head node of each cluster in the self-organizing network. The cluster head node also serves as the regional controller, responsible for managing communication and resource allocation among nodes within the cluster. The command node configures the regional controller based on the size of each cluster, node distribution, and service requirements. This configuration process includes assigning control tasks, setting communication parameters, and defining management policies.

[0148] Step S1.2.4: The regional controller collects the status data of the nodes in the cluster and performs preprocessing;

[0149] Specifically, the regional controller collects status data from cluster nodes in real time and preprocesses the data, extracting key information and compressing it to reduce the amount of data transmitted to the command node. Cluster node status data includes energy consumption, communication load, and task execution progress. The preprocessing process includes data cleaning and feature extraction.

[0150] Step S1.2.5: Continuously monitor the status of the self-organizing network;

[0151] Specifically, the intelligent control layer continuously monitors the state of the self-organizing network and, upon changes in the self-organizing network state, instructs the nodes to adjust their control strategies and generate new control instructions. Control strategy adjustments include reconfiguring regional controller tasks, adjusting link bandwidth, and optimizing data transmission paths. Changes in the self-organizing network state include node additions and departures, changes in link quality, and fluctuations in service demand.

[0152] Step S1.2.6: Distribute the control instructions in step S1.2.5 to ordinary nodes in the cluster and execute them.

[0153] Specifically, the command node generates new control instructions based on the adjusted control strategy. The control instructions are sent to the regional controller via the phased array antenna. The regional controller converts the control instructions into specific configuration instructions and distributes them to ordinary nodes in the cluster. After receiving the configuration instructions, the ordinary nodes perform the corresponding operations and feed back the execution results to the regional controller. After summarizing the feedback information, the regional controller reports it to the command node, forming a closed-loop control.

[0154] Step S1.3: Dynamic task deployment and management at the business execution layer.

[0155] In this step, the command node receives external commands and parses them to obtain task parameters, including task type, execution time, target region, latency, and bandwidth requirements. Tasks are then added to the scheduling queue based on their priority. Qualified network nodes are then selected based on the task's resource requirements. Finally, the subnet topology is planned based on the task type and network node distribution, allocating resources such as communication frequency bands, time slots, and bandwidth to the network nodes. Once configuration is complete, the network nodes enter the activated state, the subnet becomes operational, and the task begins executing.

[0156] In the present invention, step S1.3 is implemented as follows:

[0157] Step S1.3.1: receiving and parsing external instructions;

[0158] Specifically, the command node receives external mission instructions. External mission instructions are instructions issued by entities outside the ad hoc network's service execution layer, directing the ad hoc network to perform specific tasks or operations. The command node parses the mission instructions and extracts key mission parameters. These parameters include the mission type (reconnaissance, relay, attack), the mission execution time window, the target area's location, the mission's latency, and the required bandwidth.

[0159] Step S1.3.2: Task scheduling and resource matching;

[0160] Specifically, tasks are added to the scheduling queue based on their priority and urgency. Then, based on the task parameters, the network resource requirements for the task are analyzed, including spectrum resources, computing power, storage resources, and communication bandwidth. Based on the network resources required for the task, subnet nodes that meet the task's resource requirements are matched.

[0161] Step S1.3.3: Create and configure the task subnet;

[0162] Specifically, based on the task requirements and the distribution of subnet nodes selected in step S1.3.2, a task subnet is created and allocated independent space-time-frequency resource slices. Simultaneously, the command node sends configuration instructions to the task subnet. The subnet nodes are configured accordingly and enter a task-ready state. The task subnet topology can be star or mesh; configuration instructions include setting communication parameters, loading the task execution program, and allocating resources.

[0163] Step S1.3.4: the task subnet executes the task;

[0164] Specifically, after configuration is complete, the task subnet begins executing the task. Meanwhile, the command node monitors the task execution status in real time through the regional controller. The task execution status includes data transmission progress, node energy consumption, and task completion status.

[0165] Step S1.3.5: Task completion and resource recovery.

[0166] Specifically, a task is considered complete when its objectives are achieved or the task execution time expires. Upon completion, the command node issues a resource reclaim instruction, and the subnet nodes within the task subnet release allocated resources, including spectrum, time slots, and computing resources. After resource reclaim, the subnet nodes return to an idle state and can be reused for other tasks.

[0167] Thus far, various embodiments of the present invention have been described in detail. To avoid obscuring the concept of the present invention, some details well known in the art have been omitted. Based on the above description, those skilled in the art will fully understand how to implement the technical solutions of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A dynamic slicing architecture for self-organizing networks based on software-defined radio, characterized by: Including layered three-dimensional network plane system; The layered three-dimensional network plane system includes a basic communication layer, an intelligent control layer, and a service execution layer; The basic communication layer adopts a multi-distribution multi-center clustering architecture, and the cluster nodes of the basic communication layer include cluster head nodes and ordinary nodes other than cluster head nodes; The cluster structure is generated based on the comprehensive performance of cluster nodes, and the cluster head node is elected based on the comprehensive performance of cluster nodes using a multi-round voting mechanism; Comprehensive performance includes signal strength, computing power, storage resources, and remaining energy; The intelligent control layer includes multiple command nodes, each of which is equipped with a phased array antenna and an AI decision engine; When the basic communication layer detects that the self-organizing network structure or configuration information has changed, the cluster head node sends the change information to the control node of the intelligent control layer; Each command node of the intelligent control layer sends control instructions to each cluster head node of the basic communication layer according to the change information to re-plan the self-organizing network structure, thereby realizing the management of the basic communication layer by the intelligent control layer; The business execution layer is a microservice-based architecture that encapsulates different network functions and business logic into independent microservice modules. The intelligent control layer issues control instructions, and the business execution layer receives and executes the control instructions. At the same time, the intelligent control layer monitors the execution process of the business execution layer and receives feedback signals from the business execution layer. The microservice module is composed of subnet nodes, which work together to form a microservice module that performs specific task functions and businesses; the microservice module includes a reconnaissance task module, a relay task module, and an attack task module.

2. The software-defined radio-based ad hoc network dynamic slicing architecture according to claim 1, characterized in that: Each cluster node in the basic communication layer has a built-in GPS receiving module, and receives satellite signals in real time through the built-in GPS receiving module to obtain the location information of the cluster node; an Ethernet-based communication link is established between each cluster node to realize information exchange between cluster nodes; During information exchange, the information packet sent by the cluster node includes the time and location information of the cluster node. After receiving the information packet, the other cluster nodes synchronize the time through the adaptive filtering algorithm and multipath compensation algorithm to make the time base of the cluster nodes the same; ensuring that each cluster node communicates and works collaboratively under the same time base.

3. The software-defined radio-based ad hoc network dynamic slicing architecture according to claim 1, characterized in that: The communication protocol of the self-organizing network adopts a hybrid routing protocol, and the hybrid routing protocol includes an optimized link state routing protocol and an on-demand distance vector routing protocol-enhanced; During the operation of the self-organizing network, one of the above two communication protocols is dynamically selected as the dominant routing protocol; the dynamic selection rules are as follows: when the topology of the self-organizing network is stable, the optimized link state routing protocol is selected as the main protocol to establish the global optimal route. At this time, the on-demand distance vector routing protocol-enhanced auxiliary main communication protocol is used; when rapid changes in the local topology of the self-organizing network or sudden bursts of specific business traffic are detected, the on-demand distance vector routing protocol-enhanced rapid response is used as the dominant communication protocol.

4. The software-defined radio-based ad hoc network dynamic slicing architecture according to claim 1, characterized in that: The communication signal transmission waveforms of the self-organizing network include OFDM and SC-FDE, and the two signal transmission waveforms are switched according to real-time channel status information; when the RMS delay spread is greater than 100ns and the available spectrum resources are less than 30%, OFDM is selected; when the carrier frequency offset is greater than 10% of the subcarrier spacing, it switches to SC-FDE; the real-time channel status information includes multipath strength, signal-to-noise ratio, and frequency offset.

5. The software-defined radio-based ad hoc network dynamic slicing architecture according to claim 1, characterized in that: The business execution layer adopts a microservice-based architecture. In the business execution layer, the self-organizing network subnet nodes responsible for different network functions and business logics are encapsulated into independent microservice modules. The microservice modules responsible for the same business can form a task subnet. The task subnet loads and unloads the corresponding microservice modules according to the creation and destruction requirements of new tasks; the microservice modules include reconnaissance task module, relay task module and attack task module; the reconnaissance task module, relay task module and attack task module are responsible for image acquisition and transmission, data relay and target attack respectively.

6. A method for managing and controlling a dynamic slicing architecture of a self-organizing network based on software-defined radio, characterized in that: The control method comprises the following steps: Step S1: Establish the architecture of basic communication layer, intelligent control layer, and business execution layer; Step S2: Determine the spectrum range and access mode of different planes according to the requirements of different planes of the self-organizing network and the plane-level resource allocation rules; Step S3: Monitor node status and link quality, and perform survivability enhancements upon node failure or link termination. Specifically, the self-organizing network continuously monitors node status and link quality. Upon detecting a node failure or link interruption, it immediately triggers a routing repair algorithm to recalculate the link path. After calculating the new link path, the routing table of the affected node is updated, and the updated information is broadcast to all relevant nodes. Simultaneously, the affected node caches data and, after a new path is established, retransmits the cached data to the target node.

7. The method for controlling dynamic slicing architecture of a self-organizing network based on software-defined radio according to claim 6, characterized in that: The step S1 of establishing the architecture of the basic communication layer, intelligent control layer, and service execution layer is implemented by the following steps: Step S1.1: Node initialization and topology construction of the basic communication layer; Each node in the self-organizing network periodically broadcasts a broadcast frame containing its own identity, location, and energy status. Adjacent nodes receive the broadcast frames, extract, and record the information of the sending node. At the same time, each node sends an active detection frame to its neighboring nodes, and the neighboring nodes respond to the active detection frame. Nodes send and receive node information with each other, and based on the collected neighbor node information, the initial topology of the self-organizing network is constructed. Step S1.2: Dynamic architecture and optimization of the intelligent control layer; The command node uses a phased array antenna to scan the entire self-organizing network and collect the topology information of the self-organizing network; based on the topology information, it generates an initial command node-region controller-ordinary node control link; based on the initial control link, the command node identifies the cluster head node of each cluster in the self-organizing network, and the command node configures the regional controller served by the cluster head node according to the scale of each cluster, the distribution of nodes and business needs. The regional controller collects the status data of the nodes in the cluster in real time, and pre-processes the collected data, extracts key information and compresses the data. The control layer continuously monitors the status of the self-organizing network. When the status of the self-organizing network changes, the command node adjusts the control strategy and generates new control instructions. The control instructions are sent to the regional controller via the phased array antenna. The regional controller converts the control instructions into specific configuration instructions and distributes them to the ordinary nodes in the cluster. After receiving the configuration instructions, the ordinary nodes execute the instructions and feedback the execution results to the regional controller. The regional controller summarizes the feedback information and reports it to the command node to form a closed-loop control. Step S1.3: Dynamic task deployment and management at the business execution layer; The command node receives external commands and parses them to obtain task parameters, including task type, execution time, target area, latency, and bandwidth requirements. It then adds tasks to the scheduling queue based on their priority. It then selects qualified network nodes based on the task's resource requirements. Finally, it plans the subnet topology based on the task type and network node distribution, allocating resources such as communication frequency bands, time slots, and bandwidth to the network nodes. Once configuration is complete, the network nodes enter an activated state, the subnet becomes operational, and the task begins executing.

8. The method for controlling a dynamic slicing architecture of a self-organizing network based on software-defined radio according to claim 6, characterized in that: In step S1.1, the node initialization and topology construction of the basic communication layer are implemented according to the following steps: Step S1.1.1: Self-discovery of self-organizing network nodes and construction of neighbor node lists; Nodes in a self-organizing network periodically broadcast frames containing their own identification, location, and energy status. After receiving the broadcast frames, neighboring nodes extract and record the information of the sending node, including the identification, location, and energy status, and build a preliminary list of neighbor nodes. Step S1.1.2: Initialize the cluster structure and elect the cluster head node; All nodes randomly select an initial cluster and send a join request to the selected cluster head node to initially form a cluster structure; After the initial cluster structure is formed, each cluster node in the cluster votes based on the comprehensive performance indicators of the cluster head candidate cluster node. The cluster nodes with the highest number of votes jointly serve as cluster heads, forming a multi-center architecture with multiple cluster head nodes in the cluster.

9. The method for controlling dynamic slicing architecture of a self-organizing network based on software-defined radio according to claim 6, characterized in that: The dynamic architecture and optimization of the intelligent control layer in step S1.2 are implemented as follows: Step S1.2.1: Initialization and scanning of the accusation node; During the construction of the intelligent control layer, the command node is activated. The command node uses the phased array antenna to scan the entire network and collect the self-organizing network topology information. The phased array antenna quickly obtains the spatial location information of all nodes in the self-organizing network, the node's remaining energy and other status information. Step S1.2.2: Generate an initial control link based on the self-organizing network topology information in step S1.2.1; After obtaining the full network topology information of the self-organizing network, the initial control link of the command node-region controller-ordinary node is generated; Step S1.2.3: The command node identifies the cluster head node of each cluster in the self-organizing network, elects the regional controller and configures it; After generating the initial control link, the command node identifies the cluster head node of each cluster in the self-organizing network. The cluster head node also serves as the regional controller. The command node configures the regional controller based on the size of each cluster, the distribution of nodes, and business needs. The configuration process includes assigning control tasks, setting communication parameters, and defining management policies. Step S1.2.4: The regional controller collects the status data of the nodes in the cluster and performs preprocessing; The regional controller collects the status data of the nodes in the cluster in real time and pre-processes the collected data to extract key information and compress the data to reduce the amount of data transmitted to the command node. The status data of the nodes in the cluster includes the node's energy consumption, communication load, and task execution progress. The preprocessing process includes data cleaning and feature extraction; Step S1.2.5: Continuously monitor the status of the self-organizing network; The intelligent control layer continuously monitors the state of the self-organizing network and, upon changes in the self-organizing network state, instructs the nodes to adjust their control strategies and generate new control instructions. Control strategy adjustments include reconfiguring regional controller tasks, adjusting link bandwidth, and optimizing data transmission paths. Self-organizing network state changes include node additions and departures, changes in link quality, and fluctuations in service demand. Step S1.2.6: Distribute the control instructions in step S1.2.5 to ordinary nodes in the cluster and execute them; The command node generates new control instructions based on the adjusted control strategy. The control instructions are sent to the regional controller through the phased array antenna. The regional controller converts the control instructions into specific configuration instructions and distributes them to ordinary nodes in the cluster. After receiving the configuration instructions, the ordinary nodes perform the corresponding operations and feed back the execution results to the regional controller. After summarizing the feedback information, the regional controller reports it to the command node, forming a closed-loop control.

10. The method for controlling a dynamic slicing architecture of a self-organizing network based on software-defined radio according to claim 6, characterized in that: Step S1.3 of dynamically deploying and managing tasks at the business execution layer is implemented as follows: Step S1.3.1: receiving and parsing external instructions; The command node receives instructions from entities outside the self-organizing network service execution layer to instruct the self-organizing network to perform specific tasks or operations. At the same time, the command node parses the task instructions and extracts key task parameters. Step S1.3.2: Task scheduling and resource matching; Add tasks to the scheduling queue based on their priority and urgency, and then analyze the network resource requirements of the tasks based on their parameters. Step S1.3.3: Create and configure the task subnet; Based on the task requirements in step S1.3.2 and the distribution of the selected subnet nodes, a task subnet is created and an independent space-time-frequency resource slice is allocated to the task subnet. At the same time, the command node sends configuration instructions to the task subnet. The subnet nodes are configured accordingly according to the configuration instructions and enter the task-ready state. The topology of the task subnet can be star or mesh. The configuration instructions include communication parameter settings, task execution program loading, and resource allocation. Step S1.3.4: the task subnet executes the task; After the configuration is completed, the task subnet executes the task, and the command node monitors the task execution status in real time through the regional controller. The task execution status includes data transmission progress, node energy consumption, and task completion status; Step S1.3.5: Task completion and resource recovery; When the task goal is achieved or the task execution time ends, the task is considered completed. After the task is completed, the command node issues a resource recovery instruction, and the subnet nodes in the task subnet release the allocated resources, including spectrum, time slots, and computing resources. After resource recovery, the subnet nodes return to the idle state and can be reused by other tasks.

Images