A distributed detection method and system for key nodes of a wireless multi-hop network

Abstract

A kind of distributed detection method and system of wireless multi-hop network key node, method includes: each node constructs adjacent information with set time interval, including: broadcast neighbor collection frame and neighborhood collection frame, receive neighbor set and neighborhood set, merge duplicate node, calculate degree of connection;Broadcast degree collection frame, receive neighbor degree of connection and neighborhood information, merge degree value, calculate key performance score;When sending, receiving and forwarding data each time, the traffic count value is added 1;Server traverses all nodes, accesses node using depth-first mode, collects key performance score and traffic count value, builds and updates score set and traffic set;When the maximum traffic count value in traffic set exceeds set threshold ratio, server broadcasts data contraction frame, node compresses traffic count value, and uploads the count result compressed in next traversal;According to score set and traffic set, key node is marked, and the distributed detection of key node is completed.

Description

Technical Field

[0001] This invention belongs to the field of wireless multi-hop network technology, and more specifically, relates to a distributed detection method and system for key nodes in wireless multi-hop networks. Background Technology

[0002] The application of wireless multi-hop networks is becoming increasingly widespread. They are used in many fields such as mine disaster early warning, agricultural information monitoring, drone collaboration, and military command and control. They can transmit information between nodes through self-organizing networks, avoiding the need for pre-fixed deployment of centralized networks. They are highly flexible and versatile.

[0003] An example of a wireless multi-hop network is as follows Figure 1 As shown, multiple nodes are interconnected through a limited radio coverage area. Node 14 is the server, while the other nodes 1-13 are sensors or distributed system devices. Each device can transmit data to the server step by step through other devices as stepping stones.

[0004] However, wireless multi-hop networks also have certain drawbacks. Due to the possibility of network nodes failing, moving, or shutting down, their distribution density varies greatly. Moreover, the radio coverage of a single network node is limited. If some network nodes stop working or move, the entire originally connected network will be broken into multiple scattered and isolated networks, which will result in information not being able to be traversed and transmitted. These nodes are called critical nodes.

[0005] For example Figure 1 If any one of the nodes (4-8) in the network fails or shuts down, the network will be split in two, and data from some nodes will be unable to be transmitted to the server. Furthermore, critical nodes act as bridges connecting multiple sub-networks; information exchange between these sub-networks will be concentrated at these critical nodes, leading to data congestion, which is detrimental to the reliable and efficient operation of the entire network.

[0006] For the reasons mentioned above, it is necessary to identify the key nodes of the entire wireless multi-hop network through certain strategies, monitor the data traffic density on the key nodes in real time, and equip the key nodes with redundant devices to avoid the network being divided into multiple complementary and interconnected sub-networks due to the failure, movement, or shutdown of the key nodes, thereby ensuring the efficient and reliable flow of information.

[0007] Traditional key node identification in wireless multi-hop networks is centralized. This centralized approach relies on acquiring the entire network topology before analysis and processing. However, acquiring and processing the overall network topology is inefficient, consuming excessive computational and network resources. Furthermore, due to the variability of wireless multi-hop networks, this process needs to be repeated periodically, resulting in significant network communication and computational loads.

[0008] Currently, some studies have attempted to use distributed detection methods to identify key nodes in order to reduce the dependence of centralized solutions on the entire network topology and the consumption of computing resources. However, these methods generally suffer from problems such as low detection accuracy, untimely updates, and weak response mechanisms, and are still difficult to meet the actual needs of wireless multi-hop networks for dynamic identification and high-frequency monitoring of key nodes.

[0009] Therefore, there is an urgent need for a distributed detection method for key nodes that does not require centralized control, has real-time update capabilities, and consumes fewer resources, in order to adapt to practical application scenarios where the structure of wireless multi-hop networks changes frequently. Summary of the Invention

[0010] To address the shortcomings of existing technologies, this invention provides a distributed detection method and system for key nodes in wireless multi-hop networks. The aim is to quickly and accurately detect key nodes in wireless multi-hop networks. Using a distributed approach, each node performs calculations and detection locally using data from itself and neighboring nodes. This provides network administrators with fast and effective data support, enabling them to perform targeted network redundancy backups or scheduling adjustments, improving the connection reliability of wireless multi-hop networks, and ensuring more efficient data flow.

[0011] The present invention adopts the following technical solution.

[0012] The first aspect of the present invention provides a distributed detection method for key nodes in a wireless multi-hop network, including statistical processing logic for a single node and statistical processing logic for the entire network, and the step of detecting key nodes;

[0013] The statistical processing logic for a single node includes: each node broadcasts neighbor collection frames and neighborhood collection frames to the outside world, and the nodes connected to the first hop reply with frame data. The degree of each node is calculated in a distributed manner. Based on the degree value of each node and the degree values ​​of all nodes connected to the first hop, the key performance score of a single node is calculated. The traffic count value of each node is also counted.

[0014] The overall network statistical processing logic and the steps for detecting key nodes include: the server traverses each node, iteratively compares the key performance score and traffic count value of each node, and selects the top-ranked nodes in terms of key performance score and traffic count value as key nodes.

[0015] Preferably, each node broadcasts neighbor collection frames and neighborhood collection frames, and nodes with one-hop connections reply with frame data, thereby distributively calculating the degree value of each node, including:

[0016] In a wireless multi-hop network, each node periodically broadcasts a neighbor collection frame. Nodes that receive a neighbor collection frame reply with the sending node ID and the receiving node ID. The replies to the neighbor collection frames are counted to form a set containing the neighbor node IDs.

[0017] The system broadcasts neighborhood collection frames to the outside world, counts the replies to neighborhood collection frames, and forms a set of neighborhood node IDs containing the sending node ID, the receiving node ID, and the IDs of its neighboring nodes.

[0018] Each node deduplicates the collected set of neighboring node IDs and calculates the degree of each node.

[0019] Preferably, the calculation of the key performance score of a single node based on the degree value of each node and the degree values ​​of all nodes connected to it via its one-hop includes:

[0020] After each node calculates its own degree value, it broadcasts a degree collection frame to the outside world and counts the degree collection frame responses from neighboring nodes.

[0021] First, merge the maximum values ​​of the degree values;

[0022] The sum of all nodes in the neighbor set after merging the maximum value is calculated by subtracting 1 from the degree values ​​of all nodes. The sum is then added to 1 and used as the denominator. The square of the degree value of each node is used as the denominator. The ratio of the two is the critical performance score of a single node.

[0023] Preferably, the statistical flow count values ​​for each node include:

[0024] Whenever data is sent, received, transmitted, or forwarded on a node, the node's traffic count is incremented by 1.

[0025] Whenever a node receives a data shrinking frame, it performs a square root operation on the node's traffic count value and updates the previous traffic count value.

[0026] Preferably, the server traverses all nodes at set time intervals, accesses nodes using a depth-first approach, collects key performance scores and traffic counts, and constructs a score set and a traffic set, including:

[0027] The server sends a traversal statistics frame, which includes three parts: statistical values, a set of key performance scores, and a set of traffic counts. The statistical values ​​are the set target number of key nodes, and the set of key performance scores and the set of traffic counts are initially empty.

[0028] The server accesses each node sequentially. After receiving the traversal statistics frame, each node checks whether the number of nodes recorded in the key performance score set and the traffic count set is less than the statistical value. If the key performance score set is not full, the data pair consisting of the current node identifier and the key performance score is added to the key performance score set. If the traffic count set is not full, the data pair consisting of the current node identifier and the traffic count value is added to the traffic count set.

[0029] After traversing the statistical frames and completing all node visits, the system backtracks to the server level by level along the original traversal path. When receiving the backtracking frames, the intermediate nodes along the statistical frames update the copies of the key performance score set and traffic count set stored locally on the node, which will be used as the initial reference in the next round of traversal.

[0030] Preferably, the step of updating the score set and traffic set based on the key performance scores and traffic counts uploaded by the nodes, replacing the node data with the smallest key performance score in the set when the set is full, and replacing the node data with the smallest traffic count when multiple nodes have the same score, includes:

[0031] Compare the current node’s critical performance score with the critical performance scores of each data pair in the critical performance score set. If the current critical performance score is larger, replace the data pair with the smallest critical performance score in the critical performance score set.

[0032] If multiple pairs of data have the same key performance score, continue to compare the corresponding traffic count values ​​and replace the pairs with smaller traffic count values ​​with the pairs with larger traffic count values.

[0033] The same logic applies to the traffic count set. If the current traffic count value is greater than the traffic count value of a certain data pair in the set, then the data pair consisting of the current node identifier and the traffic count value is used to replace it.

[0034] Preferably, when the maximum traffic count value in the traffic set exceeds a set threshold ratio, the server broadcasts a data shrinking frame, including:

[0035] After completing a node traversal and statistics, the server extracts the traffic count values ​​of all nodes in the current traffic set and calculates the maximum value among them;

[0036] If the maximum traffic count exceeds the set percentage threshold of the maximum allowed value of stored data, the server will broadcast a data contraction frame to all nodes in a traversal manner, triggering each node to perform traffic compression processing.

[0037] Preferably, the node performs compression processing on the traffic count value, including:

[0038] After receiving the data shrinking frame broadcast by the server, the current node performs a square root operation on the current traffic count value and replaces the original traffic count value of the current node with the result of the operation; during the next traversal access of the server, it uploads the compressed traffic count value.

[0039] Preferably, after the traversal is completed, the server marks key nodes according to the score set and the traffic set, completing the distributed detection of key nodes, including:

[0040] After the statistical frames are traversed and fall back to the server, the server reads the node identifiers in the score set and the traffic set respectively to confirm all the nodes contained in the set; and marks all the nodes in the score set and the traffic set as key nodes.

[0041] After the detection process is completed, a new round of node information construction and traversal process is started, forming a distributed detection mechanism for key nodes that is executed in a loop.

[0042] The second aspect of the present invention provides a distributed detection system for key nodes in a wireless multi-hop network, which operates the distributed detection method for key nodes in a wireless multi-hop network as described in the first aspect, including a server and distributed single-node statistical processing logic modules.

[0043] The distributed single-node statistical processing logic module broadcasts neighbor collection frames and neighborhood collection frames to the outside world. Nodes with one-hop connections reply with frame data, thereby calculating the degree of each node in a distributed manner. Based on the degree value of each node and the degree values ​​of all nodes connected to it in one hop, the module calculates the key performance score of each node and counts the traffic count value of each node.

[0044] The server is used to traverse each node, iteratively compare the key performance scores and traffic counts of each node, and select the top-ranked nodes by key performance scores and traffic counts as key nodes.

[0045] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0046] This invention constructs adjacency information through local communication between nodes and improves the representativeness and accuracy of key node identification by combining structural features and traffic information in a joint evaluation method. It employs a server-based traversal collection mechanism to construct key performance score sets and traffic sets, and dynamically updates the set content based on replacement rules, ensuring the continuity and adaptability of the detection results. By combining real-time data traffic statistics and threshold judgment, it triggers data contraction control operations to alleviate data transmission pressure on high-frequency nodes during network operation, avoiding connection failures caused by local congestion. Finally, through periodic traversal and judgment, it constructs a cyclically executable distributed key node detection process that does not rely on centralized control, achieving low-overhead, high-stability dynamic monitoring capabilities for key nodes, suitable for wireless multi-hop network scenarios with frequent structural changes.

[0047] The method of this invention can quickly and accurately identify key nodes in the entire wireless multi-hop network. Its identification method takes into account both the local structural scoring of nodes and traffic data, resulting in a more comprehensive evaluation. Since the structural scoring uses a distributed method, it avoids the process of acquiring the topology of the entire wireless multi-hop network. The overall operation is efficient and simple, with wide applicability; it only requires a full network traversal and statistical analysis at regular intervals, and the communication overhead for detecting key nodes in the entire network is also greatly reduced. Attached Figure Description

[0048] Figure 1 This is a diagram of a wireless multi-hop network connection based on existing technology;

[0049] Figure 2 This is a flowchart of the processing logic for a single node provided according to an embodiment of the present invention;

[0050] Figure 3 This is a flowchart of the overall processing logic of a wireless multi-hop network provided in accordance with an embodiment of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0052] Embodiment 1 of the present invention provides a distributed detection method for key nodes in a wireless multi-hop network, including statistical processing logic for a single node and statistical processing logic for the entire network, and a step for detecting key nodes. The statistical processing logic for a single node uses distributed computing to obtain the key performance scores of each node. Based on the statistical processing logic for a single node, the steps of the processing logic for the entire network count traffic and score data. The two work together to detect key nodes, thus fully realizing the distributed detection of key nodes in a multi-hop wireless network.

[0053] Specifically, such as Figure 2 As shown, the specific statistical processing logic for a single node includes:

[0054] Step A.1: Each node in the wireless multi-hop network periodically broadcasts a neighbor collection frame. The node that receives the neighbor collection frame replies with the sending node ID and the receiving node ID.

[0055] Specifically, each node in a wireless multi-hop network periodically broadcasts information to its one-hop connected neighbor set H. u The nodes in the process send neighbor collection frames, with node u representing the external broadcasting node. The neighbor collection frame contains the ID of the sending node u. The specific frame structure can be of any type, as long as the node receiving the frame can recognize it as a neighbor collection frame.

[0056] Let node v represent the node that receives the neighbor collection frame. Node v only needs to reply and does not need to forward the neighbor collection frame. The reply frame data needs to indicate the ID of the node u that sent the neighbor collection frame and the ID of the node v that replied to the neighbor collection frame.

[0057] Step A.2: Count the neighbor collection frame replies to form a set containing neighbor node IDs; broadcast the neighbor collection frame externally, count the neighbor collection frame replies to form a set of neighbor node IDs containing the sending node ID, the receiving node ID and its neighbor node IDs.

[0058] Specifically, node u counts the neighbor set H it receives. u After receiving neighbor collection frames from each one-hop neighbor node, the neighbor node IDs are aggregated and formed into a set J of neighbor node IDs. u Then node u broadcasts to its one-hop connected neighbor set H. u The nodes in the process send neighborhood collection frames. The specific frame structure can be of any type, as long as the node receiving the frame can identify it as a neighborhood collection frame, and the neighborhood collection frame must contain the ID of the sending node u.

[0059] Node v that receives the neighborhood collection frame only needs to reply and does not need to forward the neighborhood collection frame. The reply frame data needs to indicate the ID of the node u that sent the neighborhood collection frame, the ID of the node v that replied to the neighborhood collection frame, and the set of its neighbor node IDs J. v .

[0060] Step A.3: Each node performs deduplication on the responses to the collected neighboring collection frames, that is, deduplicatively processes the set of neighboring node IDs and calculates the degree of each node.

[0061] Specifically, it is necessary to traverse the neighbor set J of each neighbor node v. v If two neighboring nodes each contain the other in their respective neighbor sets, then these two nodes are considered as one node. After deduplication, the degree of node u is calculated using the following formula:

[0062] DGR u = The number of neighbor nodes of u after deduplication

[0063] In the formula:

[0064] DGR u This represents the degree of node u.

[0065] Step A.4: After each node calculates its own degree, it broadcasts a degree collection frame to the outside world and counts the degree collection frame replies from neighboring nodes.

[0066] Specifically, after node u calculates its own degree, it broadcasts it to its one-hop connected neighbor set H. u The nodes in the process send degree collection frames. The degree collection frame contains the ID of the sending node u, and the specific frame structure can be of any type, as long as the node receiving the frame can identify it as a degree collection frame.

[0067] Node v that receives the degree collection frame needs to reply instead of forwarding it. The reply frame data needs to indicate the ID of the node u that sent the degree collection frame and the calculated degree DGR of the replying node v. v And the set J of neighboring nodes of node v in the response collection frame. v .

[0068] Step A.5: Each node calculates its own performance score based on the degree of its response.

[0069] Specifically, after node u receives the degree of its one-hop neighbor, it first performs maximum merging on the degree and then calculates the score. Maximum merging involves merging the degrees of nodes that are each other's one-hop neighbors, treating two nodes as one node, and the degree value equals the maximum of the degrees of the two mutually one-hop neighbor nodes, expressed by the following formula:

[0070] DGR 合并 = max(degree DGR of each node that is a one-hop neighbor)

[0071] In the formula:

[0072] DGR 合并 This represents the degree after merging the maximum values.

[0073] Let set M represent the set of neighboring nodes after the maximum value merging process. Calculate the key performance score of node u using the following formula:

[0074]

[0075] In the formula:

[0076] SCORE u This represents the key performance score of node u.

[0077] Step A.6: Whenever data is sent, received, transmitted, or forwarded on node u, the traffic count value LL of node u is incremented. u Add 1.

[0078] Step A.7: Whenever node u receives a data contraction frame, the traffic count value of node u is square-rooted as follows, expressed by the following formula:

[0079]

[0080] In the formula:

[0081] LL u This represents the flow count value of node u.

[0082] Repeat the above steps to continuously process individual nodes.

[0083] Specifically, such as Figure 3 As shown, the statistical processing logic of the overall network and the steps for detecting key nodes specifically include:

[0084] Step B.1: The server periodically sends traversal statistics frames. The initial traversal statistics frame contains the statistical value K, the traffic set L, and the score set P.

[0085] Specifically, the server periodically sends traversal statistics frames at regular intervals. The traversal method can be any of the following: depth-first traversal or breadth-first traversal. The key is to add a forward and backward mechanism to the traversal process. Forward refers to the process of visiting the child node after visiting the parent node, while backward refers to the process of sending feedback information to the parent node of the child node after visiting the child node.

[0086] The initial traversal statistics frame sent by the server contains a statistical value K, a traffic set L, and a rating set P, where the initial values ​​of the traffic set L and the rating set P are both empty.

[0087] Step B.2: When the child node u visited in the forward traversal direction receives the traversal statistics frame, it will compare and count the number of tuples in the set traffic set L and the set score set P.

[0088] If the number of tuples in the traffic set L is less than the statistical value K, then the ID of node u and the traffic data are combined into a tuple and added to the traffic set L; if the number of tuples in the rating set P is less than the statistical value K, then the ID of node u and the rating data are combined into a tuple and added to the rating set P; if the number of tuples in the traffic set L is greater than or equal to the statistical value K, continue to step B.3.

[0089] Step B.3: Compare the flow of each tuple in the flow set L with the flow of node u. If the flow of node u is greater than the flow count in a certain tuple in the flow set L, then replace the original tuple in the flow set L with the tuple consisting of node u's ID and flow. If the number of tuples in the rating set P is greater than or equal to the statistical value K, then compare the rating of each tuple in the rating set P with the rating of node u. If the rating of node u is greater than the rating in a certain tuple in the rating set P, then replace the original tuple in the rating set P with the tuple consisting of node u's ID and rating.

[0090] Step B.4: During the traversal rollback access process, when the parent node v receives the rollback data of the traversal statistics frame, it directly transmits the latest statistical value K, traffic set L, and score set P contained in the rollback access frame, and updates the data of statistical value K, traffic set L, and score set P stored locally on the parent node.

[0091] Step B.5: When a traversal and statistics are completed, the traversal and statistics frame will fall back to the server. The server will count the tuples in the traffic set L and the score set P, and then mark the corresponding nodes of ID in the traffic set L and the score set P as key nodes. Administrators can query the specific information of key nodes on the server.

[0092] Step B.6: After the server detects that a traversal and statistics have been completed, it will check the tuples in the traffic set L. If the largest traffic value in the tuple exceeds a certain percentage of the maximum allowed value of stored data, preferably but not limited to exceeding 80%, the server will send a data shrinking frame in a traversal manner.

[0093] The node that receives the data contraction frame takes the square root of the traffic data, and the specific processing logic is the same as in step 1.7.

[0094] Repeat steps B.1-B.6 to continuously complete the statistical processing of the entire network.

[0095] Example 2 of the present invention provides a distributed detection method for key nodes in a wireless multi-hop network, comprising the following steps:

[0096] Step 1: Broadcast neighbor collection frames and neighborhood collection frames, receive neighbor sets and neighborhood sets, merge duplicate nodes, and calculate connectivity.

[0097] Preferably, step 1 specifically includes:

[0098] Step 1.1: The current node broadcasts a neighbor collection frame to its one-hop neighbor node. After receiving the frame, the one-hop neighbor node returns an adjacency response frame containing its one-hop neighbor information. The current node receives multiple adjacency response frames and merges them to form the current node's neighbor set.

[0099] Specifically, each node in a wireless multi-hop network periodically broadcasts information to its one-hop connected neighbor set H. u Nodes in the network send neighbor collection frames, with u representing the name of the node. The neighbor collection frame contains the ID of the sending node u. The specific frame structure can be of any type, as long as the node receiving the frame can recognize it as a neighbor collection frame.

[0100] Let v represent the node that receives the neighbor collection frame. Node v only needs to reply and does not need to forward the neighbor collection frame. The reply frame data needs to indicate the ID of the node u that sent the neighbor collection frame and the ID of the node v that replied to the neighbor collection frame. Node u counts the received H u After receiving neighbor collection frames from each one-hop neighbor node, the neighbor node IDs are aggregated and formed into a set J of neighbor node IDs. u .

[0101] Step 1.2: The current node broadcasts a neighborhood collection frame to each neighboring node in its neighbor set. After receiving the frame, each neighboring node returns a neighborhood response frame containing the neighboring node's identifier and degree value. The current node merges multiple neighborhood response frames to form its neighborhood set.

[0102] Specifically, node u broadcasts to its one-hop connected neighbor set H. u The nodes in the process send neighborhood collection frames. The specific frame structure can be of any type, as long as the node receiving the frame can identify it as a neighborhood collection frame, and the neighborhood collection frame must contain the ID of the sending node u.

[0103] Node v that receives the neighborhood collection frame only needs to reply and does not need to forward the neighborhood collection frame. The reply frame data needs to indicate the ID of the node u that sent the neighborhood collection frame, the ID of the node v that replied to the neighborhood collection frame, and the set of its neighbor node IDs J. v .

[0104] Step 1.3: After receiving the neighborhood set, the current node identifies the neighboring node identifiers. If there are neighboring nodes with the same identifier, they are merged into a one-hop neighboring node. When multiple neighboring nodes are one-hop adjacent to each other, the neighboring node with the highest retention value is retained, and the connectivity information of the other neighboring nodes is discarded. All the retained neighboring nodes are used as the first neighboring set, where the connectivity is the number of neighboring nodes retained after merging.

[0105] Specifically, it involves traversing the neighbor set J of each neighbor node v. v If two neighboring nodes each contain the other in their respective neighbor sets, then these two nodes are considered as one node. After deduplication, the connectivity degree (DGR) of node u is calculated. u It can be expressed by the following formula:

[0106] DGR u = The number of neighboring nodes of node u after deduplication

[0107] In the formula:

[0108] DGR u This represents the connectivity of node u.

[0109] Step 2: Broadcast degree collection frames, receive neighbor connectivity and neighborhood information, merge degree value, and calculate performance score.

[0110] Preferably, step 2 specifically includes:

[0111] Step 2.1: The current node broadcasts a degree collection frame to its one-hop neighboring nodes, and the neighboring nodes return response information containing their own connectivity degree and their own set of one-hop neighboring nodes;

[0112] Specifically, after node u calculates its own degree, it broadcasts it to its one-hop connected neighbor set H. u The nodes in the process send degree collection frames. Each degree collection frame contains the ID of the sending node u. The specific frame structure can be of any type, as long as the receiving node can recognize it as a degree collection frame. The node v that receives the degree collection frame needs to reply, but does not need to forward it. The reply frame data needs to indicate the ID of the node u that sent the degree collection frame and the calculated degree DGR of the replying node v. v And the set J of neighboring nodes of node v in the response collection frame. v .

[0113] Step 2.2: After the current node receives all response information, it constructs a response adjacency set; if there are multiple nodes that are adjacent to each other in the response adjacency set, the node with the largest degree value is retained and the rest are discarded;

[0114] Step 2.3: The key performance score is obtained by the ratio between the square of the current node's connectivity and the sum of the degree values ​​of all nodes in the response adjacency set minus 1.

[0115] Specifically, after node u receives the degree of its one-hop neighbor, it first performs maximum merging on the degree and then calculates the score.

[0116] The maximum value merging process merges the degrees of nodes that are one-hop neighbors of u, treating two nodes as one node, and the degree value is equal to the maximum degree of the two one-hop neighbor nodes.

[0117] DGR merge = MAX(DGR of each node that is a one-hop neighbor)

[0118] Assuming the set of neighboring nodes after merging the maximum values ​​is M, the score of node u is calculated using the following formula:

[0119]

[0120] In the formula:

[0121] SCORE u This represents the score of node u.

[0122] Furthermore, the traffic count is incremented by 1 each time data is sent, received, or forwarded.

[0123] Step 3: The server traverses all nodes at set time intervals, visits nodes in a depth-first manner, collects key performance scores and traffic counts, and constructs a score set and a traffic set.

[0124] Preferably, step 3 specifically includes:

[0125] Step 3.1: The server sends a traversal statistics frame, which includes three parts: statistical values, key performance score set, and traffic count set. The statistical values ​​are the set target number of key nodes, and the key performance score set and traffic count set are initially empty.

[0126] Step 3.2: The server accesses each node sequentially. After receiving the traversal statistics frame, each node determines whether the number of nodes recorded in the key performance score set and the traffic count set is less than the statistical value. If the key performance score set is not full, the data pair consisting of the current node identifier and the key performance score is added to the key performance score set. If the traffic count set is not full, the data pair consisting of the current node identifier and the traffic count value is added to the traffic count set.

[0127] Step 3.3: After traversing the statistics frame and completing all node visits, backtrack to the server level by level according to the original traversal path. When the intermediate nodes along the statistics frame receive the backtracking frame, they update the copies of the key performance score set and traffic count set stored locally on the node, which will be used as the initial reference in the next round of traversal.

[0128] Specifically, the server periodically sends traversal statistics frames at regular intervals. The traversal can employ various methods such as depth-first search or breadth-first search. The key is to add forward and backward mechanisms during the traversal process. Forward refers to the process of visiting the child nodes after visiting the parent node, while backward refers to the process of sending feedback information to the child node's parent node after visiting the child node. The initial traversal statistics frame sent by the server contains the statistical value K, the traffic set L (initially empty), and the rating set P (initially empty).

[0129] When a child node u visited in the forward traversal direction receives a traversal statistics frame, it will compare and count the number of tuples in the traffic set L and the rating set P. If the number of tuples in the traffic set L is less than the statistical value K, then node u's ID and traffic data will be combined into a tuple and added to the traffic set L; if the number of tuples in the rating set P is less than the statistical value K, then node u's ID and rating data will be combined into a tuple and added to the rating set P.

[0130] Step 4: Update the score set and traffic set based on the key performance score and traffic count value uploaded by the node. When the set is full, replace the node data with the smallest key performance score value in the set. When there are multiple nodes with the same score, replace the node data with the smallest traffic count value.

[0131] Preferably, step 4 specifically includes:

[0132] Step 4.1: Compare the current node's critical performance score with the critical performance scores of each data pair in the critical performance score set. If the current critical performance score is larger, replace the data pair with the smallest critical performance score in the critical performance score set.

[0133] Step 4.2: If there are multiple data pairs with the same key performance score, continue to compare the corresponding traffic count values ​​and replace the data pairs with smaller traffic count values ​​with the data pairs with larger traffic count values.

[0134] Step 4.3: Perform the same logic on the traffic count set. If the current traffic count value is greater than the traffic count value of a certain data pair in the set, then replace it with the data pair consisting of the current node identifier and the traffic count value.

[0135] Specifically, when a child node u visited in the forward traversal direction receives a traversal statistics frame, it will compare and count the number of tuples in the traffic set L and the rating set P. If the number of tuples in the traffic set L is greater than or equal to the statistical value K, then the traffic of each tuple in the traffic set L is compared with the traffic of node u. If the traffic of node u is greater than the traffic value in a certain tuple in the traffic set L, then the original tuple in the traffic set L is replaced with a tuple consisting of node u's ID and traffic. If the number of tuples in the rating set P is greater than or equal to the statistical value K, then the rating of each tuple in the rating set P is compared with the rating of node u. If the rating of node u is greater than the rating value in a certain tuple in the rating set P, then the original tuple in the rating set P is replaced with a tuple consisting of node u's ID and rating.

[0136] Step 5: When the maximum traffic count value in the traffic set exceeds the set threshold ratio, the server broadcasts a data shrinking frame.

[0137] Preferably, step 5 specifically includes:

[0138] Step 5.1: After completing one node traversal and statistics, the server extracts the traffic count values ​​of all nodes in the current traffic set and calculates the maximum value among them;

[0139] Step 5.2: If the maximum traffic count exceeds the set percentage threshold of the maximum allowed value of stored data, the server broadcasts a data contraction frame to all nodes in a traversal manner, triggering each node to perform traffic compression processing.

[0140] Specifically, after the server completes one round of traversal and constructs a traffic count set L, the server extracts the traffic count values ​​of all nodes from the traffic count set L and counts the maximum traffic value. If this maximum value exceeds the set compression threshold ratio, that is, it is greater than the product of the maximum acceptable traffic count value of the node and the preset ratio factor θ, then the server determines that there are high-frequency communication nodes in the current network, and there is a risk of congestion.

[0141] Step 6: Based on the received data shrink frame, the node performs compression processing on the traffic count value and uploads the compressed count result in the next traversal.

[0142] Preferably, but not limitingly, step 5 specifically includes:

[0143] After receiving the data shrinking frame broadcast by the server, the current node performs a square root operation on the current traffic count value and replaces the original traffic count value of the current node with the result of the operation; during the next traversal access of the server, it uploads the compressed traffic count value.

[0144] Specifically, when the server finds that a traversal and statistics have been completed, it will check the tuples in the traffic set L. If the largest traffic value in the tuple exceeds a certain percentage (e.g., 80%) of the maximum allowed value of stored data, the server will send a data shrinking frame in a traversal manner. The node that receives the data shrinking frame will take the square root of the traffic data.

[0145] Step 7: After the traversal is completed, the server marks key nodes based on the score set and traffic set, thus completing the distributed detection of key nodes.

[0146] Preferably, but not limitingly, step 7 specifically includes:

[0147] Step 7.1: Traverse the statistical frames. After falling back to the server, the server reads the node identifiers in the score set and the traffic set respectively to confirm all nodes contained in the set; and marks all nodes in the score set and the traffic set as key nodes.

[0148] Step 7.2: After the detection process is completed, a new round of node information construction and traversal process is started to form a distributed detection mechanism for key nodes that is executed in a loop.

[0149] Specifically, when a traversal and statistics are completed, the traversal and statistics frame will fall back to the server. The server will count the tuples in the traffic set L and the score set P, and then mark the corresponding nodes of ID in the traffic set L and the score set P as key nodes. Administrators can query the specific information of key nodes on the server.

[0150] Example 3 of the present invention provides a distributed detection system for key nodes in a wireless multi-hop network, which runs the distributed detection method for key nodes in a wireless multi-hop network as described in Example 1 or 2, including a server and a distributed single-node statistical processing logic module.

[0151] The distributed single-node statistical processing logic module broadcasts neighbor collection frames and neighborhood collection frames to the outside world. Nodes with one-hop connections reply with frame data, thereby calculating the degree of each node in a distributed manner. Based on the degree value of each node and the degree values ​​of all nodes connected to it in one hop, the module calculates the key performance score of each node and counts the traffic count value of each node.

[0152] The server is used to traverse each node, iteratively compare the key performance scores and traffic counts of each node, and select the top-ranked nodes by key performance scores and traffic counts as key nodes.

[0153] Compared with existing technologies, this invention constructs adjacency information through local communication between nodes and improves the representativeness and accuracy of key node identification by combining structural features and traffic information in a joint evaluation method. It employs a server-based traversal collection mechanism to construct key performance score sets and traffic sets, and dynamically updates the set content based on replacement rules, ensuring the continuity and adaptability of the detection results. By combining real-time data traffic statistics and threshold judgment, it triggers data contraction control operations to alleviate data transmission pressure on high-frequency nodes during network operation, avoiding connection failures caused by local congestion. Finally, through periodic traversal and judgment, it constructs a cyclically executable distributed key node detection process that does not rely on centralized control, achieving low-overhead, high-stability dynamic monitoring capabilities for key nodes, suitable for wireless multi-hop network scenarios with frequent structural changes.

[0154] This invention can quickly and accurately identify key nodes in a wireless multi-hop network. Its identification method considers both local structural scoring and traffic data of nodes, resulting in a more comprehensive evaluation. Because the structural scoring uses a distributed method, it avoids the process of acquiring the topology of the entire wireless multi-hop network. The overall operation is efficient and simple, with wide applicability; it only requires a full network traversal and statistical analysis at regular intervals, significantly reducing the communication overhead for detecting key nodes throughout the network.

[0155] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A distributed detection method for key nodes in a wireless multi-hop network, characterized in that, This includes statistical processing logic for individual nodes and statistical processing logic for the entire network, as well as steps for detecting key nodes; The statistical processing logic for a single node includes: each node broadcasts neighbor collection frames and neighborhood collection frames to the outside world; nodes with one-hop connections reply with frame data; and the degree of each node is calculated in a distributed manner. u The degree is equal to the degree after deduplication. u The critical performance score of a single node is calculated based on the number of neighboring nodes and the degree values ​​of all nodes connected to it via its one-hop. This includes: after each node calculates its own degree value, it broadcasts a degree collection frame and counts the degree collection frame responses from neighboring nodes; the degree values ​​are first merged by maximizing the maximum value; the degree values ​​of all nodes in the neighboring node set after maximizing the maximum value are subtracted by 1 and then summed, and the sum is used as the denominator after adding 1. The square of the node's degree value is used as the denominator, and the ratio of the two is the critical performance score of the single node; and the traffic count value of each node is also counted. The overall network statistical processing logic, and the detection of key nodes, includes: the server traverses each node, iteratively compares the key performance scores and traffic counts of each node, and selects the top-ranked nodes in terms of key performance scores and traffic counts as key nodes.

2. The distributed detection method for key nodes in a wireless multi-hop network according to claim 1, characterized in that: Each node broadcasts neighbor collection frames and neighborhood collection frames to the outside world. Nodes with one-hop connections reply with frame data, thereby distributively calculating the degree value of each node, including: In a wireless multi-hop network, each node periodically broadcasts a neighbor collection frame, and the node that receives the neighbor collection frame replies to the sending node. ID and receiving node ID Collect frame responses from neighbors to form a list containing neighbor nodes. ID gather; Broadcast neighbor collection frames externally, collect responses to neighbor collection frames, and form a list containing the sending nodes. ID Receiver node ID and its neighboring nodes ID Neighboring nodes ID gather; Each node collects neighboring nodes ID The set is deduplicated, and the degree of each node is calculated.

3. The distributed detection method for key nodes in a wireless multi-hop network according to claim 1, characterized in that: The traffic count values ​​for each node include: Whenever data is sent, received, transmitted, or forwarded on a node, the node's traffic count is incremented by 1. Whenever a node receives a data shrinking frame, it performs a square root operation on the node's traffic count value and updates the previous traffic count value.

4. The distributed detection method for key nodes in a wireless multi-hop network according to claim 1, characterized in that: The server traverses all nodes at set time intervals, using a depth-first approach to access nodes, collecting key performance scores and traffic counts, and constructing a score set and a traffic set, including: The server sends a traversal statistics frame, which includes three parts: statistical values, a set of key performance scores, and a set of traffic counts. The statistical values ​​are the set target number of key nodes, and the set of key performance scores and the set of traffic counts are initially empty. The server accesses each node sequentially. After receiving the traversal statistics frame, each node checks whether the number of nodes recorded in the key performance score set and the traffic count set is less than the statistical value. If the key performance score set is not full, the data pair consisting of the current node identifier and the key performance score is added to the key performance score set. If the traffic count set is not full, the data pair consisting of the current node identifier and the traffic count value is added to the traffic count set. After traversing the statistical frames and completing all node visits, the system backtracks to the server level by level along the original traversal path. When receiving the backtracking frames, the intermediate nodes along the statistical frames update the copies of the key performance score set and traffic count set stored locally on the node, which will be used as the initial reference in the next round of traversal.

5. The distributed detection method for key nodes in a wireless multi-hop network according to claim 4, characterized in that: Update the score set and traffic set based on the key performance scores and traffic counts uploaded by the nodes. When the set is full, replace the data of the node with the smallest key performance score in the set. When multiple nodes have the same score, replace the data of the node with the smallest traffic count, including: Compare the current node’s critical performance score with the critical performance scores of each data pair in the critical performance score set. If the current critical performance score is larger, replace the data pair with the smallest critical performance score in the critical performance score set. If multiple pairs of data have the same key performance score, continue to compare the corresponding traffic count values ​​and replace the pairs with smaller traffic count values ​​with the pairs with larger traffic count values. The same logic applies to the traffic count set. If the current traffic count value is greater than the traffic count value of a certain data pair in the set, then the data pair consisting of the current node identifier and the traffic count value is used to replace it.

6. The distributed detection method for key nodes in a wireless multi-hop network according to claim 5, characterized in that: When the maximum traffic count in the traffic set exceeds a set threshold ratio, the server broadcasts a data shrinking frame, including: After completing a node traversal and statistics, the server extracts the traffic count values ​​of all nodes in the current traffic set and calculates the maximum value among them; If the maximum traffic count exceeds the set percentage threshold of the maximum allowed value of stored data, the server will broadcast a data contraction frame to all nodes in a traversal manner, triggering each node to perform traffic compression processing.

7. The distributed detection method for key nodes in a wireless multi-hop network according to claim 6, characterized in that: The node performs compression processing on the traffic count value, including: After receiving the data shrinking frame broadcast by the server, the current node performs a square root operation on the current traffic count value and replaces the original traffic count value of the current node with the result of the operation; during the next traversal access of the server, it uploads the compressed traffic count value.

8. The distributed detection method for key nodes in a wireless multi-hop network according to claim 7, characterized in that: After the traversal is complete, the server marks key nodes based on the score set and traffic set, completing the distributed detection of key nodes, including: After the statistical frames are traversed and fall back to the server, the server reads the node identifiers in the score set and the traffic set respectively to confirm all the nodes contained in the set; and marks all the nodes in the score set and the traffic set as key nodes. After the detection process is completed, a new round of node information construction and traversal process is started, forming a distributed detection mechanism for key nodes that is executed in a loop.

9. A distributed detection system for key nodes in a wireless multi-hop network, comprising the distributed detection method for key nodes in a wireless multi-hop network as described in any one of claims 1-8, characterized in that, Includes server and distributed single-node statistical processing logic modules; The distributed, single-node statistical processing logic module broadcasts neighbor collection frames and neighborhood collection frames. Nodes with one-hop connections reply with frame data, thereby calculating the degree of each node in a distributed manner. u The degree is equal to the degree after deduplication. u The critical performance score of a single node is calculated based on the number of neighboring nodes and the degree values ​​of all nodes connected to it via its one-hop. This includes: after each node calculates its own degree value, it broadcasts a degree collection frame and counts the degree collection frame responses from neighboring nodes; the degree values ​​are first merged by maximizing the maximum value; the degree values ​​of all nodes in the neighboring node set after maximizing the maximum value are subtracted by 1 and then summed, and the sum is used as the denominator after adding 1. The square of the node's degree value is used as the denominator, and the ratio of the two is the critical performance score of the single node; and the traffic count value of each node is also counted. The server is used to traverse each node, iteratively compare the key performance scores and traffic counts of each node, and select the top-ranked nodes by key performance scores and traffic counts as key nodes.

Images