A distributed clustering method, device and medium on wireless sensor network

Abstract

The present invention discloses a distributed clustering method, device, and medium for wireless sensor networks. The method comprises: each sensor uses NP optimal data from its own database to train a local proxy model; synthesizes a global proxy model through global model fusion; uses the global proxy model to replace true evaluation for evolutionary optimization, and sends the found candidate solutions to unterminated sensors; the unterminated sensors receive the candidate solutions, and if the candidate solutions are within the sensor's sensing range, the fitness of the candidate solutions is truly evaluated using a clustering evaluation criterion, and the database and local proxy models are updated; new offspring are obtained from each sensor, and the local model is used to assist in the evaluation of the new offspring individuals and the historical optimal particles; the new offspring individuals whose predicted values ​​are better than the predicted values ​​of the historical optimal particles are truly evaluated, and the local database is updated. The present invention protects the privacy of clustered data on sensors and can be widely applied in the field of wireless network technology.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless network, and in particular to a distributed clustering method, device and medium on a wireless sensor network. Background Art

[0002] Wireless sensor networks (WSNs) are a group of sensors that collaboratively sense, detect, and collect information about objects in various environments. After collaborative information processing, they transmit useful information to users. Distributed clustering of wireless sensor data is widely used in fields such as environmental protection, healthcare, aviation, military, and industry. Data collected by a single sensor is distributed, insecure, prone to errors, and localized. Clustering using only data from a single sensor can result in significant errors due to the sensor's limited sensing area. Centrally transmitting data from multiple sensors to a central server can lead to data privacy leaks, limited communication bandwidth, and high latency. With the recent development of the Internet of Things and information technology, network security issues have received increasing attention. Collaborative distributed clustering among wireless sensor networks belonging to different social organizations can yield more accurate and effective results. However, in distributed clustering scenarios, due to national policies and commercial competition, the original cluster objects on wireless sensors must be protected and cannot be freely transmitted to maintain data privacy.

[0003] Existing algorithms for solving distributed clustering problems in distributed wireless sensor networks include distributed k-Means and its improved versions, as well as distributed firefly flame algorithms. These algorithms often require transmitting raw clustering data and their respective cluster centers to neighboring sensors, a process that carries the risk of data leakage. Summary of the Invention

[0004] In order to solve at least one of the technical problems existing in the prior art to a certain extent, the present invention aims to provide a distributed clustering method, device and medium on a wireless sensor network.

[0005] The technical solution adopted in the present invention is:

[0006] A distributed clustering method on a wireless sensor network comprises the following steps:

[0007] S1, regularize the distributed clustering data of each sensor;

[0008] S2. Initialize the first generation of particle swarm on each sensor: Generate a local population with NP particles, evaluate each individual in the local population, and record these individuals and their evaluation values ​​as data items in the database HD i ;

[0009] S3. Each sensor uses its own database HD i NP optimal data in training local agent model After training, the local agent model Send to the central server;

[0010] S4. The central server receives the local proxy model Finally, a global proxy model is synthesized through global model fusion;

[0011] S5. The central server uses the global proxy model to replace the real evaluation to perform evolutionary optimization to find the candidate solution x p , the candidate solution x found p Issued to unterminated sensors;

[0012] S6. The unterminated sensor receives the candidate solution x p , if the candidate solution x p Within the sensor’s sensing range, the candidate solution x is truly evaluated using the clustering evaluation criterion p The fitness value of and update the database HD i and local proxy models

[0013] S7, get the new offspring of each sensor and use the local model Assisted evaluation of new offspring individuals x j (j=1,2,…,NP) and the best particle pbestX in history j (j=1,2,…,NP);

[0014] S8, the predicted value Better than the best particle prediction value in history The new offspring individual x j Conduct real-world assessments and update the local database HD i ;

[0015] S9. If the detection reaches the termination condition, then reverse normalize the optimal solution and output the optimal solution and its corresponding clustering result; otherwise, return and continue to execute step S3.

[0016] Furthermore, the step S1 specifically includes:

[0017] Regularize the distributed clustering data of each sensor, and the clustering object vector of the i-th sensor is recorded as where Z i is the number of cluster objects, each data With D-dimensional variables, by communicating the upper and lower bounds of the clustering data of all sensors and then regularizing the data, the influence of each feature dimension on the clustering results can be balanced.

[0018] Furthermore, the step S2 specifically includes:

[0019] First, a local population of NP particles is generated, each particle Contains the centers of K clusters According to particle M i,l The cluster center refers to the cluster object to the cluster where the cluster center closest to it is located; each cluster center has D-dimensional variables, so M i,l With KD-dimensional variables;

[0020] Then, the clustering evaluation criteria based on local data is used as the optimization objective function to evaluate each individual, and these individuals and their evaluation values ​​are recorded as data items in the database HD i .

[0021] Furthermore, the objective function is expressed as follows:

[0022]

[0023] Where K is the number of cluster centers, is particle M I,l represents the cluster center of the kth cluster, Represents a clustering object Belong to the particle M i,l The kth cluster of the clustering results obtained is, Refers to the Euclidean distance between a cluster object and the center of the cluster to which it belongs.

[0024] Furthermore, the expression of the global proxy model in step S4 is as follows:

[0025]

[0026] Where S is the number of sensors, w i (x) is the weight.

[0027] Furthermore, the weight w i The expression of (x) is as follows:

[0028]

[0029] Weight w i (x) and the number of sensors that can sense the location of the candidate solution vector x Related, where t i (x) represents whether sensor i can perceive candidate solution x; t i (x) = 0 means that the sensing range of sensor i fails to cover the candidate solution x, and t i(x) = 1 means that the sensing range of sensor i can cover the candidate solution x; e i (i=1,2,…,S) is the root mean square error of the prediction of the local surrogate model on the perceptron for all known data in its local database.

[0030] Furthermore, in the step of performing evolutionary optimization in step S5, the evolutionary optimizer used is a social learning particle swarm optimization algorithm.

[0031] Furthermore, the step of obtaining the new offspring of each sensor in step S7 specifically includes:

[0032] The new offspring of each sensor is obtained through particle swarm optimization;

[0033] Among them, the speed update formula of particle swarm optimization is as follows:

[0034] V=χ·(V+c1·r1·(pbestX-X)+c2·r2·(gbestX-X))

[0035] Where X is the position matrix of the population, which consists of the position vectors of NP particles, V is the velocity matrix, pbestX is the historical optimal position matrix of each particle, gbestX is the position vector of the optimal particle so far; c1 and c2 are both acceleration coefficients; χ is the contraction factor; r1 and r2 are two random real numbers between 0 and 1.

[0036] Another technical solution adopted in the present invention is:

[0037] A distributed clustering device on a wireless sensor network, comprising:

[0038] at least one processor;

[0039] at least one memory for storing at least one program;

[0040] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0041] Another technical solution adopted in the present invention is:

[0042] A computer-readable storage medium stores a program executable by a processor, wherein the program executable by the processor is used to perform the method described above when executed by the processor.

[0043] The beneficial effects of the present invention are as follows: the present invention adopts the model in machine learning as the proxy model, and uses distributed integration technology to fuse multiple proxy models, so that the algorithm can use the scattered data distributed on the wireless sensor network for clustering, and protect the privacy of the original clustering data while performing clustering efficiently and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following introduction is made to the drawings of the embodiments of the present invention or the related technical solutions in the prior art. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative work.

[0045] Figure 1 This is a flow chart of a distributed clustering method on a wireless sensor network based on a distributed agent model assisted evolutionary optimization algorithm in an embodiment of the present invention;

[0046] Figure 2 4 is a diagram illustrating an example of distributed clustering objects on a sensor in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The embodiments of the present invention are described in detail below, examples of which 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. The step numbers in the following embodiments are provided for ease of explanation only and do not limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0048] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, they cannot be understood as limitations on the present invention.

[0049] In the description of the present invention, "several" means one or more, "many" means more than two, "greater than," "less than," and "exceed" are understood to exclude the number itself, while "above," "below," and "within" are understood to include the number itself. The use of "first" and "second" in the description is solely for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance, implicitly specifying the number of the indicated technical features, or implicitly specifying the order of the indicated technical features.

[0050] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, and connecting should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.

[0051] Initial clustering is performed on local sensors, and a corresponding radial basis function network (RBF) model is trained based on the clustering criterion to serve as a proxy model for transmitting sensor information, replacing the actual clustering objects. This effectively protects the privacy of sensor data. A global proxy model is generated by performing a weighted aggregation of the local proxy models from multiple sensors on a central server. This improves the accuracy of clustering results using a proxy-model-assisted evolutionary optimization algorithm. Therefore, a distributed proxy-model-assisted evolutionary optimization algorithm is well-suited for solving the problem of distributed clustering of data on wireless sensor networks.

[0052] like Figure 1 As shown, this embodiment provides a distributed clustering method on wireless sensor networks based on a distributed agent model-assisted evolutionary optimization algorithm. This method uses a radial basis function neural network in machine learning as a proxy model to replace the transmission of clustering object data and assist in the evaluation of the evolutionary algorithm. Distributed integration technology is used to fuse multiple proxy models, enabling the algorithm to cluster dispersed data distributed across the wireless sensor network. This method performs clustering efficiently and accurately while protecting the privacy of the original clustered data. Specifically, the method includes the following steps:

[0053] S101. Regularize the distributed clustering data of each sensor. The regularized data can balance the influence of each feature dimension on the clustering result.

[0054] Regularize the distributed clustering data of each sensor, and the clustering object vector of sensor i is recorded as where Z i is the number of cluster objects, each data With D-dimensional variables, by communicating the upper and lower bounds of the clustering data of all sensors and then regularizing the data, the influence of each feature dimension on the clustering results can be balanced.

[0055] S102, initialize the first generation of particle swarms on each sensor; specifically, first generate a local population with NP particles, then evaluate each individual based on the clustering evaluation criterion of local data as the optimization objective function, and record these individuals and their evaluation values ​​as data items in the database HD i .

[0056] Initialize the first generation of particle swarms on each sensor; specifically, first generate a local population with NP particles. Contains the centers of K clusters According to particle Mi,l The cluster center can divide the cluster objects into the clusters of the cluster centers closest to them. Each cluster center has D-dimensional variables, so M i,l With KD dimension variables. Then the clustering evaluation criteria based on local data are used as the optimization objective function to evaluate each individual, and these individuals and their evaluation values ​​are recorded as data items in the database HD i .

[0057] The objective function used as the real evaluation is the clustering evaluation criterion based on local data, which can be expressed as follows:

[0058]

[0059] Where K is the number of cluster centers, is particle M I,l represents the cluster center of the kth cluster, Represents a clustering object Belong to the particle M i,l The kth cluster of the clustering results obtained is, Refers to the Euclidean distance between a cluster object and the center of the cluster to which it belongs.

[0060] S103, each sensor uses its own database HD i NP optimal data in training local surrogate models And send it to the central server.

[0061] S104: The central server receives the local proxy model Finally, a global proxy model is synthesized through global model fusion.

[0062] The process of global model fusion is composed of the following formulas. First, the prediction formula of the local agent model is used The weighted sum is used to obtain the prediction function of the global surrogate model.

[0063]

[0064] Where S is the number of sensors, weight w i (x) and the number of sensors that can sense the location of the candidate solution vector x And the root mean square error of the prediction of the local proxy model on the perceptron for all known data in its local database is e i (i=1,2,…,S) related. Where t i (x) indicates whether sensor i can perceive candidate solution x. i (x) = 0 means that the sensing range of sensor i fails to cover the candidate solution x, and t i(x) = 1 means that the sensor i can sense the range that can cover the candidate solution x. The weight coefficient w after all the cases are combined i (x) can be represented by the following piecewise function:

[0065]

[0066] When Tsum=0, the prediction function of the local model of each sensor i has a weight w i (x) is the same and is 1 / S. When Tsum=1, the weight w i (x) is equal to t i (x), if the sensing range of sensor i fails to cover the candidate solution x, w i (x) = t i (x) = 0, otherwise sensor i can sense candidate solution x and w i (x) = t i (x)=1.

[0067] S105, the central server uses the global proxy model to replace the real evaluation to perform evolutionary optimization to find promising candidate solutions x p , and find the candidate solution x p Sent to the unterminated sensors; the evolutionary optimizer used is the social learning particle swarm optimization algorithm (SL-PSO).

[0068] S106: The unterminated sensor receives the candidate solution x p If the candidate solution is within the sensor's sensing range, the candidate solution x is truly evaluated using the clustering evaluation criterion. p The fitness value of the local database HD is updated i and local proxy models

[0069] S107, each sensor obtains a new offspring through particle swarm optimization, and uses the local model to assist in evaluating the offspring individual x j (j=1,2,…,NP) and the historically optimal particle pbestX j (j=1,2,…,NP).

[0070] Among them, the speed update formula of particle swarm optimization is as follows:

[0071] V=χ·(V+c1·r1·(pbestX-X)+c2·r2·(gbestX-X))

[0072] Where X is the position matrix of the population consisting of NP particles x j(j=1,2,…,NP), V is the velocity matrix, pbestX is the historically optimal position matrix for each particle, and gbestX is the position vector of the currently optimal particle. c1 and c2 are acceleration coefficients, χ is the contraction factor, and r1 and r2 are two random real numbers between 0 and 1. The position update formula is as follows:

[0073] X=V+X

[0074] S108, the predicted value Better than the best particle prediction value in history The offspring individual x j Conduct real-world assessments and update local database HD i .

[0075] S109: If the termination condition is met, the optimal solution is normalized and the optimal solution and its corresponding clustering result are output; otherwise, the process jumps back to step S103 of training the local proxy model.

[0076] The above method is explained in detail below with reference to the accompanying drawings and specific embodiments.

[0077] See also Figure 1 This embodiment provides a distributed clustering method on a wireless sensor network based on a distributed agent model assisted evolutionary optimization algorithm, comprising the following steps:

[0078] S201, the clustering objects of sensors 1 to 3 are respectively as follows Figure 2 (a), Figure 2 (b), Figure 2 (c) As shown in the figure, the three sensors respectively regularize their own distributed clustering data, and the clustering object vector of sensor i is recorded as where Z i is the number of cluster objects, Figure 2 It can be seen that Z i =25, each data With two-dimensional variables, by communicating the upper and lower bounds of the clustering data of all sensors and then regularizing the data, the influence of each feature dimension on the clustering results can be balanced.

[0079] S202, initialize the first generation of particle swarm on each sensor; specifically, first generate a local population with NP particles, where NP=50. Contains the centers of 5 clusters According to particle M i,l The cluster center can divide the cluster objects into the clusters of the cluster centers closest to them. Each cluster center has a 2-dimensional variable, so M i,lWith 10-dimensional variables. Then the cluster evaluation criteria based on local data are used as the optimization objective function to evaluate each individual, and these individuals and their evaluation values ​​are recorded as data items in the database HD i .

[0080] The objective function used as the real evaluation is the clustering evaluation criterion based on local data, which can be expressed as follows:

[0081]

[0082] Where K is the number of cluster centers, is particle M I,l represents the cluster center of the kth cluster, Represents the cluster object o i z Belong to the particle M i,l The kth cluster of the clustering results obtained is, Refers to the Euclidean distance between a cluster object and the center of the cluster to which it belongs.

[0083] S203, each sensor uses its own database HD i The 50 best data in training local surrogate models And send it to the central server.

[0084] S204: The central server receives the local proxy model Finally, a global proxy model is synthesized through global model fusion.

[0085] The process of global model fusion is composed of the following formulas. First, the prediction formula of the local agent model is used The weighted sum is used to obtain the prediction function of the global surrogate model.

[0086]

[0087] Where S is the number of sensors, weight w i (x) and the number of sensors that can sense the location of the candidate solution vector x And the root mean square error of the prediction of the local proxy model on the perceptron for all known data in its local database is e i (i=1,2,3) related. i (x) indicates whether sensor i can perceive candidate solution x. i (x) = 0 means that the sensing range of sensor i fails to cover the candidate solution x, and t i (x)=1 indicates that the sensing range of sensor i can cover the candidate solution x. Figure 2 In this case, the range that sensor i can sense is always the same, so the weight coefficient w i(x)=1 / 3.

[0088] S205, the central server uses the global proxy model to replace the real evaluation to perform evolutionary optimization to find a promising candidate solution x p , and find the candidate solution x p Sent to the unterminated sensors; the evolutionary optimizer used is the social learning particle swarm optimization algorithm (SL-PSO).

[0089] S206: The unterminated sensor receives the candidate solution x p If the candidate solution is within the sensor's sensing range, the candidate solution x is truly evaluated using the clustering evaluation criterion. p The fitness value of the local database HD is updated i and local proxy models

[0090] S207, each sensor obtains a new offspring through particle swarm optimization, and uses a local model to assist in evaluating the offspring individuals and the historical optimal particles. The speed update formula of particle swarm optimization is as follows:

[0091] V=χ·(V+c1·r1·(pbestX-X)+c2·r2·(gbestX-X))

[0092] Where X is the position matrix of the population consisting of NP particles x j (j=1,2,…,50), V is the velocity matrix, pbestX is the historically optimal position matrix for each particle, and gbestX is the position vector of the currently optimal particle. c1 and c2 are acceleration coefficients set to 2.05, χ is the shrinkage factor set to 0.7298, and r1 and r2 are two random real numbers between 0 and 1. The position update formula is as follows:

[0093] X=V+X

[0094] S208, the predicted value Better than the best particle prediction value in history The offspring individual x j Conduct real-world assessments and update local database HD i .

[0095] S209: If the termination condition is met, the optimal solution gbestX is normalized to obtain the coordinates of the five cluster centers. Based on these cluster center coordinates, the cluster objects are divided into the clusters with the cluster centers closest to them and the clustering results are output. The expected clustering results are as follows: Figure 2 As shown in (d), the same graphic representation belongs to the same cluster, otherwise jump back to step S203 of training the local proxy model.

[0096] In summary, this embodiment has the following advantages and beneficial effects compared to the prior art:

[0097] (1) The method of this embodiment protects the privacy of clustered data on sensors and replaces the transmission of original data with the transmission of proxy models.

[0098] (2) The distributed clustering results obtained in this embodiment are better than the existing methods using two clustering evaluation indicators.

[0099] This embodiment also provides a distributed clustering device on a wireless sensor network, including:

[0100] at least one processor;

[0101] at least one memory for storing at least one program;

[0102] When the at least one program is executed by the at least one processor, the at least one processor implements Figure 1 The method shown.

[0103] A distributed clustering device on a wireless sensor network in this embodiment can execute the method provided by the embodiment of the present invention. Figure 1 The method shown can execute any combination of implementation steps of the method embodiment and have the corresponding functions and beneficial effects of the method.

[0104] The present application also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs Figure 1 The method shown.

[0105] This embodiment also provides a storage medium storing instructions or programs that can execute a distributed clustering method on a wireless sensor network provided by an embodiment of the method of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the corresponding functions and beneficial effects of the method can be obtained.

[0106] In some optional embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided in an exemplary manner for the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operation and logic flow presented herein. Optional embodiments are contemplated in which the order of the various operations is changed and the sub-operations described as a part of a larger operation are performed independently.

[0107] Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise indicated, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It will also be understood that a detailed discussion of the actual implementation of each module is not necessary for understanding the present invention. More specifically, given the properties, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the module will be understood within the ordinary skill of an engineer. Therefore, a person skilled in the art using ordinary skill will be able to implement the present invention set forth in the claims without undue experimentation. It will also be understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0108] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0109] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0110] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0111] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0112] In the above description of this specification, reference to the terms "one embodiment / example," "another embodiment / example," or "certain embodiments / examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0113] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

[0114] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A distributed clustering method on a wireless sensor network, characterized in that: The following steps are involved: S1, regularize the distributed clustering data of each sensor; S2. Initialize the first generation of particle swarm on each sensor: Generate a local population with NP particles, evaluate each individual in the local population, and record these individuals and their evaluation values ​​as data items in the database HD i ; S3. Each sensor uses its own database HD i NP optimal data in training local agent model After training, the local agent model Send to the central server; S4. The central server receives the local proxy model Finally, a global proxy model is synthesized through global model fusion; S5. The central server uses the global proxy model to replace the real evaluation to perform evolutionary optimization to find the candidate solution x p , the candidate solution x found p Issued to unterminated sensors; S6. The unterminated sensor receives the candidate solution x p , if the candidate solution x p Within the sensor’s sensing range, the candidate solution x is truly evaluated using the clustering evaluation criterion p The fitness value of and update the database HD i and local proxy models S7, get the new offspring of each sensor and use the local model Assisted evaluation of new offspring individuals x j And the historical best particle pbestX j ; S8, the predicted value Better than the best particle prediction value in history The new offspring individual x j Conduct real-world assessments and update the local database HD i ; S9. If the detection reaches the termination condition, then reverse normalize the optimal solution and output the optimal solution and its corresponding clustering result; otherwise, return to and continue to execute step S3; The step S2 specifically includes: First, a local population of NP particles is generated, each particle Contains the centers of K clusters According to particle M i,l The cluster center refers to the cluster object to the cluster where the cluster center closest to it is located; each cluster center has D-dimensional variables, so M i,l With KD-dimensional variables; Then, the clustering evaluation criteria based on local data is used as the optimization objective function to evaluate each individual, and these individuals and their evaluation values ​​are recorded as data items in the database HD i ; The objective function is expressed as follows: Where K is the number of cluster centers, is particle M I,l represents the cluster center of the kth cluster, Represents a clustering object Belong to the particle M i,l The kth cluster of the clustering results obtained is, Refers to the Euclidean distance between the cluster object and the center of the cluster to which it belongs; The expression of the global proxy model in step S4 is as follows: Where S is the number of sensors, w i (x) is the weight; Weight w i The expression of (x) is as follows: Weight w i (x) and the number of sensors that can sense the location of the candidate solution vector x Related, where t i (x) represents whether sensor i can perceive candidate solution x; t i (x) = 0 means that the sensing range of sensor i fails to cover the candidate solution x, and t i (x) = 1 means that the sensing range of sensor i can cover the candidate solution x; e i is the root mean square error of the prediction of the local surrogate model on the perceptron for all known data in its local database, i = 1, 2, …, S.

2. The distributed clustering method on a wireless sensor network according to claim 1, characterized in that: The step S1 specifically includes: Regularize the distributed clustering data of each sensor, and the clustering object vector of the i-th sensor is recorded as where Z i is the number of cluster objects, each data With D-dimensional variables, by communicating the upper and lower bounds of the clustering data of all sensors and then regularizing the data, the influence of each feature dimension on the clustering results can be balanced.

3. The distributed clustering method on a wireless sensor network according to claim 1, characterized in that: In the step of performing evolutionary optimization in step S5, the evolutionary optimizer used is a social learning particle swarm optimization algorithm.

4. The distributed clustering method on a wireless sensor network according to claim 1, characterized in that: The step of obtaining the new offspring of each sensor in step S7 specifically includes: The new offspring of each sensor is obtained through particle swarm optimization; Among them, the speed update formula of particle swarm optimization is as follows: V=χ·(V+c1·r1·(pbestX-X)+c2·r2·(gbestX-X)) Where X is the position matrix of the population, which consists of the position vectors of NP particles, V is the velocity matrix, pbestX is the historical optimal position matrix of each particle, gbestX is the position vector of the optimal particle so far; c1 and c2 are both acceleration coefficients; χ is the shrinkage factor; r1 and r2 are two random real numbers between 0 and 1.

5. A distributed clustering device on a wireless sensor network, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 4.

6. A computer-readable storage medium storing a program executable by a processor, characterized in that: The processor-executable program is used to perform the method according to any one of claims 1 to 4 when executed by the processor.

Images