Intelligent Routing Planning Method and System for Self-Organizing Networks

Abstract

This invention relates to the field of wireless communication technology, specifically to an intelligent routing planning method and system for ad hoc networks. The method includes: constructing an ad hoc network strategy with a dual-link mechanism, achieving central node election and network construction through a first bidirectional link and a second unidirectional link mechanism; introducing signal encryption and authentication mechanisms to ensure communication security; constructing a network maintenance strategy to support automatic access of new nodes, switching when a central node fails, and periodic network structure updates; constructing a graph model based on node connection relationships and signal strength, and introducing artificial intelligence graph algorithms and reinforcement learning strategies to achieve optimal routing planning under multi-hop paths. The method provided in this application can improve the security, stability, and routing intelligence of ad hoc networks, achieving efficient and reliable wireless communication network planning.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of wireless communication, in particular to a self-organizing network intelligent routing planning method and system. BACKGROUND

[0002] With the development of wireless network technology, self-organizing network has become an important network organization method, and has the characteristics of automatic configuration, dynamic networking and strong node mobility, and can quickly establish a communication network in an environment without fixed infrastructure, and facilitates short-distance wireless communication.

[0003] At present, the existing self-organizing network technology still has some problems and challenges. In the process of self-organizing network construction, how to efficiently elect a central node and establish a link relationship between nodes is a key factor affecting network performance. The communication security problem between nodes in the self-organizing network cannot be ignored, and a proper security mechanism needs to be introduced to protect the network from malicious attacks. The self-organizing network also needs to have the ability of network maintenance, which can handle the access of new nodes, the exit of nodes, and the dynamic update of network structure.

[0004] After the self-organizing network is established, how to select the best routing path in the network to transmit data is crucial to improve the transmission efficiency and reliability of the network. Traditional routing algorithms are usually based on pre-designed rules or simple measurement indicators, which are difficult to adapt to the dynamic characteristics and complex network environment of the self-organizing network.

[0005] In view of this, the application provides a self-organizing network intelligent routing planning method and system. SUMMARY

[0006] To achieve the above purpose, the application provides a self-organizing network intelligent routing planning method and system, and the specific technical solutions are as follows:

[0007] The self-organizing network intelligent routing planning method comprises:

[0008] The self-organizing network strategy of creating twice link routing nodes comprises constructing a first bidirectional link mechanism and a second unidirectional link mechanism, and performing central node election and self-organizing network between routing nodes;

[0009] In the self-organizing network process of the routing nodes, a signal encryption and identity verification mechanism is introduced for the wireless signal transmitted by the routing nodes;

[0010] A network maintenance strategy is constructed, including an automatic access mechanism of new nodes, an exit judgment processing mechanism of existing nodes, and a periodic network structure update mechanism;

[0011] According to the node connection relationship and signal strength obtained by the link mechanism, a graph model of the routing nodes is constructed, and an artificial intelligence-based graph algorithm is used to calculate the best transmission path of the signal between the routing nodes.

[0012] Preferably, the first bidirectional linking mechanism comprises: a routing node sending a first linking signal; after a neighboring routing node receives the signal, recording the sending node identity and the received signal strength to form a neighbor information set;

[0013] Each routing node sends the neighbor information set to its neighboring nodes as a first linking response signal; after a node receives the first linking response signal, it updates the neighbor information set and records the strength information of the signal received by the neighboring node.

[0014] Preferably, the second unidirectional linking mechanism comprises: each routing node transmitting the signal strength received and fed back by the neighboring nodes as second linking information;

[0015] The receiving node calculates a connection degree score according to the neighbor information set, and the connection degree score is equal to the sum of the received signal strength and the sum of the received signal strength;

[0016] According to the connection degree score, the node with the highest score is elected as the center node, and the node with the second highest score is elected as the backup center node; other nodes update the local network parameters according to the center node configuration information and start relay forwarding to form a multi-hop relay network topology.

[0017] Preferably, each routing node generates a random session key, encrypts the key using a private key, and sends a linking signal;

[0018] The linking signal is encrypted using the session key, and is sent to the neighboring nodes together with the encrypted session key; the neighboring nodes decrypt the session key through the public key of the sending node, and then decrypt the linking signal using the session key to complete the encrypted communication.

[0019] Preferably, when a new routing node requests to join an existing network, it submits an identity request containing a digital certificate to the center node;

[0020] After the center node verifies the validity of the identity certificate, it generates a random challenge value and sends it encrypted; the node decrypts the challenge value and signs it using a private key, and returns the signature and the decrypted value;

[0021] The center node verifies whether the signature matches, and if it does, it allows the new node to join the network.

[0022] Preferably, the new node finds the center node through the already constructed network, and sends a join request containing a digital certificate to the center node;

[0023] The center node performs identity verification on the digital certificate of the new node, adds the new node information to the list of trusted nodes after verification, sends a confirmation message to the new node, and broadcasts a network topology update message to all existing nodes;

[0024] The existing routing node receives the update message, updates the trusted node list and the routing table;

[0025] When the existing routing node exits, if the exiting routing node is not the center node, the network remains unchanged; if it is the center node, the standby center node is enabled as the new center node, and the new standby center node is re-elected based on the original standby node.

[0026] Preferably, all routing nodes periodically re-execute the ad hoc network strategy, calculate the new connection degree score of each routing node according to the updated connection relationship and signal strength;

[0027] According to the new connection degree score, the center node and the standby center node are re-elected.

[0028] Preferably, based on the node connection relationship and signal strength obtained in the ad hoc network process, a routing node undirected weighted graph model is constructed;

[0029] The graph neural network is used for feature learning of the graph model, and the routing node hidden feature representation is learned by iteratively aggregating the routing node's own features and neighbor node features;

[0030] Based on the learned node features, the attention mechanism is used to calculate the attention weight between nodes, representing the importance and relevance between nodes.

[0031] Preferably, a routing strategy based on reinforcement learning is designed, and a state space, an action space and a reward function are defined;

[0032] The Q-learning algorithm is used to learn the optimal routing strategy, and the optimal next hop node is selected by interacting with the environment and updating the Q value table to achieve the best transmission path of signals between routing nodes.

[0033] The ad hoc network intelligent routing planning system is used to implement the ad hoc network intelligent routing planning method, and includes an ad hoc network module, an encryption verification module, a network maintenance module and a path planning module.

[0034] The ad hoc network module creates a routing node ad hoc network strategy with two links, including constructing a first bidirectional link mechanism and a second unidirectional link mechanism, and electing a center node and an ad hoc network between routing nodes.

[0035] The encryption verification module introduces a signal encryption and identity verification mechanism for the wireless signals transmitted by the routing nodes during the ad hoc network process of the routing nodes.

[0036] The network maintenance module is used to construct a network maintenance strategy, including an automatic access mechanism for new nodes, an exit judgment processing mechanism for existing nodes, and a periodic network structure update mechanism.

[0037] The path planning module constructs a graph model of the routing nodes according to the node connection relationship and signal strength obtained by the linking mechanism, and calculates the best transmission path of the signal between the routing nodes by using a graph algorithm based on artificial intelligence.

[0038] The application has the beneficial effects that: by constructing bidirectional and unidirectional linking mechanisms and combining with center node election, the application realizes efficient ad hoc networking between routing nodes, improves the connection flexibility and robustness of the network, and adapts to the rapid deployment requirements of different network topologies.

[0039] The application introduces encryption and identity verification mechanisms in the process of wireless communication, improves the data security and node credibility in the communication process of the routing nodes, prevents illegal access and signal interception, and ensures the overall security of the network.

[0040] The application realizes dynamic self-adaptation and stable operation of the network structure by constructing a perfect network maintenance strategy, improves the response capability of the network to topology changes, and enhances the expansibility and continuous service capability of the network.

[0041] The application constructs a graph model based on connection relationship and signal strength, and uses an artificial intelligence graph algorithm for path calculation, which improves the efficiency and stability of signal transmission, realizes intelligent optimization of the routing path, and reduces communication delay and energy consumption. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 The application is an ad hoc network intelligent routing planning method flowchart;

[0043] Figure 2 The application provides a routing node ad hoc networking strategy flowchart based on twice linking;

[0044] Figure 3 The application provides an encryption and identity verification mechanism flowchart;

[0045] Figure 4 The application provides a dynamic network maintenance mechanism diagram;

[0046] Figure 5 The application provides an intelligent routing optimization mechanism diagram based on a graph neural network;

[0047] Figure 6 The application is an ad hoc network intelligent routing planning system structure diagram. DETAILED DESCRIPTION

[0048] In order to make the above-mentioned purposes, features and advantages of the application more obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the accompanying drawings.

[0049] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention can also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0050] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0051] Example 1

[0052] Reference Figure 1 This is the first embodiment of the present invention, which provides an intelligent routing planning method for self-organizing networks.

[0053] Step 1: Construct a self-organizing network strategy with two links. Through a first bidirectional link mechanism and a second unidirectional link mechanism, central node election and self-organizing network are performed among the routing nodes. See [link / details]. Figure 2 This is a flowchart of the self-organizing network strategy for routing nodes based on two links in this step.

[0054] Each routing node periodically transmits a first link signal, which contains the routing node's unique identifier ID. i After receiving the first link signal from neighboring router node j, router node i records the neighbor router node's identifier ID. j With the received signal strength S ij This forms the set N of neighboring routing node information for routing node i. i :N i ={(ID) j ,S ij )|j∈neighborr(i)}; where neighbor(i) represents the set of neighboring routing nodes of routing node i; routing node i sets its neighbor routing node information set N i It is transmitted to all neighboring routing nodes, forming the first link response signal.

[0055] Through the first link mechanism, each routing node can obtain the identifier and signal strength of its neighboring routing nodes, providing necessary information for subsequent central node election and routing planning.

[0056] After receiving the first link response signal from neighboring router node j, router node i calculates the signal strength S of the signal received by neighboring router node j from this router node. ji The set of neighbor routing node information for updated routing node i is: N i={(ID) j ,S j ,S ji )|j∈neighbor(i)}.

[0057] The first link response signal enables nodes to know the strength of the signals received by neighboring nodes, thereby expanding the range of information perception; this mechanism helps nodes to comprehensively evaluate the network topology and optimize ad hoc network decisions.

[0058] Router node i transmits the signal strength information {S} received by its neighboring router nodes. ji |j∈neighbor(i)} is transmitted as the second link information to all neighboring routing nodes; after receiving the second link information of neighboring routing node j, routing node i calculates the connectivity score C of routing node j. j :

[0059] C j =∑ k∈neighbor(j) S jk +∑ k∈neighbor(j) S kj ;

[0060] Where, ∑ k∈neighbor(j) S jk ∑ represents the sum of the signal strengths of all neighboring routing nodes received by routing node j. k∈neighbor(j) S kj This represents the sum of the signal strengths of all neighboring routing nodes of routing node j received from routing node j.

[0061] After the second one-way link mechanism is implemented, each routing node can obtain the communication information between neighboring nodes and then calculate the connectivity score of each node. Taking into account the signal connection strength between nodes, it provides an important basis for the election of the central node. Nodes with high connectivity scores are usually located in the center of the network and have better signal connections with other nodes, making them more suitable as the central node.

[0062] Based on connectivity scores, routing nodes select the highest-scoring node as the central node and the second-highest-scoring node as a candidate central node. After selecting a central node, all routing nodes automatically switch to the same network identification parameters (such as network name, channel number, and authentication method) and enable relay forwarding. Each node connects to the central node directly or indirectly via hop-by-hop relay links from other nodes, constructing multi-hop relay links to achieve regional extension of network signal and enhanced communication coverage. In this topology, the central node acts as the primary signal source, and other nodes, simulating the central node's network configuration, wirelessly extend the central node's signal and perform distributed routing relay, ensuring the entire network forms a self-organizing wireless mesh structure with unified identification and access policies.

[0063] The central node is responsible for maintaining the network topology, coordinating data transmission and routing decisions; the backup central node maintains low activity to monitor the network status and periodically synchronizes information with the central node; when the central node fails, the backup central node can quickly take over the responsibilities of the central node to ensure the stable operation of the network.

[0064] This step, through the election mechanism of central node and alternative nodes, selects the most important and most connected node in the network as the central node, which helps improve network management efficiency and data transmission performance. At the same time, the existence of alternative central nodes improves the robustness and reliability of the network and avoids network paralysis caused by the failure of the central node.

[0065] The central node transmits network configuration information, including network name, channel parameters, authentication configuration, etc., to all nodes. The central node first sends the configuration information to its directly connected neighbor nodes, and the neighbor nodes forward it to their neighbor nodes after receiving it, until all nodes have received the configuration information.

[0066] After receiving the network configuration information from the central node, all nodes perform local settings according to the configuration information, join the network, and form a unified network topology. The nodes initialize routing paths according to the routing table, establish connections with other nodes, and prepare for subsequent data transmission.

[0067] By building a unified network, all nodes join the network according to a unified configuration, forming a standardized and orderly network topology, which improves the network's organizational efficiency. Unified IP address allocation and routing table settings simplify the communication process between nodes and reduce data transmission overhead.

[0068] This step, through a two-link self-organizing network strategy, achieves self-organizing connections between routing nodes and initializes routing paths, completing the network construction between routing nodes. This self-organizing network approach fully utilizes local and neighbor information of nodes, efficiently and reliably constructing the network topology through distributed linking and election mechanisms, laying the foundation for subsequent route optimization and data transmission. Simultaneously, the introduction of a central node and alternative central nodes improves network management efficiency and robustness, ensuring stable network operation.

[0069] Step 2: During the self-organizing network process of the routing nodes, a signal encryption and authentication mechanism is introduced for the wireless signals emitted by the routing nodes. See [link / reference] Figure 3 This is a flowchart of the signal encryption and authentication mechanism for this step.

[0070] When any node i sends a link signal, a random session key SK is first generated. ij The private key SK is used for encryption of communication with neighboring node j; then the session key SK is encrypted using node i's private key. ij Encryption is performed to obtain the encrypted session key. Node i then uses session key SK ij For link signal M i Encryption is performed to obtain the encrypted link signal.

[0071] Node i will use the encrypted session key and encrypted link signal Together, they are sent to neighbor node j; after receiving the encrypted link signal, neighbor node j uses node i's public key PK. i Encrypted session key Decrypt to obtain the session key SK ij Neighbor node j uses session key SK ij Encrypted link signals Decryption is performed to obtain the original link signal.

[0072] Encrypting link signals using session keys prevents unauthorized nodes from eavesdropping on and tampering with them, ensuring the confidentiality and integrity of the link signals. Furthermore, since each node uses different session keys to communicate with different neighboring nodes, even if one session key is cracked, it will not affect the security of other sessions, thus improving the overall security of the network.

[0073] In the new routing node ID e When joining an established network, a digital certificate-based authentication mechanism is introduced to verify the legitimacy of routing nodes; the specific steps are as follows:

[0074] Step a: Issue digital certificates to verified routing nodes based on a trusted Certificate Authority (CA); the digital certificate contains the node's identity information ID. e Public key PK e And CA's digital signature Sig CA .

[0075] Step b: Node e sends an authentication request to the central node M, the request containing node e's digital certificate Cert. e .

[0076] Step c: After receiving the authentication request, the central node M first verifies the digital certificate Cert. i The validity of the digital certificate is verified by the following steps: verifying the format and integrity of the digital certificate; verifying whether the digital certificate is valid; and verifying the CA signature Sig on the digital certificate using the CA's public key. CA Is it effective?

[0077] Step d: If the digital certificate Certe is verified, the central node M extracts the identity information ID of node e from the digital certificate. e PK with public key e .

[0078] Step e: The central node M generates a random challenge value R and uses node e's public key PK. e The random challenge value is encrypted to obtain the encrypted challenge value E. e .

[0079] Step f: The central node M transmits the encrypted challenge value E e Send to node e.

[0080] Step g: Node e receives the encrypted challenge value E e Then, use your private key SK e Decrypt it to obtain the random challenge value R.

[0081] Step h: Node e uses its own private key SK e Sign the random challenge value R to obtain the signature value Sig. e ;

[0082] Step i: Node e will decrypt the random challenge value R and the signature value Sig. e Send to the central node M;

[0083] Step j: The central node M receives the random challenge value R and signature value Sig returned by node i. e Then, use node e's public key PK. e For the signature value Sig ePerform verification; if verification passes, authentication is successful and node i is allowed to join the network; otherwise, authentication fails and node e is refused to join the established network.

[0084] By introducing digital certificates and challenge-response mechanisms, stronger identity verification capabilities can be provided. Digital certificates are issued by trusted third-party CAs, which can ensure the authenticity and legitimacy of node identities. At the same time, the introduction of random challenge values ​​and digital signatures can prevent man-in-the-middle attacks and identity forgery attacks, further enhancing the security of identity verification.

[0085] The method provided in this step, compared to simple authentication, introduces certain computational and communication overhead through a digital certificate-based authentication mechanism, but it can significantly improve network security, especially in the face of complex network attacks, providing a higher level of defense. By closely integrating with signal encryption mechanisms, it can comprehensively ensure communication security and node trustworthiness during the self-organizing network process, providing a secure and reliable network environment for subsequent route discovery and data transmission.

[0086] Step 3: Construct network maintenance strategies, including automatic access mechanisms for new nodes, exit handling mechanisms for existing nodes, and periodic network structure update mechanisms. See [link / reference needed]. Figure 4 This is a diagram of the dynamic network maintenance mechanism for this step.

[0087] The mechanism for adding a new routing node involves performing the following steps to automatically join the network when a new routing node n enters the network:

[0088] New node o searches for central node M through the established network and sends a join request to central node M, which includes the digital certificate Cert of new node o. o ;

[0089] The central node M verifies the digital certificate Cert of the new node o based on the established authentication mechanism. o Perform identity verification;

[0090] If the new node o passes authentication, the central node M will send the new node o's identity information ID. o PK with public key o Add it to the list of trusted nodes and send an acknowledgment message to the new node o to indicate that the new node o has successfully joined the network;

[0091] Central node M broadcasts a network topology update message to all existing nodes, notifying them of the addition of new node o; the update message includes the identity information (ID) of the new node n. o PK with public key o .

[0092] After an existing node receives the network topology update message, it will send the identity information (ID) of the new node o. o PK with public key o Add it to your list of trusted nodes and update the routing table.

[0093] By authenticating new nodes, it is ensured that only legitimate nodes can join the network, thus improving network security and trustworthiness. Through network topology update messages, all nodes are promptly informed of the addition of new nodes, updating their own node lists and routing tables to maintain network topology consistency. Existing nodes update their trusted node lists and routing tables, ensuring that network topology information remains consistent across all nodes, facilitating subsequent communication and routing.

[0094] Establish an exit handling mechanism for existing nodes. When an existing routing node i leaves the network, perform the following steps to handle the node exit:

[0095] If the exiting node is not the central node, the already constructed network remains unchanged;

[0096] If the exiting node is the central node, a backup central node is activated, and the network of the backup central node becomes the new central network. Nodes with lower connectivity than the original backup central node are elected as the new backup central node. The specific steps are as follows:

[0097] The backup central node takes over the responsibilities of the central node, becomes the new central node, and creates a network signal. After the backup central node is activated, it immediately broadcasts a "central node change notification" message to all known routing nodes in the current network, announcing that the backup central node's identity has been switched to the new central node. After receiving the change message, each routing node actively pushes its current network signal identifier to its neighboring nodes. Based on the received new central node identifier information, the routing node copies and switches to the same network configuration parameters as the new central node (including network name, channel parameters, authentication configuration, etc.), updates its local trusted node list and routing table, resets the new central node as the network master node, completes its own access to the new network topology, and elects nodes with lower connectivity than the original backup central node as the new backup central node.

[0098] When the central node disconnects from existing network communication, the backup central node can quickly restore normal network operation and reduce the impact of central node failure on the network. Simultaneously, the backup central node sends central node change messages, ensuring all nodes update network topology information promptly and maintaining network consistency. The exit of non-central nodes does not affect the overall network structure, and the network remains stable.

[0099] Create a periodic network structure update mechanism. At fixed intervals, regardless of whether any nodes enter or leave, perform the following steps to update the network structure:

[0100] All nodes re-execute the self-organizing network policy, including the first and second connection processes;

[0101] Based on the updated node connectivity and signal strength, a new connectivity score for each node is recalculated.

[0102] Based on the new connectivity score, a new central node and a candidate central node are re-elected; the newly elected central node broadcasts a network topology update message to all nodes, conveying the identities of the new central node and the candidate central node to all nodes;

[0103] After receiving the network topology update message, all nodes update their trusted node list and routing table to adapt to the new network structure.

[0104] By re-executing the self-organizing network strategy, updating the connection relationships and signal strength information between nodes, the network environment can be adapted to changes. Based on the latest network information, the importance of nodes can be reassessed, providing a basis for optimizing the network structure.

[0105] This step utilizes a periodic network structure update mechanism to dynamically optimize the network topology, adapting to factors such as node movement and changes in channel conditions. It also ensures timely updates to the central and backup central nodes, guaranteeing network robustness and performance. Regular network structure updates also help to promptly detect and handle node anomalies, improving network fault tolerance and reliability. Through the effective combination of these mechanisms, dynamic management and maintenance of ad hoc networks can be achieved, ensuring long-term stable operation and efficient service.

[0106] Step 4: Based on the node connectivity and signal strength obtained from the link mechanism, construct a graph model of the routing nodes, and use an AI-based graph algorithm to calculate the optimal transmission path for the signal between the routing nodes. (See [link / reference]) Figure 5 This is a diagram of the intelligent routing optimization mechanism based on graph neural networks in this step.

[0107] Based on the node connection relationships and signal strength information obtained during the self-organizing network process, an undirected weighted graph model G = (V, E, W) of routing nodes is constructed, where V represents the set of routing nodes, E represents the set of connection relationships between nodes, and W represents the set of signal strength weights for the connections; for any two routing nodes v i and v j If node v i and v j If there is a connection between two points, add an edge e to the graph. ij The weight w of the edgeij This represents the signal strength between two routing nodes.

[0108] By abstracting network topology into a graph model, complex network relationships are transformed into mathematical models, facilitating analysis and optimization using graph theory algorithms. Weighted undirected graph models accurately represent the connectivity and signal quality between nodes, providing crucial input information for subsequent route optimization.

[0109] Graph Neural Networks (GNNs) are used to learn and represent the graph model of routing nodes, learning the hidden feature representation of each node. GNNs iteratively aggregate the node's own features and the features of its neighbors to update the node's hidden representation, capturing the interaction relationships between nodes and network structure information. The aggregation and update functions of the GNN are designed as follows:

[0110] Aggregate functions:

[0111] Update function:

[0112] in, Let w represent the aggregation information of node v in the k-th round of aggregation, N(v) represent the set of neighboring nodes of node v, and w represent the aggregation information of node v in the k-th round of aggregation. uv d represents the edge weight between nodes u and v. u and d v Let u and v represent the degrees of nodes u and v, respectively. This represents the hidden feature representation of node u after the (k-1)th round update.

[0113] W represents the new hidden feature representation of node v after the k-th round update. (k) and b (k) Let represent the weight matrix and bias vector updated in the k-th round, respectively, and σ represent the activation function (such as the ReLU function).

[0114] Through multiple rounds of iterative aggregation and updates, GNNs can learn high-order feature representations of nodes, capturing complex interaction relationships and network structure information between nodes; the learned node feature representations serve as inputs for subsequent routing path decisions, improving the effectiveness of routing optimization.

[0115] Based on the learned node feature representations, an attention mechanism is used to calculate attention weights between nodes, representing the importance and relevance between nodes. For routing node v i and v j Node v i and v j The attention weights between them are calculated as follows:

[0116]

[0117] Among them, h i and h j They represent nodes v respectively i and v j Feature representation, W q and W k These represent the query weight matrix and the key weight matrix, respectively. 'a' represents the attention calculation function (such as a dot product function or a parameterized function), and 'e' represents the key weight matrix. ij Represents node v i For node v j Attention score, α ij This represents the normalized attention weights.

[0118] The attention mechanism can be used to calculate the correlation and importance between nodes based on their feature representations, thus obtaining the attention weights between nodes. These attention weights can serve as an important basis for routing decisions, helping to select the optimal next-hop node and improving routing efficiency and quality.

[0119] Design a routing strategy based on reinforcement learning. By learning the optimal routing decision, achieve the best transmission path for signals between routing nodes. Define the state space, action space, and reward function of the reinforcement learning, including: State space S = {s1, s2, ..., s...} n}, where s i Indicates the node v where the current signal is located. i The state includes information such as the node's feature representation and the distance from the current node to the target node.

[0120] Action space A = {a1, a2, ..., a} m}, where a j This indicates that at the current node v i Select the next hop node v j As a routing decision.

[0121] Reward function: Wherein d(v i ,v t ) represents node v i to target node v t The distance, r p Indicates a positive reward, r n This indicates a negative reward; if the selected next hop node v k Compared to the current node v i Closer to the target node v t If the reward is positive, a positive reward will be given; otherwise, a negative reward will be given.

[0122] The optimal routing strategy is learned using the Q-learning algorithm, and the Q-value table is updated accordingly.

[0123] Q(s i ,a j )←Q(s i ,a j )+α·[R(s i ,a j ,s k )+γ·max a Q(s k ,a)-Q(s i ,a j )];

[0124] Where Q(s) i a j ) indicates that in state s i Choose action a j The Q value is given by α, where α represents the learning rate and γ represents the discount factor.

[0125] By continuously interacting with the environment and updating the Q-value table, reinforcement learning algorithms can learn optimal routing strategies, select the best next-hop node, and achieve the best transmission path for signals between routing nodes. The learned routing strategies can adapt to dynamic changes in network topology, exhibiting adaptability and robustness.

[0126] During data transmission, the network status and signal quality between routing nodes are continuously monitored, and the routing strategy is dynamically adjusted according to changes in the network status to achieve adaptive routing optimization.

[0127] Based on changes in network state indicators, the reward function and state representation in the reinforcement learning algorithm are dynamically adjusted so that the routing strategy can adapt to changes in network state. For example, when the signal quality between routing nodes deteriorates, the weight of negative rewards is increased to encourage the selection of paths with better signals; when the network is congested, link bandwidth utilization information is added to the state representation to guide the routing strategy to avoid congested links.

[0128] By continuously monitoring network status and dynamically adjusting routing strategies, adaptive routing optimization can be achieved, enabling routing decisions to respond promptly to changes in network status and ensuring the quality and efficiency of data transmission.

[0129] The routing node dynamically selects the next-hop node based on the learned optimal routing strategy and forwards the data packet to the next-hop node until the data packet reaches the target node.

[0130] During data transmission, each routing node selects the next-hop node with the largest Q value as the forwarding decision based on its current state and the learned Q-value table. The routing node sends the data packet to the selected next-hop node and updates the routing information of the data packet, recording the nodes that have been passed and the next-hop node. After receiving the data packet, the next-hop node repeats the above process until the data packet reaches the target node.

[0131] By dynamically selecting the optimal next-hop node, data packets can be transmitted along the optimal path between routing nodes, avoiding inefficient or congested links and improving data transmission efficiency and reliability. Since routing decisions are based on learned strategies, they can adapt to changes in network topology and state, exhibiting good adaptability and robustness.

[0132] This step leverages artificial intelligence technology, using graph neural networks to learn the network topology, employing attention mechanisms to calculate the correlations between nodes, and using reinforcement learning algorithms to learn the optimal routing strategy. This method fully utilizes the network's structural and state information to dynamically adjust routing decisions, achieving intelligent, efficient, and reliable data transmission. Compared to traditional fixed routing protocols, this method exhibits stronger adaptability and optimization capabilities, enabling it to cope with complex and ever-changing network environments and improve overall network performance.

[0133] Example 2

[0134] Reference Figure 6 This is the second embodiment of the present invention, which provides a self-organizing network intelligent routing planning method system.

[0135] The system includes: a self-organizing network module, an encryption verification module, a network maintenance module, and a path planning module.

[0136] The self-organizing network module establishes a self-organizing network strategy for routing nodes through two links, including constructing a first bidirectional link mechanism and a second unidirectional link mechanism, to elect a central node and form a self-organizing network among the routing nodes.

[0137] The encryption and verification module introduces a signal encryption and authentication mechanism for the wireless signals emitted by the routing nodes during the self-organizing network process of the routing nodes.

[0138] The network maintenance module is used to construct network maintenance strategies, including an automatic access mechanism for new nodes, a mechanism for judging and handling the exit of existing nodes, and a periodic network structure update mechanism.

[0139] The path planning module constructs a graph model of routing nodes based on the node connection relationships and signal strength obtained from the link mechanism, and uses an artificial intelligence-based graph algorithm to calculate the optimal transmission path of the signal between the routing nodes.

[0140] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings or direct couplings or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0141] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the claims. All of these variations are within the protection scope of the present invention.

Claims

1. A method for intelligent routing planning in an ad hoc network, characterized in that, The application relates to a twice-linking routing node self-organizing network strategy, which comprises a first bidirectional linking mechanism and a second unidirectional linking mechanism, center node election and self-organizing network among routing nodes. The first bidirectional linking mechanism comprises that a routing node sends a first linking signal; after a neighbor routing node receives the signal, the neighbor routing node records the sending node identifier and the received signal strength to form a neighbor information set; Each routing node sends the neighbor information set to the neighbor node as a first linking response signal; after a node receives the first linking response signal, the node updates the neighbor information set and records the strength information of the signal received by the neighbor node; The second unidirectional linking mechanism comprises that each routing node transmits the signal strength received and fed back by the neighbor node as second linking information; A receiving node calculates a connection degree score according to the neighbor information set, and the connection degree score is equal to the sum of the received signal strength and the sum of the received signal strength; According to the connection degree score, a node with the highest score is elected as a center node, a node with the second highest score is elected as a backup center node, and other nodes update local network parameters according to the center node configuration information and start relay forwarding to form a multi-hop relay network topology; In the self-organizing network process of the routing node, a signal encryption and identity authentication mechanism is introduced to the wireless signal transmitted by the routing node; A network maintenance strategy is constructed, which comprises an automatic access mechanism of a new node, an exit judgment processing mechanism of an existing node and a periodic network structure updating mechanism; According to the node connection relationship and signal strength obtained by the linking mechanism, a graph model of the routing node is constructed, and an artificial intelligence-based graph algorithm is used to calculate the best transmission path of the signal among the routing nodes. Each routing node generates a random session key, encrypts the key by using a private key and sends a linking signal; the linking signal is encrypted by using the session key and is sent to a neighbor node together with the encrypted session key; the neighbor node decrypts the session key by using a sending node public key and then decrypts the linking signal by using the session key, so that encrypted communication is completed.

2. The method of claim 1, wherein, When a new routing node requests to join an existing network, the new routing node submits an identity request containing a digital certificate to a center node; 3. The method of claim 2, wherein, After the center node verifies the validity of the identity certificate, the center node generates a random challenge value and sends the challenge value after encryption; the node decrypts the challenge value and signs the challenge value by using a private key, and returns the signature and the decrypted value; the center node verifies whether the signature is matched, and if the verification is passed, the new node is allowed to join the network. The new node finds the center node through the constructed network, sends a joining request containing a digital certificate to the center node; the center node verifies the identity of the new node through the digital certificate, adds the new node information to a trusted node list after the verification, sends a confirmation message to the new node, and broadcasts a network topology updating message to all existing nodes; 4. The method of claim 3, wherein, After receiving the updating message, the existing routing node updates the trusted node list and the routing table; when an existing routing node exits, if the exiting routing node is not the center node, the network remains unchanged; if the exiting routing node is the center node, a backup center node is enabled as a new center node, and a new backup center node is re-elected on the basis of the original backup node. ​ 5. The method of claim 4, wherein, All routing nodes periodically re-perform the ad hoc network strategy, calculate the new connection degree score of each routing node according to the updated connection relationship and signal strength, and re-elect the center node and the backup center node according to the new connection degree score.

6. The method of claim 4, wherein, According to the node connection relationship and signal strength obtained by the ad hoc network process, a routing node undirected weighted graph model is constructed; Using graph neural network to learn the features of the graph model, the routing node's own features and neighbor node features are iteratively aggregated to learn the routing node's hidden feature representation; Based on the learned node features, use attention mechanism to calculate the attention weight between nodes to represent the importance and relevance between nodes.

7. The method of claim 6, wherein, Design a routing strategy based on reinforcement learning, define state space, action space and reward function; Using Q-learning algorithm to learn the optimal routing strategy, through interaction with the environment and updating Q value table, the optimal next hop node is selected to realize the best transmission path of signal between routing nodes.

8. A self-organizing network intelligent routing planning system for implementing the self-organizing network intelligent routing planning method of any one of claims 1 to 7, characterized in that, It includes: Ad hoc network module, encryption verification module, network maintenance module and path planning module; The ad hoc network module creates a twice-linked routing node ad hoc network strategy, including building a first bidirectional link mechanism and a second unidirectional link mechanism, and electing a center node and an ad hoc network between routing nodes; The encryption verification module introduces a signal encryption and identity verification mechanism for the wireless signal transmitted by the routing node during the ad hoc network process of the routing node; The network maintenance module is used to construct a network maintenance strategy, including an automatic access mechanism for new nodes, an exit judgment processing mechanism for existing nodes and a periodic network structure updating mechanism; The path planning module constructs a graph model of routing nodes according to the node connection relationship and signal strength obtained by the link mechanism, and uses an artificial intelligence-based graph algorithm to calculate the best transmission path of signal between routing nodes.

Images