CN113078958B - Network node distance vector synchronization method based on transfer learning - Google Patents

Network node distance vector synchronization method based on transfer learning Download PDF

Info

Publication number
CN113078958B
CN113078958B CN202110333523.8A CN202110333523A CN113078958B CN 113078958 B CN113078958 B CN 113078958B CN 202110333523 A CN202110333523 A CN 202110333523A CN 113078958 B CN113078958 B CN 113078958B
Authority
CN
China
Prior art keywords
network
node
data
distance vector
vgg
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110333523.8A
Other languages
Chinese (zh)
Other versions
CN113078958A (en
Inventor
高明生
潘羿航
李建
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hohai University HHU
Original Assignee
Hohai University HHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hohai University HHU filed Critical Hohai University HHU
Priority to CN202110333523.8A priority Critical patent/CN113078958B/en
Publication of CN113078958A publication Critical patent/CN113078958A/en
Application granted granted Critical
Publication of CN113078958B publication Critical patent/CN113078958B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B11/00—Transmission systems employing sonic, ultrasonic or infrasonic waves
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B13/00—Transmission systems characterised by the medium used for transmission, not provided for in groups H04B3/00 - H04B11/00
    • H04B13/02—Transmission systems in which the medium consists of the earth or a large mass of water thereon, e.g. earth telegraphy
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Algebra (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a network node distance vector synchronization method based on transfer learning, which comprises the following steps: s1: according to the node density of the network and the Eckman spiral ocean current model, determining node distance vector synchronization period data by utilizing computer simulation; s2: performing fusion processing on the obtained data, and outputting the data characteristics serving as a learning model; s3: according to the node density of the network and the Ackerman spiral ocean current model, further classifying the data, inputting the data as training data into a VGG + CNN network for training, and then performing fine adjustment and correction by using a transfer learning model; s4: obtaining an optimal distance vector synchronization period according to the node density and the ocean current condition of the current network and the corrected model; s5: and initiating node distance vector synchronization by the sink node according to the obtained optimal period. The invention has better network bandwidth utilization rate and energy efficiency, prolongs the life cycle of the underwater acoustic network, and is beneficial to long-time information collection and monitoring of the ocean target.

Description

Network node distance vector synchronization method based on transfer learning
Technical Field
The invention relates to a network node distance vector synchronization method based on transfer learning, and belongs to the field of underwater acoustic network data transmission.
Background
Due to the highly dynamic marine environment and the non-stable, asymmetric nature of the underwater acoustic channel, packet loss can easily occur in underwater acoustic networks. In order to improve the probability of the data packet delivery of the underwater acoustic network and improve the reliability of the route, the conventional underwater acoustic network adopts an opportunistic route which has higher protocol efficiency and is based on opportunistic transmission.
The working principle of opportunistic routing is as follows: when the relay node is selected, a plurality of adjacent nodes are selected according to certain indexes (such as energy, depth, hop count and the like), and are sequenced according to priority, and then the relay nodes forward data packets in sequence; once a certain relay node successfully forwards the data packet, other relay nodes stop forwarding, and bandwidth and energy waste caused by repeated forwarding is avoided; if the data packet is unsuccessful, the data packet is forwarded by the next relay node in sequence, and so on, thereby greatly increasing the probability of data packet delivery.
And an opportunistic routing protocol based on the hop count, wherein the hop count refers to the hop count from the sensor node to the sink (buoy) node. On the premise that hop count information of the nodes is accurate, the opportunistic routing based on hop counts can effectively solve the problem of common holes in the underwater acoustic network, and meanwhile, the opportunistic routing can improve the reliability of the routing.
Most of traditional node distance vector synchronization methods are that sink nodes periodically synchronize in a flooding mode at fixed time intervals, so that information implosion is easily generated, and bandwidth utilization rate and energy efficiency are low.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides a network node distance vector synchronization method based on transfer learning, which adopts the transfer learning to determine the network node distance vector synchronization period to carry out distance vector synchronization on network nodes under the new ocean current condition, thereby not only accelerating and optimizing the learning efficiency of a new model, but also ensuring the timeliness and the accuracy of the network node distance vectors, having good network bandwidth utilization rate and energy efficiency and prolonging the life cycle of a network.
The invention mainly adopts the technical scheme that:
a network node distance vector synchronization method based on transfer learning specifically comprises the following construction steps:
s1: simulating node densities and an Ackerman spiral ocean current model of different networks by using a computer simulation technology to obtain optimal node distance vector synchronization period data under different conditions;
s2: performing feature fusion on the data obtained in the step S1 by adopting a K-means clustering algorithm to obtain an original data set; extracting common features of the data, classifying the common features according to the features, and outputting the data features serving as a learning model;
s3: dividing the original data set obtained in the step S2 into training data and test data according to a certain proportion, inputting the training data into a VGG + CNN network for training, then performing fine tuning correction by using a transfer learning model, and testing the corrected VGG + CNN network by using the test data;
s4: obtaining an optimal distance vector synchronization period according to the node density and the ocean current condition of the current network and the corrected model;
s5: and initiating node distance vector synchronization by the sink node according to the obtained optimal period.
Preferably, the specific steps of S3 are as follows:
s31, taking the fusion data as training data to be input into the VGG + CNN network, and performing spatial transformation on the training data by using a feature extraction function of the VGG + CNN network, as shown in formula (1):
F(I)＝R (1)；
wherein F (-) represents a feature extraction function, I represents input training data, R represents output data of the VGG network, and R is equal to RA×BRepresenting a real number space, A representing the row number of a matrix, and B representing the column number of the matrix;
the feature extraction function F (-) transforms input training data into features of A multiplied by B dimension; the output of the feature extraction function is then dimension transformed using the full join function, as shown in equation (2):
D(R)＝W (2)；
d (-) represents a full-connection function, W represents output data of a full-connection network, the full-connection function D (-) converts A multiplied by B dimensional data into AB multiplied by 1 dimensional data, and a full-connection layer adopts a sigmoid activation function;
s32: performing feature extraction on the output data subjected to space and dimension transformation by using a convolution function of the VGG + CNN network, as shown in formula (3):
Figure BDA0002996394440000031
wherein Y (·) represents a convolution function, i represents a row index of the matrix, j represents a column index of the matrix, K represents a convolution kernel, R represents output data of the VGG network, m represents a row number of the matrix, and n represents a column number of the matrix;
s33: converging the data after feature extraction by using a pooling function of the VGG + CNN network, and performing fine adjustment and correction on the matched model by using a transfer learning algorithm, as shown in formula (4):
Figure BDA0002996394440000032
wherein Y represents the output data of the convolutional layer and is used as the input data of the pooling layer,
Figure BDA0002996394440000033
represents the output data of the pooling layer, and Ave (-) represents the averaging function.
Preferably, in S4, based on the node density of the network, under the condition of determining the maximum point-to-point transmission distance, based on a fixed medium access control protocol of a link layer, in combination with the current ackermann spiral ocean current situation, an optimal distance vector synchronization period is obtained through a modified model, which includes the specific steps of:
s41: inputting the node density and the Ackerman spiral ocean current model of the current network into the VGG + CNN network after training as test data, and calculating the node density of the input network and the probability S of each category of the Ackerman spiral ocean current by using the softmax classification layer of the VGG + CNN networkhThe calculation formula is shown as formula (5):
Figure BDA0002996394440000041
where h denotes the h-th neuron of the output layer, ShFor output, t represents the total number of output layer neurons;
s42: and sorting the magnitude values of the probabilities of the categories, and taking the category with the maximum probability value as an output classification result to finish the identification of the distance vector period.
Preferably, the specific steps of S5 are as follows:
s51: in the channel access process, the sink node broadcasts a distance vector update binary group (turn, distance) to adjacent nodes, wherein the turn represents the synchronization of the distance vector initiated by the sink node for the second time, and the distance represents the hop count of the node from the sink node;
s52: each node updates the own distance vector according to the distance vector information of the adjacent node obtained in the channel interception and channel access processes, and broadcasts the distance vector carried by the node in the channel access process with the adjacent node, and so on;
s53: and when all the nodes finish the distance vector updating, the distance vector synchronization initiated by the sink node in the current round is finished.
Has the advantages that: the invention provides a network node distance vector synchronization method based on transfer learning, which can adaptively realize optimal synchronization of node densities of different networks and network node distance vectors under different ocean current conditions, has better network bandwidth utilization rate and energy efficiency, prolongs the life cycle of an underwater acoustic network, and is favorable for long-time information collection and monitoring of ocean targets.
Drawings
Fig. 1 is a flow chart of the network node distance vector synchronization based on the transfer learning of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application are clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
A network node distance vector synchronization method based on transfer learning is disclosed, as shown in FIG. 1, and specifically comprises the following construction steps:
s1: simulating node densities and an Ekman Spiral sea (Ekman Spiral) flow model of different networks by using a computer simulation technology to obtain optimal node distance vector synchronization period data under different conditions; the method comprises the following specific steps: and determining the optimal period data of node distance vector synchronization by computer simulation under the condition of a certain communication distance according to the node density of the network and the Ekman Spiral ocean current condition. If the period is too long, the hop count information of the node distance sink node may be inaccurate in consideration of the dynamic property of the marine environment; if the period is too short, the information in the flooding mode is imploded, and bandwidth and energy are wasted.
S2: performing feature fusion on the data obtained in the step S1 by adopting a K-means clustering algorithm to obtain an original data set; extracting common features of the data, classifying the common features according to the features, and outputting a data feature set serving as a learning model;
s3: dividing the original data set obtained in the step S2 into training data and test data according to a certain proportion, inputting the training data into a VGG + CNN network for training, then performing fine tuning correction by using a transfer learning model, and testing the corrected VGG + CNN network by using the test data;
in the invention, the VGG + CNN network comprises a trained VGG16 network and a CNN network which are connected in sequence, wherein the VGG16 network is a trained primary feature extraction network, and the CNN network is a secondary feature extraction network. The CNN network comprises a full connection layer, a plurality of convolution layers, a pooling layer and a softmax classification layer. These two networks consist of the following quintuple functions: m ═ V, D, C, P, S; wherein V is a feature extraction function, D is a full-link function, C is a convolution function, P is a pooling function, and S is a softmax classification function; the feature extraction function V is completed through a VGG16 network, the VGG16 network comprises 13 convolution layers, 5 pooling layers and 3 full-connection layers, wherein the sizes of all convolution kernels are 3 multiplied by 3, all the pooling layers adopt 2 multiplied by 2 pooling kernels, and finally, the probability of each category is output through the 3 full-connection layers and the Softmax layer; in the VGG-CNN network model, the last pooling layer is cut off, and the full connection layer and the softmax layer behind the pooling layer are discarded.
The specific steps of S3 in the invention are as follows:
s31, dividing the original data set into training data and test data according to a certain proportion (such as 8: 2), taking the training data as the input of the VGG + CNN network, and performing spatial transformation on the training data by using the feature extraction function of the VGG + CNN network, as shown in formula (1):
F(I)＝R (1)；
wherein F (-) represents a feature extraction function, I represents input training data, R represents output data of the VGG network, and R is equal to RA×BRepresenting a real number space, A representing the row number of a matrix, and B representing the column number of the matrix;
the feature extraction function F (-) transforms input training data into features of A multiplied by B dimension; the output of the feature extraction function is then dimension transformed using the full join function, as shown in equation (2):
D(R)＝W (2)；
wherein D (-) represents a full connection function, W represents output data of a full connection network, the full connection function D (-) converts A × B dimensional data into AB × 1 dimensional data, and a full connection layer adopts a sigmoid activation function;
s32: performing feature extraction on the output data subjected to space and dimension transformation by using a convolution function of the VGG + CNN network, as shown in formula (3):
Figure BDA0002996394440000071
wherein Y (-) represents a convolution function, i represents a row index of the matrix, j represents a column index of the matrix, K represents a convolution kernel, R represents output data of the VGG network, m represents a row number of the matrix, and n represents a column number of the matrix;
s33: converging the data after feature extraction by using a pooling function of the VGG + CNN network, and performing fine adjustment and correction on the matched model by using a transfer learning algorithm, as shown in formula (4):
Figure BDA0002996394440000072
wherein Y represents the output data of the convolutional layer, again as a pooling layerThe data is input into the data processing system,
Figure BDA0002996394440000073
represents the output data of the pooling layer, and Ave (-) represents the averaging function.
The invention utilizes the test data to test the corrected VGG + CNN network, belongs to the conventional technical means, and is not detailed.
In the invention, the convolution function is used for carrying out feature extraction again on the output data of the VGG network. The function of the pooling function is to converge the output of each convolution layer into a final distance vector periodic characteristic, remove redundant information and reduce the calculation amount.
S4: according to the node density of the current network, under the condition of determining the maximum point-to-point transmission distance, on the basis of a fixed Medium Access Control (MAC) protocol of a link layer and in combination with the current situation of the Ickman spiral ocean current, the optimal distance vector synchronization period is obtained through a modified model, and the method specifically comprises the following steps:
s41: inputting the node density of the current network and the Ackerman spiral ocean current model into the VGG + CNN network after training as test data, and calculating the node density of the input network and the probability S of each category of the Ackerman spiral ocean current model by using the softmax classification layer of the VGG + CNN networkhThe calculation formula is shown as formula (5):
Figure BDA0002996394440000081
where h denotes the h-th neuron of the output layer, ShFor output, t represents the total number of output layer neurons;
s42: and sorting the magnitude values of the probabilities of the categories, and taking the category with the maximum probability value as an output classification result to finish the identification of the distance vector period.
The current network node density + the ackermann spiral ocean current constitutes a particular datum (i.e., the current test datum). In the invention, the trained VGG + CNN network is used for testing the current test data and judging which synchronization isTo which the period most likely corresponds, i.e. which synchronization period probability SpAnd maximum, namely, the identification of the optimal distance vector period is completed.
S5: initiating node distance vector synchronization by the sink node according to the obtained optimal period, and specifically comprising the following steps:
s51: in the channel access process, the sink node broadcasts a distance vector update binary group (turn, distance) to adjacent nodes, wherein the turn represents the synchronization of the distance vector initiated by the sink node for the second time, and the distance represents the hop count of the node from the sink node;
s52: each node updates the own distance vector according to the distance vector information of the adjacent node obtained in the channel interception and channel access processes, and broadcasts the distance vector carried by the node in the channel access process with the adjacent node, and so on;
s53: and when all the nodes finish the distance vector updating, the distance vector synchronization initiated by the sink node in the current round is finished.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims (4)

1. A network node distance vector synchronization method based on transfer learning is characterized by comprising the following specific construction steps:
s1: simulating node densities and an Ackerman spiral ocean current model of different networks by using a computer simulation technology to obtain optimal node distance vector synchronization period data under different conditions;
s2: performing feature fusion on the data obtained in the step S1 by adopting a K-means clustering algorithm to obtain an original data set;
s3: dividing the original data set obtained in the step S2 into training data and test data according to a certain proportion, inputting the training data into a VGG + CNN network for training, then performing fine tuning correction by using a transfer learning model, and testing the corrected VGG + CNN network by using the test data;
s4: obtaining an optimal distance vector synchronization period according to the node density and the ocean current condition of the current network and the corrected VGG + CNN network;
s5: and initiating node distance vector synchronization by the sink node according to the obtained optimal period.
2. The method for synchronizing the distance vectors between the network nodes based on the transfer learning of claim 1, wherein the step S3 is as follows:
s31, taking the training data as the input of the VGG + CNN network, and performing spatial transformation on the training data by using the feature extraction function of the VGG + CNN network, as shown in formula (1):
F(I)＝R (1)；
wherein F (-) represents a feature extraction function, I represents input training data, R represents output data of the VGG network, and R is equal to RA ×BRepresenting a real number space, A representing the row number of a matrix, and B representing the column number of the matrix;
the feature extraction function F (-) transforms input training data into features of A multiplied by B dimension; the output of the feature extraction function is then dimension transformed using the full join function, as shown in equation (2):
D(R)＝W (2)；
wherein D (-) represents a full connection function, W represents output data of a full connection network, the full connection function D (-) converts A × B dimensional data into AB × 1 dimensional data, and a full connection layer adopts a sigmoid activation function;
s32: performing feature extraction on the output data subjected to space and dimension transformation by using a convolution function of the VGG + CNN network, as shown in formula (3):
Figure FDA0003532131230000021
wherein Y (-) represents a convolution function, i represents a row index of the matrix, j represents a column index of the matrix, K represents a convolution kernel, R represents output data of the VGG network, m represents a row number of the matrix, and n represents a column number of the matrix;
s33: converging the data after feature extraction by using a pooling function of the VGG + CNN network, and performing fine adjustment and correction on the matched model by using a transfer learning algorithm, as shown in formula (4):
Figure FDA0003532131230000022
wherein Y represents the output data of the convolutional layer and is used as the input data of the pooling layer,
Figure FDA0003532131230000023
represents the output data of the pooling layer, and Ave (-) represents the averaging function.
3. The method according to claim 2, wherein in S4, based on the node density of the network, under the condition of determining the maximum transmission distance between nodes, based on a fixed medium access control protocol of a link layer, in combination with a current ackermann spiral ocean current condition, an optimal distance vector synchronization period is obtained through a modified model, and the method specifically includes:
s41: inputting the node density and the Ackerman spiral ocean current model of the current network into the VGG + CNN network after training as test data, and calculating the node density of the input network and the probability S of each category of the Ackerman spiral ocean current by using the softmax classification layer of the VGG + CNN networkhThe calculation formula is shown as formula (5):
Figure FDA0003532131230000031
where h denotes the h-th neuron of the output layer, ShFor output, t represents the total number of output layer neurons;
s42: and sorting the magnitude values of the probabilities of the categories, and taking the category with the maximum probability value as an output classification result to finish the identification of the distance vector period.
4. The method for network node distance vector synchronization based on transfer learning of claim 3, wherein the specific steps of S5 are as follows:
s51: in the channel access process, the sink node broadcasts a distance vector updating binary group to adjacent nodes, wherein the binary group comprises a turn and a distance, the turn represents the synchronization of a distance vector initiated by the sink node for the second time, and the distance represents the hop count of the node from the sink node;
s52: each node updates the own distance vector according to the distance vector information of the adjacent node obtained in the channel interception and channel access processes, and broadcasts the distance vector carried by the node in the channel access process with the adjacent node, and so on;
s53: and when all the nodes finish the distance vector updating, the distance vector synchronization initiated by the sink node in the current round is finished.
CN202110333523.8A 2021-03-29 2021-03-29 Network node distance vector synchronization method based on transfer learning Active CN113078958B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110333523.8A CN113078958B (en) 2021-03-29 2021-03-29 Network node distance vector synchronization method based on transfer learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110333523.8A CN113078958B (en) 2021-03-29 2021-03-29 Network node distance vector synchronization method based on transfer learning

Publications (2)

Publication Number Publication Date
CN113078958A CN113078958A (en) 2021-07-06
CN113078958B true CN113078958B (en) 2022-04-29

Family

ID=76611092

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110333523.8A Active CN113078958B (en) 2021-03-29 2021-03-29 Network node distance vector synchronization method based on transfer learning

Country Status (1)

Country Link
CN (1) CN113078958B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117313993B (en) * 2023-09-15 2024-07-12 湖南省烟草公司湘西自治州公司 Flue-cured tobacco cultivation management method based on temperature and altitude

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110148466A (en) * 2019-05-15 2019-08-20 东北大学 A kind of heart impact signal atrial fibrillation computer aided diagnosing method based on transfer learning
US10674399B2 (en) * 2018-06-20 2020-06-02 MeshPlusPlus, Inc. Clustered synchronization within a distance-vector wireless network
CN112183332A (en) * 2020-09-28 2021-01-05 成都希盟泰克科技发展有限公司 Yak face identification method based on transfer learning

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107786958A (en) * 2017-10-12 2018-03-09 中国科学院合肥物质科学研究院 A kind of data fusion method based on deep learning model

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10674399B2 (en) * 2018-06-20 2020-06-02 MeshPlusPlus, Inc. Clustered synchronization within a distance-vector wireless network
CN110148466A (en) * 2019-05-15 2019-08-20 东北大学 A kind of heart impact signal atrial fibrillation computer aided diagnosing method based on transfer learning
CN112183332A (en) * 2020-09-28 2021-01-05 成都希盟泰克科技发展有限公司 Yak face identification method based on transfer learning

Also Published As

Publication number Publication date
CN113078958A (en) 2021-07-06

Similar Documents

Publication Publication Date Title
CN111462282B (en) Scene graph generation method
WO2023280065A1 (en) Image reconstruction method and apparatus for cross-modal communication system
CN110263280B (en) Multi-view-based dynamic link prediction depth model and application
CN112181971A (en) Edge-based federated learning model cleaning and equipment clustering method, system, equipment and readable storage medium
CN106993320B (en) Wireless sensor network cooperative transmission routing method based on multiple relays and multiple hops
CN111563275A (en) Data desensitization method based on generation countermeasure network
CN108289285B (en) Method for recovering and reconstructing lost data of ocean wireless sensor network
CN113190939B (en) Large sparse complex network topology analysis and simplification method based on polygon coefficient
CN106535279B (en) Vehicle-mounted ad hoc network intersection route prediction method based on CP neural network
CN113078958B (en) Network node distance vector synchronization method based on transfer learning
CN115358487A (en) Federal learning aggregation optimization system and method for power data sharing
CN111224905B (en) Multi-user detection method based on convolution residual error network in large-scale Internet of things
CN115526316A (en) Knowledge representation and prediction method combined with graph neural network
CN112165401A (en) Edge community discovery algorithm based on network pruning and local community expansion
CN114615118A (en) Modulation identification method based on multi-terminal convolution neural network
CN108470251B (en) Community division quality evaluation method and system based on average mutual information
CN114301910A (en) Cloud-edge collaborative computing task unloading method in Internet of things environment
CN113923123A (en) Underwater wireless sensor network topology control method based on deep reinforcement learning
CN110661566B (en) Unmanned aerial vehicle cluster networking method and system adopting depth map embedding
Chen et al. An efficient neural network-based next-hop selection strategy for multi-hop VANETs
CN114979017B (en) Deep learning protocol identification method and system based on original flow of industrial control system
CN111160536A (en) Convolution embedding representation reasoning method based on fragmentation knowledge
CN115438722A (en) Big data product index clustering method based on distributed clustering
CN112163673B (en) Population routing method for large-scale brain-like computing network
CN112423400A (en) Ethernet communication link scheduling method based on improved firework algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant