Distributed network clustering method and system for heterogeneous mobile ad hoc network

Abstract

The invention relates to a heterogeneous mobile ad hoc network distributed network clustering method and system, and the method comprises the steps: responding to the starting of an enhanced node or a common node, controlling the enhanced node or the common node to broadcast the attribute information of the enhanced node or the common node, and receiving the attribute information of a neighbor node to obtain time delay distance data and priority data; based on the time delay distance data and the priority data of the enhanced nodes or the common nodes, priority core nodes are determined, and an initial cluster structure is constructed in a distributed mode through cluster affiliation relation establishment of the enhanced nodes or the common nodes and density correlation information exchange between the priority core nodes; screening enhanced nodes from the priority core nodes as cluster head candidates, and electing a cluster head from the cluster head candidates; and based on the clustering mechanism of the enhanced node and the common node, controlling the enhanced node or the common node to join in a cluster and distribute a network address through a cluster head. According to the method, intra-cluster autonomy and inter-cluster interconnection efficiency is improved, and ordered and efficient clustering control and stable structure in a dynamic environment are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile ad hoc networks, and in particular to a distributed network clustering method and system for heterogeneous mobile ad hoc networks. Background Art

[0002] Mobile ad hoc networks are networks without fixed infrastructure, consisting of mobile nodes that support wireless communication. Nodes perform both data transmission and routing functions and can dynamically adjust their topology. They are suitable for scenarios such as disaster relief, military communications, drone swarms, and vehicular communications. Clustering algorithms improve network energy efficiency, routing efficiency, and scalability by dividing nodes into clusters and assigning cluster heads to manage communications. However, existing clustering methods have the following shortcomings:

[0003] Chinese patent CN118265104A discloses a cluster routing method with an optimized forwarding mechanism. This method uses non-cluster head nodes within a cluster to forward data to the cluster head and uses multi-hop forwarding rules between clusters to simplify routing and extend network life. However, the cluster head node is overloaded and prone to failure due to rapid energy consumption, limiting cluster expansion. Single-path forwarding within the cluster lacks flexibility, limiting communication performance. In dynamic environments, the method lacks a mechanism to adapt to topology changes and node failures, impacting network stability.

[0004] Another Chinese patent, CN116321350A, discloses a clustering method for in-vehicle self-organizing networks based on shared nearest neighbors. This method divides clusters by node similarity, reducing the number of clusters. However, its cluster head selection fails to consider node stability, energy, and processing power, which can easily lead to the selection of highly mobile or low-capability nodes as cluster heads, resulting in frequent cluster switching and performance degradation. The lack of energy management and load balancing mechanisms shortens cluster lifespan. Furthermore, the SNN mechanism has limited applicability and insufficient scalability in heterogeneous or highly dynamic scenarios. Summary of the Invention

[0005] The object of the present invention is to provide a distributed network clustering method and system for heterogeneous mobile ad hoc networks to solve the existing problem of clustering enhanced nodes and ordinary nodes in a mixed scenario in heterogeneous mobile ad hoc networks.

[0006] To achieve one of the above-mentioned objectives, an embodiment of the present invention provides a distributed network clustering method for heterogeneous mobile ad hoc networks, the method comprising:

[0007] In response to the enhanced node or the ordinary node being started, controlling the enhanced node or the ordinary node to broadcast its attribute information, and receiving attribute information of neighboring nodes to obtain delay distance data and priority data;

[0008] Based on the delay distance data and priority data of enhanced nodes or ordinary nodes, priority core nodes are determined, and the initial cluster structure is constructed in a distributed manner by establishing cluster affiliation relationships of enhanced nodes or ordinary nodes and exchanging density association information between priority core nodes;

[0009] The enhanced nodes are selected from the priority core nodes as cluster head candidates, and the cluster head is elected from the cluster head candidates; based on the clustering mechanism of the enhanced nodes and ordinary nodes, the cluster head controls the enhanced nodes or ordinary nodes to join the cluster and allocates network addresses.

[0010] As a further improvement of an embodiment of the present invention, the method further includes: the calculation formula of the priority data is:

[0011]

[0012] in, To enhance the node collection, is the indicator function, if For enhanced nodes Take 1, otherwise Take 0; is the average damage probability of the node per unit time; is the current remaining energy of the node, is the maximum energy capacity of the node; is the energy attenuation coefficient; is the node mobility index, is the maximum mobility index; is the weight coefficient, and ;

[0013] The average probability of damage The calculation formula is:

[0014]

[0015] Where N is the number of threat source types, The mth type of threat to the node The failure probability per unit time represents the reliability of the node under external threats;

[0016] The energy attenuation coefficient The calculation formula is:

[0017]

[0018] in, For nodes The energy loss rate, are the coefficients of the preset linear function and exponential function, It is a preset energy loss threshold used to measure the nonlinear characteristics of node energy consumption;

[0019] The node mobility index The calculation formula is:

[0020]

[0021] in, For nodes The velocity at time t, is the maximum speed, is the velocity variance, is the maximum velocity variance, is the weight coefficient used to evaluate the stability and dynamics of node movement, and .

[0022] As a further improvement of an embodiment of the present invention, the method further includes: the delay distance includes an uplink delay distance and downlink delay distance ;

[0023] The downlink delay distance The calculation formula is:

[0024]

[0025] in, is the propagation delay, To handle delay, is the downlink transmission delay, expressed as:

[0026]

[0027] in, is the downlink traffic, is the actual downlink throughput, expressed as:

[0028]

[0029] in, is the node throughput empirical coefficient, is the downlink channel communication capacity, expressed as:

[0030]

[0031] in, The downlink channel bandwidth is determined by the spectrum division of the communication system. value, is the signal power, N is the thermal noise power, expressed as:

[0032]

[0033] in, is the Boltzmann constant, T is the ambient temperature, B is the system bandwidth, and F is the noise factor of the signal receiver at this node, which is used to describe the degree of degradation of the system noise;

[0034] The uplink delay distance The calculation formula is:

[0035]

[0036] in, is the uplink transmission delay, expressed as:

[0037]

[0038] in, is the uplink traffic, is the actual uplink throughput, expressed as:

[0039]

[0040] in, is the uplink channel capacity, which is evaluated by the channel fading model and is approximately expressed as:

[0041]

[0042] in, is the uplink channel bandwidth, is the transmit power, are the antenna gains of the signal transmitter and receiver respectively, c is the speed of light, f is the communication frequency, and x is the Euclidean distance between nodes.

[0043] As a further improvement of an embodiment of the present invention, the method further includes: the step of constructing the initial cluster network in a distributed manner includes:

[0044] Based on the delay distance, determine the node of Neighborhood Set , and according to the Number of neighborhood set nodes Not less than the preset minimum node density threshold MinPDs and with a priority lower than node The number of neighbor nodes The condition that the priority level is not less than the preset minimum priority threshold MinPPs is used to determine the node It is a priority core node;

[0045] described Neighborhood Set Defined as:

[0046]

[0047] in, and are the thresholds of uplink and downlink delay distances respectively, and the enhanced node The value is smaller than that of ordinary nodes;

[0048] Control priority core nodes to The non-priority core node in the neighborhood broadcasts an affiliation invitation to establish a cluster affiliation relationship with the non-priority core neighbor node;

[0049] Density association information is exchanged between priority core nodes to establish density direct, density reachable, and density connected relationships, and to construct a local priority core node association data structure.

[0050] As a further improvement of an embodiment of the present invention, the method further includes that the density direct, density reachable and density connected relationships include:

[0051] The density direct relationship is defined as:

[0052] If the node Located at the node of In the neighborhood, and is the priority core node, then the node Depend on The density is directly reached, and the relationship does not satisfy the symmetry;

[0053] The density reachability relation is defined as:

[0054] If there is a node sequence ,in , ,and can be Density direct access, except the last node in the node sequence In addition, all intermediate nodes are priority core nodes, then the node Depend on Density is reachable, the relation satisfies transitivity but not symmetry;

[0055] The density connection relationship is defined as:

[0056] If there is a priority core node , so that the node and All by Density is reachable, then the node and The density is connected, and the relationship satisfies symmetry.

[0057] As a further improvement of an embodiment of the present invention, the method further includes that the clustering mechanism based on ordinary nodes and enhanced nodes controls the node to join the cluster and allocates a network address through the cluster head, including:

[0058] The clustering mechanism of the common nodes includes:

[0059] Determine the status of common nodes;

[0060] In the wandering state, it broadcasts its own attributes and determines whether it is a priority core node; if it is a priority core node, it The non-priority core nodes in the neighborhood send a home invitation and update the local association data structure after receiving the confirmation message;

[0061] In the belonging state, it receives the cluster joining invitation sent by the belonging priority core node, obtains the cluster identifier and network address, and switches to the cluster joining state;

[0062] If a node fails to join the cluster within the specified time limit, it will actively send a request to the nearby cluster nodes to join;

[0063] In the cluster joining state, it processes the broadcast or cluster joining request of the new node and forwards it to the priority core node to which it belongs;

[0064] The clustering mechanism of the enhanced nodes includes:

[0065] Determine the status of the enhanced node;

[0066] In the negotiation state, broadcast attributes and determine whether it is a priority core node; if it is a priority core node, send a home invitation and process the confirmation message and density reachability association information;

[0067] Based on the local priority core node association data, the priority core enhanced nodes with direct density are merged, and the cluster head is elected through the shortest path optimization and weighted connection index;

[0068] The cluster head negotiates the non-conflicting cluster identifier and network address resource pool, allocates its own network address, switches to the cluster entry state, and sends Neighboring nodes send cluster-joining invitations;

[0069] The enhanced nodes that have joined the cluster report the uplink and downlink throughput of the adjacent enhanced nodes outside the cluster to the cluster head. The cluster head selects the enhanced node with the highest throughput as the cluster gateway based on the weighted average throughput. If there is only one enhanced node in the cluster, the cluster head also serves as the gateway.

[0070] If a non-clustered enhanced node fails to join within the specified time limit, it will actively request nearby clustered nodes to join.

[0071] The clustered enhanced nodes process the broadcast or cluster joining requests of new nodes.

[0072] As a further improvement of an embodiment of the present invention, the method further includes dynamically maintaining the cluster structure, specifically including:

[0073] When the remaining energy of the cluster head is lower than the preset threshold or is damaged, the enhanced nodes in the cluster re-elect the cluster head based on the priority and connection index, and notify the nodes in the cluster to update the cluster head information;

[0074] When the remaining energy of the cluster gateway is lower than the preset threshold or is damaged or leaves the cluster, the cluster head re-elects the gateway based on the weighted average throughput;

[0075] The keep-alive mechanism detects nodes in the cluster. If a node does not respond for a preset number of consecutive times, it is considered to have left the cluster, triggering it to send a join request to a nearby node in the cluster.

[0076] For enhanced nodes in negotiation state, ordinary nodes in wandering state or nodes that have lost contact with the cluster head, periodically initiate cluster entry requests to surrounding cluster entry nodes;

[0077] When the change of the priority core nodes in the cluster exceeds a preset change ratio threshold, the cluster head triggers local re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster;

[0078] At preset time intervals, all cluster heads synchronously trigger global re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster.

[0079] To achieve one of the above-mentioned objects of the invention, an embodiment of the present invention further provides a distributed network clustering system for heterogeneous mobile ad hoc networks, the system comprising a broadcast module, a cluster construction module and a cluster management module;

[0080] The broadcast module is used to control the enhanced node or the ordinary node to broadcast its attribute information in response to the startup of the enhanced node or the ordinary node, and receive the attribute information of the neighboring node to obtain the delay distance data and priority data;

[0081] The cluster construction module is used to determine the priority core node based on the delay distance data and priority data of the enhanced nodes or ordinary nodes, and to construct the initial cluster structure in a distributed manner by establishing the cluster affiliation relationship of the enhanced nodes or ordinary nodes and exchanging density association information between the priority core nodes;

[0082] The cluster management module is used to screen enhanced nodes from priority core nodes as cluster head candidates and elect a cluster head from the cluster head candidates; based on the clustering mechanism of enhanced nodes and ordinary nodes, the cluster head controls the enhanced nodes or ordinary nodes to join the cluster and allocates network addresses.

[0083] In order to achieve one of the above-mentioned purposes of the invention, an embodiment of the present invention also provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps in the above-mentioned heterogeneous mobile ad hoc network distributed network clustering method are implemented.

[0084] To achieve one of the above-mentioned objects of the invention, an embodiment of the present invention further provides a storage medium, wherein the storage medium stores a computer program, and is characterized in that when the computer program is executed by a processor, the steps in the above-mentioned heterogeneous mobile ad hoc network distributed network clustering method are implemented.

[0085] Compared with the existing technology, the present invention provides a distributed network clustering method and system for heterogeneous mobile ad hoc networks. Through distributed clustering based on priority core nodes and density reachability, the present invention adapts to the characteristics of heterogeneous nodes and improves the efficiency of intra-cluster autonomy and inter-cluster interconnection. A polymorphic state machine model and a cluster splitting guarantee mechanism are designed to ensure orderly and efficient clustering control and structural stability in a dynamic environment. The clustering strategy is optimized based on delay distance and centrality to build a stable information backbone, taking into account both communication quality and service efficiency, significantly improving the energy efficiency, stability and scalability of large-scale heterogeneous mobile ad hoc networks, and outperforming traditional static or neighbor-oriented clustering methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] Figure 1 It is an overall flow chart of the distributed network clustering method for heterogeneous mobile ad hoc networks described in the present invention.

[0087] Figure 2 It is a topology diagram of a distributed heterogeneous network clustering algorithm based on density and priority competition in the distributed network clustering method of heterogeneous mobile ad hoc networks described in the present invention.

[0088] Figure 3 The present invention provides a schematic diagram of a spanning tree-based IP address pool partitioning mechanism for the distributed network clustering method of heterogeneous mobile ad hoc networks.

[0089] Figure 4 This is a schematic diagram of the cluster selection process of the heterogeneous mobile ad hoc network distributed network clustering method described in the present invention.

[0090] Figure 5 It is a schematic diagram of the architecture of the heterogeneous mobile ad hoc network distributed network clustering system described in the present invention. DETAILED DESCRIPTION

[0091] The present invention will be described in detail below with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional changes made by those skilled in the art based on these embodiments are all within the scope of protection of the present invention.

[0092] The embodiments of the present invention are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and are not to be construed as limiting the present invention.

[0093] In the first embodiment of the present invention, the present invention provides a distributed network clustering method for heterogeneous mobile ad hoc networks, such as Figure 1 As shown, the method includes,

[0094] S1: In response to the start-up of an enhanced node or a common node, the enhanced node or the common node is controlled to broadcast its attribute information and receive attribute information of neighboring nodes to obtain delay distance data and priority data;

[0095] S2: Based on the delay distance data and priority data of enhanced nodes or ordinary nodes, priority core nodes are determined. The initial cluster structure is constructed in a distributed manner by establishing cluster affiliation relationships of enhanced nodes or ordinary nodes and exchanging density association information between priority core nodes.

[0096] S3: Select enhanced nodes from the priority core nodes as cluster head candidates, and elect a cluster head from the cluster head candidates; based on the clustering mechanism of enhanced nodes and ordinary nodes, control the enhanced nodes or ordinary nodes to join the cluster and allocate network addresses through the cluster head.

[0097] In a specific embodiment of the present invention, the priority calculation formula is:

[0098]

[0099] in, To enhance the node collection, is the indicator function, if For enhanced nodes Take 1, otherwise Take 0; is the average damage probability of the node per unit time; is the current remaining energy of the node, is the maximum energy capacity of the node; is the energy attenuation coefficient; is the node mobility index, is the maximum mobility index; is the weight coefficient, and ;

[0100] The average probability of damage The calculation formula is:

[0101]

[0102] Where N is the number of threat source types, The mth type of threat to the node The failure probability per unit time represents the reliability of the node under external threats;

[0103] The energy attenuation coefficient The calculation formula is:

[0104]

[0105] in, For nodes The energy loss rate, are the coefficients of the preset linear function and exponential function, It is a preset energy loss threshold used to measure the nonlinear characteristics of node energy consumption;

[0106] The node mobility index The calculation formula is:

[0107]

[0108] in, For nodes The velocity at time t, is the maximum speed, is the velocity variance, is the maximum velocity variance, is the weight coefficient used to evaluate the stability and dynamics of node movement, and .

[0109] It should be noted that the present invention sets all nodes in the network as , the enhanced nodes with low mobility, low damage rate and high processing capability are set as , the set of common nodes is , and by the formula Compute nodes Priority is used to measure the ability of enhanced nodes to become cluster heads / gateways, or for ordinary nodes to measure their comparative advantages with other ordinary nodes;

[0110] It should be noted that in heterogeneous mobile ad hoc networks, node priority calculation comprehensively considers node type, reliability, energy status, and mobility characteristics to ensure that highly reliable, low-mobility, and energy-efficient nodes are prioritized for key roles (such as core nodes or cluster heads) during clustering. Enhanced nodes, due to their low mobility, high communication capabilities, and low damage rate, receive higher initial weights in the priority calculation through an indicator function, reflecting their dominant role in the network. Ordinary nodes, due to their high mobility and low communication capabilities, receive lower weights. The average damage probability comprehensively assesses a node's failure risk under various threat sources. Enhanced nodes have higher reliability than ordinary nodes, facilitating the selection of priority core nodes. The remaining energy ratio and energy decay coefficient jointly measure a node's energy availability and consumption characteristics. Enhanced nodes generally have a lower energy loss rate, extending their service life. The mobility metric assesses dynamics through node speed and speed variance. The high mobility of ordinary nodes results in a lower stability score, while the low mobility of enhanced nodes increases their priority contribution. The weighting coefficient flexibly balances energy and mobility factors to adapt to the needs of diverse network scenarios. Through this priority calculation mechanism, the present invention can effectively distinguish the heterogeneous characteristics of enhanced nodes and ordinary nodes, optimize the rationality and stability of cluster structure construction, and provide a reliable basis for subsequent priority core node determination and cluster head election. It is particularly suitable for dynamic topology and large-scale heterogeneous network environments.

[0111] In a specific embodiment of the present invention, the delay distance includes the uplink delay distance and downlink delay distance ;

[0112] The downlink delay distance The calculation formula is:

[0113]

[0114] in, is the propagation delay, To handle delay, is the downlink transmission delay, expressed as:

[0115]

[0116] in, is the downlink traffic, is the actual downlink throughput, expressed as:

[0117]

[0118] in, is the node throughput empirical coefficient, is the downlink channel communication capacity, expressed as:

[0119]

[0120] in, The downlink channel bandwidth is determined by the spectrum division of the communication system. value, is the signal power, N is the thermal noise power, expressed as:

[0121]

[0122] in, is the Boltzmann constant, T is the ambient temperature, B is the system bandwidth, and F is the noise factor of the signal receiver at this node, which is used to describe the degree of degradation of the system noise;

[0123] The uplink delay distance The calculation formula is:

[0124]

[0125] in, is the uplink transmission delay, expressed as:

[0126]

[0127] in, is the uplink traffic, is the actual uplink throughput, expressed as:

[0128]

[0129] in, is the uplink channel capacity, which is evaluated by the channel fading model and is approximately expressed as:

[0130]

[0131] in, is the uplink channel bandwidth, is the transmit power, are the antenna gains of the signal transmitter and receiver respectively, c is the speed of light, f is the communication frequency, and x is the Euclidean distance between nodes.

[0132] It should be noted that delay distance calculation is a core component of the clustering process in heterogeneous mobile ad hoc networks. It is used to assess the quality of the communication link between enhanced nodes and standard nodes, supporting neighborhood determination and priority core node selection. Delay distance includes uplink and downlink delay distances, which are composed of transmission delay, propagation delay, and processing delay, respectively. Transmission delay is calculated based on traffic volume and actual throughput. Enhanced nodes, due to their high communication capabilities and low loss rates, typically have higher channel capacity and throughput, resulting in lower transmission delay, especially in communications between enhanced nodes and standard nodes, or between enhanced nodes (e.g., low traffic demand from enhanced nodes to standard nodes). Standard nodes, due to their high mobility and low communication capabilities, have lower channel capacity, resulting in higher transmission delay. Channel capacity is evaluated using bandwidth, signal power, and thermal noise power. Combined with a channel fading model to reflect the effects of Euclidean distance between nodes and communication frequency, enhanced nodes have higher channel efficiency than standard nodes. Propagation delay and processing delay are determined based on node type and network environment to ensure that delay calculation fully reflects the communication characteristics of heterogeneous nodes. This mechanism quantifies link latency, supports the construction of a cluster structure centered on enhanced nodes, optimizes communication efficiency and network stability, and is particularly suitable for dynamic topologies and large-scale heterogeneous network scenarios.

[0133] Furthermore, the uplink channel capacity is limited by the signal power of the signal from the node to the destination node, which is affected by channel fading. If the signal power is directly detected from the receiving node, it will generate additional communication burden. Therefore, we use the channel fading model to evaluate the uplink channel capacity. The node obtains the geographic location of the node that generated the message from the initial broadcast message of other node j. By combining the geographic location of the node itself, the Euclidean distance can be calculated as , let the received power of the signal from this node after it propagates to node j be ,have:

[0134]

[0135] in, is a fixed value and does not need to be obtained from node i; the node receives power Affected by both large-scale and small-scale channel fading, large-scale fading is primarily due to path loss caused by increasing communication distance, and typically changes slowly with increasing distance. Small-scale fading, on the other hand, is caused by rapid signal fluctuations due to multipath effects. This type of fading changes rapidly and can be noticeable within a short period of time or distance.

[0136] Furthermore, the free space loss model is used to characterize large-scale fading, and we have . Using random variables represents small-scale fading that obeys the Rayleigh distribution. The instantaneous signal-to-noise ratio of the channel is:

[0137]

[0138] The average channel capacity is a random variable Perform mathematical expectation calculation based on Rayleigh fading It obeys the exponential distribution (mean is 1). According to Shannon's formula, the average uplink channel capacity is calculated as follows:

[0139]

[0140] In a specific embodiment of the present invention, the step of constructing the initial cluster network in a distributed manner is specifically as follows:

[0141] Based on the delay distance, determine the node of Neighborhood Set , and according to the Number of neighborhood set nodes Not less than the preset minimum node density threshold MinPDs and with a priority lower than node The number of neighbor nodes The condition that the priority level is not less than the preset minimum priority threshold MinPPs is used to determine the node It is a priority core node;

[0142] described Neighborhood Set Defined as:

[0143]

[0144] in, and are the thresholds of uplink and downlink delay distances respectively, and the enhanced node The value is smaller than that of ordinary nodes;

[0145] Control priority core nodes to The non-priority core node in the neighborhood broadcasts an affiliation invitation to establish a cluster affiliation relationship with the non-priority core neighbor node;

[0146] Density association information is exchanged between priority core nodes to establish density direct, density reachable, and density connected relationships, and to construct a local priority core node association data structure.

[0147] It should be noted that the steps of constructing the initial cluster network are intended to form a stable cluster structure in heterogeneous mobile ad hoc networks through a distributed mechanism, making full use of the high communication capability and low mobility of enhanced nodes and the high dynamic characteristics of ordinary nodes. Based on the received attribute information broadcast message, enhanced nodes and ordinary nodes calculate the delay distance (including uplink and downlink) with neighboring nodes to reflect the quality of the communication link between nodes. Enhanced nodes usually have lower delay due to their high channel capacity. In the neighborhood set, enhanced nodes, due to their lower threshold (reflecting high communication efficiency), are more likely to form larger neighborhoods and are therefore selected as priority core nodes. The determination of priority core nodes is based on a comparison of the number and priority of neighboring nodes, ensuring that enhanced nodes, due to their high priority and stable connectivity, dominate cluster formation. As priority core nodes, enhanced nodes send membership invitations, while standard nodes, due to their high mobility, primarily respond as subordinate nodes, forming a cluster structure centered around enhanced nodes. Density-direct, density-reachable, and density-connected relationships are established through message exchanges between priority core nodes. Enhanced nodes leverage their high reliability to establish stable density associations, generating a local association data structure in the form of an undirected graph. This provides a topological basis for subsequent cluster selection and cluster structure optimization. This mechanism effectively adapts to differences in communication capabilities and dynamics among heterogeneous nodes, improving the robustness of cluster partitioning and network scalability.

[0148] In a specific embodiment of the present invention, the relationship between density direct, density reachable and density connected is specifically,

[0149] The density direct relationship is defined as:

[0150] If the node Located at the node of In the neighborhood, and is the priority core node, then the node Depend on The density is directly reached, and the relationship does not satisfy the symmetry;

[0151] The density reachability relation is defined as:

[0152] If there is a node sequence ,in , ,and can be Density direct access, except the last node in the node sequence In addition, all intermediate nodes are priority core nodes, then the node Depend on Density is reachable, the relation satisfies transitivity but not symmetry;

[0153] The density connection relationship is defined as:

[0154] If there is a priority core node , so that the node and All by Density is reachable, then the node and The density is connected, and the relationship satisfies symmetry.

[0155] It should be noted that the definition of density direct, density reachable and density connected relationships is the core mechanism in the clustering process of heterogeneous mobile ad hoc networks, which is used to build a distributed cluster topology dominated by enhanced nodes. The density direct relationship is determined based on the delay distance. In a neighborhood, priority core nodes are primarily enhanced nodes. Their high communication capabilities and low mobility enable them to stably connect to regular nodes within the neighborhood. Regular nodes, due to their high mobility, are often directly accessible as subordinate nodes, reflecting asymmetry. Density reachability is achieved through a node sequence consisting of priority core nodes (typically enhanced nodes). Intermediate nodes in the sequence leverage the enhanced nodes' low loss rate and high reliability to ensure transitivity. Regular nodes serve only as the starting or ending points of the sequence, reflecting their subordinate role. Density connectivity is established based on shared priority core nodes (enhanced nodes). The enhanced nodes' stability and high throughput ensure symmetrical intra-cluster connectivity, forming a stable information backbone. This relationship is implemented through the exchange of density association messages between priority core nodes, generating a local association data structure in the form of an undirected graph. This provides topological support for cluster selection and cluster structure optimization, significantly improving the dynamic adaptability and cluster partitioning efficiency of heterogeneous networks.

[0156] In a specific embodiment of the present invention, a clustering mechanism based on ordinary nodes and enhanced nodes is implemented by controlling the nodes to join the cluster and assigning network addresses, specifically,

[0157] The clustering mechanism of the common nodes includes:

[0158] Determine the status of common nodes;

[0159] In the wandering state, it broadcasts its own attributes and determines whether it is a priority core node; if it is a priority core node, it The non-priority core nodes in the neighborhood send a home invitation and update the local association data structure after receiving the confirmation message;

[0160] In the belonging state, it receives the cluster joining invitation sent by the belonging priority core node, obtains the cluster identifier and network address, and switches to the cluster joining state;

[0161] If a node fails to join the cluster within the specified time limit, it will actively send a request to the nearby cluster nodes to join;

[0162] In the cluster joining state, it processes the broadcast or cluster joining request of the new node and forwards it to the priority core node to which it belongs;

[0163] The clustering mechanism of the enhanced nodes includes:

[0164] Determine the status of the enhanced node;

[0165] In the negotiation state, broadcast attributes and determine whether it is a priority core node; if it is a priority core node, send a home invitation and process the confirmation message and density reachability association information;

[0166] Based on the local priority core node association data, the priority core enhanced nodes with direct density are merged, and the cluster head is elected through the shortest path optimization and weighted connection index;

[0167] The cluster head negotiates the non-conflicting cluster identifier and network address resource pool, allocates its own network address, switches to the cluster entry state, and sends Neighboring nodes send cluster-joining invitations;

[0168] The enhanced nodes that have joined the cluster report the uplink and downlink throughput of the adjacent enhanced nodes outside the cluster to the cluster head. The cluster head selects the enhanced node with the highest throughput as the cluster gateway based on the weighted average throughput. If there is only one enhanced node in the cluster, the cluster head also serves as the gateway.

[0169] If a non-clustered enhanced node fails to join within the specified time limit, it will actively request nearby clustered nodes to join.

[0170] The clustered enhanced nodes process the broadcast or cluster joining requests of new nodes.

[0171] It should be noted that the principle of the clustering algorithm used in the present invention is:

[0172] 1) Each distributed priority core node that has not yet been assigned to a cluster category is used as a distributed seed. The set of nodes that can be densely reached by this priority core node is then found, forming a cluster network. This process continues until all priority core nodes have been assigned to a cluster category. Once clustering is complete, the booster node with the highest weighted coverage capability within the same cluster is selected as the cluster head, and the other booster nodes in the cluster that have the highest combined throughput with booster nodes in other clusters are selected as gateways. 2) We use service latency to define a "distance" measurement method. In this patent scenario, enhanced nodes are often priority core nodes, and the latency distance between enhanced nodes is often very small. Even if we can set a smaller ϵ-ϵ' to avoid all enhanced nodes being densely reachable as priority core nodes, which will lead to too small a number of clusters or even no clustering, this method of adjusting the ϵ-ϵ' value is effective but difficult to flexibly deal with certain special situations. Therefore, we also design a bottom-line mechanism. When there are a large number of priority core nodes that are too far away from the optimal cluster head (for example, more than 6 hops), a cluster is automatically divided into multiple clusters, and the specific cluster head is determined by comparing the centering coverage capabilities of multiple enhanced nodes.

[0173] Through the above steps, this patent can enable the priority core nodes with lower damage rate, higher energy efficiency and weaker mobility to find a clustering structure that promotes higher clustering density based on "delay distance", and ultimately in such a clustering structure, the priority core nodes will play a role similar to the "veins" of leaves, while ordinary nodes follow the priority core nodes, and enhance the interconnection between the nodes and other enhanced nodes, thereby greatly improving the reliability, scalability, orderliness and resource utilization efficiency of the cluster network.

[0174] In a specific implementation scenario of the present invention, the topology diagram of the distributed heterogeneous network clustering algorithm based on density and priority competition is as follows: Figure 2 As shown, yellow nodes and enhanced nodes are priority core nodes. There are at least 5 other nodes in the neighborhood whose priority is not lower than The priorities of at least five other nodes in the neighborhood. Black hollow circles represent non-priority core nodes, and green hollow circles represent discrete nodes. All nodes directly accessible by the priority core node are within the dashed circle centered on the priority core node. Nodes not within this circle are not directly accessible by density. The priority core nodes connected by red bidirectional arrows in the figure above form a sequence of nodes that are directly accessible by density, which is a cluster. Figure 1 The topology shown is divided into three clusters. It can be observed from the figure that the priority core nodes with cross neighborhoods (dashed circles overlap) and their Nodes within a neighborhood (represented by dashed circles) form the same cluster.

[0175] Furthermore, the present invention defines three states for common nodes: wandering state, belonging state, and clustered state. Wandering state means that the node has just started or has left the old cluster network and is in a state of not belonging to a cluster. Belonging state means that the node has established a belonging relationship with other nodes, that is, it belongs to a priority core node. Neighborhood, or the node itself is a priority core node but has established an adjacency relationship with other priority core nodes, so the two Neighboring nodes should also be aggregated into a cluster structure, but the overall cluster structure has not yet been fully determined, that is, they have not yet officially joined a cluster organized by a cluster head and been informed of the cluster ID, cluster head node, cluster gateway node, and other information. The cluster entry state means that the node has joined a cluster and has been informed of the cluster ID and the IP address assigned to the node. For ordinary nodes, the clustering algorithm process mechanism logic is as follows:

[0176] ①Mechanism for ordinary nodes in wandering state to send node information broadcast message: Each ordinary node Calculate its own priority when joining the network , and periodically sends node information broadcast messages to surrounding nodes to inform their own priority , geographic location information, node type information (normal node or enhanced node).

[0177] ② Ordinary nodes in wandering state receive node information broadcast message processing mechanism: Ordinary nodes in wandering state After receiving the broadcast information of neighbor node j, calculate the uplink and downlink delay distance. Belongs to this node of neighborhood, then the node of Number of neighborhood elements Add 1 ( The initial value is 0), and the node is recorded Identification ID information (such as node MAC interface address information) to this node of Neighborhood node information collection; node Get node j priority from broadcast message ,like , then the node Number of neighboring nodes with low priority Add 1 ( The initial value is 0). After the node listens to the broadcast information for a period of time (such as 5 seconds), it compares the and , then this node is determined to be a priority core node.

[0178] ③Mechanism for ordinary nodes in wandering state and belonging to priority core nodes to send priority core node belonging invitation broadcast message: Each ordinary node If it has been determined to be a priority core node, it will broadcast a priority core node affiliation invitation message to neighboring nodes. The invitation message contains the node of The identification IDs of all nodes in the neighborhood node set.

[0179] ④ Wandering ordinary nodes receive priority core node affiliation invitation message processing mechanism: Each wandering ordinary node If a priority core node affiliation invitation message is received, the node ID list is extracted from the message. If the node ID belongs to the list, the node is judged whether it is a priority core node. Otherwise, it means that the node is not a neighbor of the priority core node that sent the message, so the message is not processed. If it is not a priority core node, it will directly return a priority core node confirmation message to the source node that sent the priority core node invitation broadcast message (the message contains the priority core node The identification ID of the node and the identification ID of the priority core node of the destination), then the node If the node receives confirmation messages from multiple source nodes, it will only respond to the message from the first source node by default. If it is a priority core node, the source node information of the priority core node belonging invitation message is recorded to the node The adjacent priority core node association set, then this node Converted to the attributed state.

[0180] ⑤ Priority core node in any state Ordinary node receives priority core node confirmation message processing mechanism: Priority core node in any state Ordinary node After receiving the confirmation message from the priority core node, the source node ID contained in the message is recorded in the local node. 's owned node set, and update the number of elements in the set to .

[0181] ⑥ Ordinary nodes in the belonging state and belonging to the priority core node send density up to the priority core node association message mechanism: each ordinary node in the belonging state If it belongs to the core priority object and has not sent an association message of the density-reachable priority core node within this clustering period, it will send it to this node. Each node in the adjacent priority core node association set sends a density-reachable priority core node association message, which contains the node Identification ID, node type (normal or enhanced node), of The priority core node ID of the neighborhood and the number of elements in the node's already owned node set , and other density-reachable priority core nodes recorded by this node and their Neighborhood priority core node ID, enhanced or ordinary node type, and the number of elements in the attributed node set corresponding to the core object .

[0182] ⑦ The receiving density of ordinary nodes belonging to the priority core node can reach the priority core node associated message processing mechanism: each ordinary node in the belonging state If it belongs to the core priority object, after receiving the density-reachable priority core node association message, the local priority core node association data structure is updated. The data structure is an undirected graph represented as , records the association relationship of all density-reachable priority core nodes obtained by this node. This patent uses the adjacency list data structure to store , maintain a dynamic list (such as a linked list or array) for each priority core node vertex to store all its directly connected Neighborhood priority core nodes, each priority core node has an additional node weight, which is the number of elements in the node set that the object has belonged to, and an additional attribute to indicate whether the node is an enhanced node or a normal node. If the local priority core node association data structure increases priority core nodes or neighbor core objects after the update (note that only the increase in information is considered during this cluster construction cycle, and the decrease in information is not considered), the information that needs to be updated will be encapsulated into a new density-reachable priority core node association message and sent to all nodes except the source node that sent the density-reachable priority core node association message. Neighborhood priority is given to core nodes.

[0183] ⑧Mechanism for ordinary nodes in the belonging state to receive cluster invitation information: Each ordinary node in the belonging state If it does not belong to the core priority object, after receiving the cluster invitation message sent by the priority core node to which it belongs (note that only one cluster invitation message is processed for multiple similar messages), it modifies the cluster ID to which the node belongs according to the cluster ID contained in the message, and modifies the IP address of the node according to the IP address allocated for the node contained in the message, and then converts to the cluster state. Belong to the core priority object, then Obtain the spanning tree topology with the current node as the root node and the densely accessible priority core nodes as branch nodes or leaf nodes from the cluster invitation information, and the assignable IP address range (the cluster invitation information can contain a single IP address to be assigned or a pool of assignable IP addresses), and select the smallest IP host address as the host address of the current node. The node will then be in the cluster state. Take out consecutive IP addresses close to the IP address of this node from the allocable IP address pool. Addresses ( is a fixed coefficient with a value range of [1, 10], providing a certain number of redundant IP addresses that can be allocated), and selects an IP address from them in turn to generate a cluster invitation message (including the cluster ID and the allocated IP address of the cluster to which this node joins), and sends it to this node. of Ordinary nodes in the neighborhood send cluster invitation information; for this node of The priority core node of the neighborhood, node First, determine the subtree structure with itself as the root node according to the spanning tree obtained from the cluster invitation information, then select a subtree with a child node as the root node in turn, calculate the sum of the number of elements in the belonging node set of all nodes in the subtree and set it as , and divide it from the remaining allocatable IP address pool The new IP address pool, the subtree information with the child node as the root node, and the cluster ID are used to generate cluster invitation information and send it to the child node (node of The priority core node in the neighborhood is sent until all child nodes are traversed.

[0184] ⑨ If a common node is still in the wandering state within a period of time (e.g., 8 seconds), it indicates that the common node is a discrete node. In this case, it needs to actively inquire about its cluster joining status from surrounding nodes. If there are cluster joining nodes around, it will send a cluster joining request message to the first cluster joining node detected. After receiving the cluster joining invitation message, it will follow step ⑧ and switch to the cluster joining state.

[0185] ⑩Each common node in the cluster state Processing mechanism for receiving node information broadcast messages sent by other nodes: This describes the node that sent this broadcast message. In the process of just starting up or just leaving the original cluster network, in order to quickly build an updated cluster network topology, if the ordinary node Is the priority core node, then determine the node Is it a node? of Neighboring node, if it is still Send a priority core node affiliation invitation message. When a new node is added to the priority core node's affiliation set, a minimum IP address is taken from the IP address pool reserved for ordinary neighboring nodes, and a cluster invitation message is constructed and sent to the node. send.

[0186] ⑪Each common node in the cluster state Received from other nodes The message sent by the request to join the cluster: If the ordinary node If it is not a priority core node, then The priority core node to which it belongs forwards the cluster entry request message; if the ordinary node It is a priority core node, then it is no longer judged whether the node is a node of Neighboring nodes, directly to the node Send a priority core node affiliation invitation message. When a new node is added to the priority core node's affiliation set, a minimum IP address is taken from the IP address pool reserved for ordinary neighboring nodes, and a cluster invitation message is constructed and sent to the node. send.

[0187] In a specific implementation scenario of the present invention, the IP address pool partitioning mechanism based on the spanning tree is as follows: Figure 3 As shown, assuming that the node Indicated by the letter A, its two child nodes are B and C. The blue area and the green area are the subtrees with B and C as the root nodes respectively. Node A Calculate the number of elements in the attributed node set of all nodes contained in tree B, which is Then, after planning the IP address pool, a cluster invitation message is generated and sent to node B. The same applies to subtree C.

[0188] Furthermore, the present invention defines two states for enhanced nodes: negotiation state and clustering state, as well as two types: gateway node and cluster head node. The negotiation state is the initial state of the enhanced node, indicating that it is receiving density-reachable priority core node association messages from nearby priority core nodes, or is negotiating with other enhanced nodes on specific cluster head selection, cluster ID determination, gateway selection, etc., and cluster division has not yet been completed. The clustering state means that the enhanced node has joined a cluster or has become a cluster head / gateway, and has been informed or negotiated to confirm the cluster ID and the IP address or subnet address assigned to the node. For enhanced nodes, the core point of the clustering algorithm process is that the enhanced node can participate in cluster construction in the same way as an ordinary node as a priority core node (with minor differences in operations), and it is necessary to evaluate the distributed clustering results to determine the appropriate clustering structure range. If necessary, it is necessary to split a cluster divided in the distributed process into multiple clusters to avoid reducing the efficiency of information transmission within the cluster. In addition, it is necessary to achieve cluster ID conflict-free determination and IP address pool division through the interconnection between cluster heads. The specific processing flow is as follows:

[0189] ① Enhanced nodes in negotiation state ( ) The mechanism for sending node information broadcast messages is the same as the ordinary node processing mechanism①.

[0190] ② The enhanced node in the negotiation state receives the node information broadcast message processing mechanism is the same as the ordinary node processing mechanism ②, but The thresholds for the neighborhoods are different.

[0191] ③ The mechanism for an enhanced node in the negotiation state and belonging to a priority core node to send a priority core node affiliation invitation broadcast message is the same as the ordinary node processing mechanism ③.

[0192] ④ The processing mechanism for the enhanced node in the negotiation state to receive the priority core node affiliation invitation message is the same as the ordinary node processing mechanism ④, but it does not need to switch to the affiliation state and remains in the negotiation state.

[0193] ⑤ The processing mechanism for the priority core node enhanced node in any state to receive the priority core node confirmation message is the same as the ordinary node processing mechanism ⑤.

[0194] ⑥ The enhanced node in the negotiation state and belonging to the priority core node sends a message with a density that can reach the priority core node. The association message mechanism is the same as the ordinary node processing mechanism ⑥.

[0195] ⑦ The enhanced node receiving density in the negotiation state and belonging to the priority core node can reach the priority core node associated message processing mechanism is the same as the ordinary node processing mechanism ⑦.

[0196] ⑧ The processing mechanism for an enhanced node in a non-cluster head negotiation state to receive a cluster invitation message is the same as the processing mechanism for a common node ⑧.

[0197] ⑨ Cluster priority selection process: If the local priority core node association data graph structure of the enhanced node that is in the negotiation state and is the core priority object does not change for a period of time (such as within 10 seconds), it is considered that the first stage of the distributed clustering process has converged, and the enhanced node begins to execute the cluster priority selection and cluster entry process. By selecting multiple suitable cluster heads, the cluster structure that may be too long, too wide, or too bloated can be adjusted secondary. Note that if there is only one core priority enhanced node in the cluster, then this node will be selected as the cluster head node. Otherwise, the following steps will be performed:

[0198] After each enhancement node i belonging to the priority core node waits for the local priority core node association data graph structure to converge, it first merges the priority core enhancement nodes that constitute density direct access, mainly linking the ordinary priority core nodes to the same enhancement node entropy. In subsequent operations, the merged nodes are all regarded as a single priority core enhancement node (this can avoid using two density direct access enhancement nodes as cluster heads in subsequent operations, resulting in unreasonable clustering). Then, the shortest path graph between all priority core enhancement nodes and other priority core nodes is calculated according to the Dijkstra algorithm.

[0199] Each enhanced node belonging to a priority core node traverses the shortest path graph with each priority core enhanced node as the destination node. Otherwise, the ratio of the number of nodes with more than MaxHopsNum hops (e.g., 6 hops) to the total number of nodes is calculated, which is recorded as RatioBeyond. The product of all node weights and the number of hops is summed, which is recorded as SumVH. After the traversal is completed, the following operations are performed based on the RatioBeyond value:

[0200]

[0201] ⑩ Cluster head nodes negotiate cluster information: After each cluster has completed the cluster head determination, each cluster head needs to negotiate to determine a non-conflicting cluster ID and IP address resource pool. In the scenario of this patent, the enhanced nodes can form a connectivity graph under the condition of no less than a certain communication link quality, that is, each enhanced node can achieve mutual communication, but the enhanced nodes may not be directly densely reachable. Based on the above conditions, each cluster head enhanced node first determines the cluster ID. The determination method is that each cluster head randomly generates a cluster ID value; then determines the address pool. The determination method is to select a cluster ID from the private network address pool according to the number of nodes in this cluster and reserve a The nodes randomly select a subnet address that meets the address quantity requirement, and then communicate their address pool and cluster ID information. If there are no duplicate IDs, the cluster ID is determined. If there are no subnet address conflicts (a subnet address is a subnet of its own subnet address or there is overlap in the address pool), the subnet address division is complete. If there are duplicate IDs, the cluster head node with the smaller node ID (such as the MAC address) regenerates the cluster ID until there are no duplicates. If there are conflicting subnet addresses, the cluster head node with the smaller node ID reselects the subnet address.

[0202] ⑪ After the cluster head completes the negotiation of cluster ID and subnet address pool, it selects the smallest IP host address from the address pool as the cluster head IP address and then changes to the cluster entry state. Then it takes out the consecutive IP addresses close to the IP address of the node from the allocable IP address pool. addresses, and selects one IP address from them in turn to generate cluster invitation information (including the cluster ID and assigned IP address of this cluster), and sends it to Ordinary nodes in the neighborhood send cluster invitation information; for the cluster head The priority core node of the neighborhood, the cluster head first determines the tree structure with the cluster head as the root node according to the shortest path tree generated with the cluster head as the destination in step ⑨ (a scheme of spanning tree. If the shortest path tree is not generated in step ⑨ because there is only one priority core enhanced node in the cluster, it is generated according to the logic in step ⑨). Then, a subtree with a child node as the root node is selected in turn, and the sum of the number of elements in the attributed node set of all nodes in the subtree is calculated and set as , and divide it from the remaining allocatable IP address pool The new IP address pool, the subtree information with the child node as the root node, and the cluster ID are used to generate cluster invitation information and send it to the child node (node of The priority core node in the neighborhood is sent until all child nodes are traversed.

[0203] ⑫ Cluster Gateway Election Mechanism for Enhanced Nodes in the Cluster Entering State: When an enhanced node transitions to the cluster entering state, it immediately reports its uplink and downlink throughput metrics with other adjacent enhanced nodes to the cluster head. If the cluster head node, transitioning to the cluster entering state, collects throughput information for adjacent enhanced nodes in the cluster from more than three sources or remains unchanged for 6 seconds, it calculates the sum of the uplink and downlink throughputs of each enhanced node in the cluster and all its adjacent non-cluster enhanced nodes. The weighted average throughput is calculated using the ratio of uplink throughput 4 to downlink throughput 6, and the coarse enhanced node with the highest weighted average throughput is selected as the cluster gateway. If the cluster has only one enhanced node, the cluster head directly serves as the gateway node.

[0204] ⑬ If the enhanced node remains in the negotiation state for a period of time (e.g., 8 seconds), it indicates that the enhanced node is a discrete node. In this case, it needs to proactively inquire about the cluster joining status of the surrounding nodes. If there are cluster joining nodes nearby, it sends a cluster joining request message to the cluster joining node detected for the first time. Upon receiving the cluster joining invitation message, it follows step ⑧ and switches to the cluster joining state.

[0205] ⑭Each enhanced node in the cluster state The node information broadcast message processing mechanism received from other nodes is the same as the ordinary node processing mechanism ⑩.

[0206] ⑮Each enhanced node in the cluster state Received from other nodes The message processing mechanism for requesting to join the cluster is the same as that for ordinary nodes⑪.

[0207] In a specific implementation scenario of the present invention, the cluster selection process diagram is as follows: Figure 4 As shown in FIG, by merging nodes A and C, the two one-hop neighbor nodes of node C are directly connected to node A. Figure 4 A clustered network structure containing two enhanced nodes that are priority core nodes is shown. When the network topology is stable, both enhanced nodes A and B need to calculate the shortest path graph with A and B as the destination nodes.

[0208] In a specific embodiment of the present invention, the cluster structure is dynamically maintained, specifically,

[0209] When the remaining energy of the cluster head is lower than the preset threshold or is damaged, the enhanced nodes in the cluster re-elect the cluster head based on the priority and connection index, and notify the nodes in the cluster to update the cluster head information;

[0210] When the remaining energy of the cluster gateway is lower than the preset threshold or is damaged or leaves the cluster, the cluster head re-elects the gateway based on the weighted average throughput;

[0211] The keep-alive mechanism detects nodes in the cluster. If a node does not respond for a preset number of consecutive times, it is considered to have left the cluster, triggering it to send a join request to a nearby node in the cluster.

[0212] For enhanced nodes in negotiation state, ordinary nodes in wandering state or nodes that have lost contact with the cluster head, periodically initiate cluster entry requests to surrounding cluster entry nodes;

[0213] When the change of the priority core nodes in the cluster exceeds a preset change ratio threshold, the cluster head triggers local re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster;

[0214] At preset time intervals, all cluster heads synchronously trigger global re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster.

[0215] It should be noted that the cluster head update switching trigger mechanism specifically includes:

[0216] ① : =( ...

[0217] ② Cluster node damage or loss of connection. At this point, after the other enhanced nodes in the cluster sense the cluster head's loss, they update the enhanced node information according to the cluster's first selection process and reselect the cluster head. The newly selected cluster head sends a cluster head update message to the nearby core priority nodes. The subsequent process is the same as in case ①. If there are no enhanced nodes in the cluster, all ordinary nodes will follow the process of leaving the cluster and rejoining a nearby cluster after losing connection with the cluster head.

[0218] It should be noted that the cluster gateway update switching trigger conditions specifically include:

[0219] ① : = is the cluster gateway's remaining energy threshold. A value below this indicates insufficient energy for the cluster gateway, making it highly unlikely to be able to handle cross-cluster communication. At this point, the cluster gateway sends a message to the cluster head, indicating that it cannot handle cross-cluster communication. The cluster head then selects another node as the cluster gateway based on the local priority core node association data and the cluster gateway selection method.

[0220] ② The cluster gateway leaves the cluster, is damaged, or loses connection. After the cluster head detects and obtains the above information, it selects another node as the cluster gateway according to the cluster gateway selection method.

[0221] It should be noted that the node leaving the cluster mechanism specifically includes:

[0222] ① A keep-alive mechanism is maintained between the nodes in the cluster and the cluster head, that is, regular communication is carried out with the cluster head. If there is no response after three consecutive keep-alive communications, it is determined that the node has left the cluster and the node joins the surrounding cluster mechanism.

[0223] ② After the node is restarted or damaged. After the node restarts, it sends a node information broadcast message, then prepares to receive the priority core node joining invitation message, and finally realizes the cluster joining.

[0224] It should be noted that the mechanism for discrete nodes to join a cluster specifically includes:

[0225] ① The enhanced node remains in the negotiation state for a period of time (e.g. 8 seconds);

[0226] ② Ordinary nodes remain in a wandering state for a period of time (e.g. 8 seconds);

[0227] ③ The node has just determined that it has lost the ability to communicate with the cluster head. If the above three situations occur, it means that the node is a discrete node. At this time, it needs to actively inquire about its cluster joining status from the surrounding nodes. If there are cluster joining nodes around, it will send a cluster joining request message to the cluster joining node detected for the first time.

[0228] Furthermore, in addition to the above-mentioned single-node scale update and maintenance mechanism, if the cluster head finds through the keep-alive mechanism of all nodes in the cluster that the priority core nodes of the cluster have changed by more than 40%, that is, the nodes in the current cluster that are still priority core nodes only account for 60% of the core priority nodes in the initial cluster structure, it means that the cluster structure maintained by the mutual connection of core priority nodes has undergone major changes. At this time, the cluster head notifies the nodes in the cluster to restore the initial state (wandering state or negotiation state) and re-execute the clustering process.

[0229] It should be noted that the single-node and local cluster-scale update mechanisms can address some special cases and avoid frequent global updates to the cluster structure, which can affect communication efficiency. However, global updates to the cluster structure can produce a more optimal and reasonable clustering structure. Therefore, the clustering logic can be re-executed after a longer period of time, improving the rationality of the clustering structure. All cluster head nodes maintain a synchronized global update cycle, with a 30-minute (minute) period (the cycle time can be adjusted based on the speed of specific node changes). Every 30 minutes, the cluster head notifies the nodes in the cluster to resume their initial state (wandering state or negotiation state) and re-execute the clustering process.

[0230] In the second embodiment of the present invention, the present invention provides a distributed network clustering system for heterogeneous mobile ad hoc networks, such as Figure 5 As shown, the system includes a broadcast module 1, a cluster construction module 2 and a cluster management module 3;

[0231] The broadcast module 1 is used to control the enhanced node or the ordinary node to broadcast its attribute information in response to the startup of the enhanced node or the ordinary node, and receive the attribute information of the neighboring node to obtain the delay distance data and priority data;

[0232] The cluster construction module 2 is used to determine the priority core node based on the delay distance data and priority data of the enhanced nodes or ordinary nodes, and to construct the initial cluster structure in a distributed manner by establishing the cluster affiliation relationship of the enhanced nodes or ordinary nodes and exchanging density association information between the priority core nodes;

[0233] The cluster management module 3 is used to select enhanced nodes from the priority core nodes as cluster head candidates and elect a cluster head from the cluster head candidates; based on the clustering mechanism of enhanced nodes and ordinary nodes, the cluster head controls the enhanced nodes or ordinary nodes to join the cluster and allocates network addresses.

[0234] In a third embodiment of the present invention, the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps in the above-mentioned heterogeneous mobile ad hoc network distributed network clustering method are implemented.

[0235] In a fourth embodiment of the present invention, the present invention provides a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps in the above-mentioned heterogeneous mobile ad hoc network distributed network clustering method.

[0236] In summary, the present invention provides a distributed network clustering method and system for heterogeneous mobile ad hoc networks. By adopting distributed clustering based on priority core nodes and density reachability, the method adapts to the characteristics of heterogeneous nodes and improves the efficiency of intra-cluster autonomy and inter-cluster interconnection. A polymorphic state machine model and a cluster splitting guarantee mechanism are designed to ensure orderly and efficient clustering control and structural stability in a dynamic environment. The method optimizes clustering strategies based on delay distance and centrality, builds a stable information backbone, and takes into account both communication quality and service efficiency. This significantly improves the energy efficiency, stability, and scalability of large-scale heterogeneous mobile ad hoc networks, surpassing traditional static or neighbor-oriented clustering methods.

[0237] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process of the modules described above can refer to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0238] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., they may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0239] In addition, the functional modules in each embodiment of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into a single module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0240] The above-mentioned integrated modules implemented in the form of software functional modules can be stored in a computer-readable storage medium. The above-mentioned software functional modules are stored in a storage medium and include a number of instructions for causing a computer system (which may be a personal computer, server, or network system, etc.) or a processor to execute some of the steps of the methods described in various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0241] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A distributed network clustering method for heterogeneous mobile ad hoc networks, characterized by: include, In response to the enhanced node or the ordinary node being started, controlling the enhanced node or the ordinary node to broadcast its attribute information, and receiving attribute information of neighboring nodes to obtain delay distance data and priority data; Based on the delay distance data and priority data of enhanced nodes or ordinary nodes, priority core nodes are determined, and the initial cluster structure is constructed in a distributed manner by establishing cluster affiliation relationships of enhanced nodes or ordinary nodes and exchanging density association information between priority core nodes; Select enhanced nodes from the priority core nodes as cluster head candidates, and elect a cluster head from the cluster head candidates; Based on the clustering mechanism of enhanced nodes and common nodes, the cluster head controls the enhanced nodes or common nodes to join the cluster and allocates network addresses.

2. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 1, characterized in that: The calculation formula of the priority data is: ； in, To enhance the node set, is the indicator function, if For enhanced nodes Take 1, otherwise Take 0; is the average damage probability of the node per unit time; is the current remaining energy of the node, is the maximum energy capacity of the node; is the energy attenuation coefficient; is the node mobility index, is the maximum mobility index; is the weight coefficient, and ; The average probability of damage The calculation formula is: ； Where N is the number of threat source types, The mth type of threat to the node The failure probability per unit time represents the reliability of the node under external threats; The energy attenuation coefficient The calculation formula is: ； in, For nodes The energy loss rate, are the coefficients of the preset linear function and exponential function, It is a preset energy loss threshold used to measure the nonlinear characteristics of node energy consumption; The node mobility index The calculation formula is: ； in, For nodes The velocity at time t, is the maximum speed, is the velocity variance, is the maximum velocity variance, is the weight coefficient used to evaluate the stability and dynamics of node movement, and .

3. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 2, characterized in that: The delay distance includes the uplink delay distance and downlink delay distance ; The downlink delay distance The calculation formula is: ； in, is the propagation delay, To handle delay, is the downlink transmission delay, expressed as: ； in, is the downlink traffic, is the actual downlink throughput, expressed as: ； in, is the node throughput empirical coefficient, is the downlink channel communication capacity, expressed as: ； in, The downlink channel bandwidth is determined by the spectrum division of the communication system. value, is the signal power, N is the thermal noise power, expressed as: ； in, is the Boltzmann constant, T is the ambient temperature, B is the system bandwidth, and F is the noise factor of the signal receiver at this node, which is used to describe the degree of degradation of the system noise; The uplink delay distance The calculation formula is: ； in, is the uplink transmission delay, expressed as: ； in, is the uplink traffic, is the actual uplink throughput, expressed as: ； in, is the uplink channel capacity, which is evaluated by the channel fading model and is approximately expressed as: ； in, is the uplink channel bandwidth, is the transmit power, are the antenna gains of the signal transmitter and receiver respectively, c is the speed of light, f is the communication frequency, and x is the Euclidean distance between nodes.

4. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 3, characterized in that: The step of constructing the initial cluster network in a distributed manner includes: Based on the delay distance, determine the node of Neighborhood Set , and according to the Number of neighborhood set nodes Not less than the preset minimum node density threshold MinPDs and with a lower priority than node The number of neighbor nodes The condition that the priority level is not less than the preset minimum priority threshold MinPPs is used to determine the node It is a priority core node; described Neighborhood Set Defined as: ； in, and are the thresholds of uplink and downlink delay distances respectively, and the enhanced node The value is smaller than that of ordinary nodes; Control priority core nodes to The non-priority core node in the neighborhood broadcasts an affiliation invitation to establish a cluster affiliation relationship with the non-priority core neighbor node; Density association information is exchanged between priority core nodes to establish density direct, density reachable, and density connected relationships, and to construct a local priority core node association data structure.

5. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 4, characterized in that: The density direct, density reachable and density connected relationships include: The density direct relationship is defined as: If the node Located at the node of In the neighborhood, and is the priority core node, then the node Depend on The density is directly reached, and the relationship does not satisfy the symmetry; The density reachability relation is defined as: If there is a node sequence ,in , ,and can be Density direct access, except the last node in the node sequence In addition, all intermediate nodes are priority core nodes, then the node Depend on Density is reachable, the relation satisfies transitivity but not symmetry; The density connection relationship is defined as: If there is a priority core node , so that the node and All by The density is reachable, then the node and The density is connected, and the relationship satisfies symmetry.

6. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 1, characterized in that: The clustering mechanism based on common nodes and enhanced nodes includes: The clustering mechanism of the common nodes includes: Determine the status of common nodes; In the wandering state, it broadcasts its own attributes and determines whether it is a priority core node; If it is a priority core node, The non-priority core nodes in the neighborhood send a home invitation and update the local association data structure after receiving the confirmation message; In the belonging state, it receives the cluster joining invitation sent by the belonging priority core node, obtains the cluster identifier and network address, and switches to the cluster joining state; If a node fails to join the cluster within the specified time limit, it will actively send a request to the nearby cluster nodes to join; In the cluster joining state, it processes the broadcast or cluster joining request of the new node and forwards it to the priority core node to which it belongs; The clustering mechanism of the enhanced nodes includes: Determine the status of the enhanced node; In the negotiation state, broadcast attributes and determine whether it is a priority core node; If it is a priority core node, it sends a joining invitation and processes the confirmation message and density reachability association information; Based on the local priority core node association data, the priority core enhanced nodes with direct density are merged, and the cluster head is elected through the shortest path optimization and weighted connection index; The cluster head negotiates the non-conflicting cluster identifier and network address resource pool, allocates its own network address, switches to the cluster entry state, and sends Neighboring nodes send cluster-joining invitations; The enhanced nodes that have joined the cluster report the uplink and downlink throughput of the adjacent enhanced nodes outside the cluster to the cluster head. The cluster head selects the enhanced node with the highest throughput as the cluster gateway based on the weighted average throughput. If there is only one enhanced node in the cluster, the cluster head also serves as the gateway. If a non-clustered enhanced node fails to join within the specified time limit, it will actively request nearby clustered nodes to join. The clustered enhanced nodes process the broadcast or cluster joining requests of new nodes.

7. The distributed network clustering method for heterogeneous mobile ad hoc networks according to claim 6, characterized in that: The method further includes dynamically maintaining the cluster structure, specifically including: When the remaining energy of the cluster head is lower than the preset threshold or is damaged, the enhanced nodes in the cluster re-elect the cluster head based on the priority and connection index, and notify the nodes in the cluster to update the cluster head information; When the remaining energy of the cluster gateway is lower than the preset threshold or is damaged or leaves the cluster, the cluster head re-elects the gateway based on the weighted average throughput; The keep-alive mechanism detects nodes in the cluster. If a node does not respond for a preset number of consecutive times, it is considered to have left the cluster, triggering it to send a join request to a nearby node in the cluster. For enhanced nodes in negotiation state, ordinary nodes in wandering state or nodes that have lost contact with the cluster head, periodically initiate cluster entry requests to surrounding cluster entry nodes; When the change of the priority core nodes in the cluster exceeds a preset change ratio threshold, the cluster head triggers local re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster; At preset time intervals, all cluster heads synchronously trigger global re-clustering, notifying the nodes in the cluster to restore the initial state and re-cluster.

8. A distributed network clustering system for heterogeneous mobile ad hoc networks, characterized by: Includes broadcast module, cluster building module and cluster management module; The broadcast module is used to control the enhanced node or the ordinary node to broadcast its attribute information in response to the startup of the enhanced node or the ordinary node, and receive the attribute information of the neighboring node to obtain the delay distance data and priority data; The cluster construction module is used to determine the priority core node based on the delay distance data and priority data of the enhanced nodes or ordinary nodes, and to construct the initial cluster structure in a distributed manner by establishing the cluster affiliation relationship of the enhanced nodes or ordinary nodes and exchanging density association information between the priority core nodes; The cluster management module is used to select enhanced nodes from the priority core nodes as cluster head candidates and elect a cluster head from the cluster head candidates; Based on the clustering mechanism of enhanced nodes and common nodes, the cluster head controls the enhanced nodes or common nodes to join the cluster and allocates network addresses.

9. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program that can be run on the processor, and when the program is executed on the processor, the steps of the distributed network clustering method for heterogeneous mobile ad hoc networks as claimed in any one of claims 1 to 7 are implemented.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the distributed network clustering method for heterogeneous mobile ad hoc networks as claimed in any one of claims 1 to 7 are implemented.

Images