CN111970716A - Fluid boundary extraction method based on classification hyperplane search in underwater acoustic sensor network - Google Patents

Fluid boundary extraction method based on classification hyperplane search in underwater acoustic sensor network Download PDF

Info

Publication number
CN111970716A
CN111970716A CN202010818276.6A CN202010818276A CN111970716A CN 111970716 A CN111970716 A CN 111970716A CN 202010818276 A CN202010818276 A CN 202010818276A CN 111970716 A CN111970716 A CN 111970716A
Authority
CN
China
Prior art keywords
nodes
event
fluid
cells
learning
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.)
Granted
Application number
CN202010818276.6A
Other languages
Chinese (zh)
Other versions
CN111970716B (en
Inventor
刘立
韩光洁
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.)
Changzhou Campus of Hohai University
Original Assignee
Changzhou Campus of Hohai University
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 Changzhou Campus of Hohai University filed Critical Changzhou Campus of Hohai University
Priority to CN202010818276.6A priority Critical patent/CN111970716B/en
Publication of CN111970716A publication Critical patent/CN111970716A/en
Application granted granted Critical
Publication of CN111970716B publication Critical patent/CN111970716B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/18—Network planning tools
    • H04W16/20—Network planning tools for indoor coverage or short range network deployment
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W40/00—Communication routing or communication path finding
    • H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • 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
    • 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)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)

Abstract

The invention discloses a fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network, which comprises a fluid coverage area search calibration stage, an elliptical envelope determination stage, an annular area division stage, a backbone unit cell selection stage and a fluid boundary extraction stage. The method can extract the fluid boundary with high precision, and has the advantages of rapid deployment, short-distance, real-time and seamless detection and tracking of the streaming media target and the like.

Description

Fluid boundary extraction method based on classification hyperplane search in underwater acoustic sensor network
Technical Field
The invention relates to a fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network, and belongs to the field of underwater acoustic sensors.
Background
With the increasing demand for energy by human life style, the frequency and scale of offshore industrial activities are increasing all the day around the world, the human economic activities in the coastal zone are highly concentrated, and the accidental leakage of petroleum, nuclear waste water and the like and the large-area outbreak of harmful algae such as red tide, water bloom and the like are easy to occur.
At present, the monitoring networking mode of the fluid target can be divided into four dimensions of earth, sea, air and sky according to the space region. However, the current monitoring of the fluid targets has many defects, such as remote sensing observation is restricted by revisit time and spatial resolution, and the fluid targets formed suddenly cannot be captured in time and the motion monitoring of the whole life cycle can not be provided; if the autonomous underwater vehicle is high in equipment cost, the autonomous underwater vehicle is only suitable for small-scale fluid tracking near ports, docks and the like, and needs to keep synchronous motion with the fluid to avoid target loss.
Disclosure of Invention
In order to overcome the defects and shortcomings of the prior art, the invention discloses a fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network.
The invention mainly adopts the technical scheme that:
a fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network comprises the following steps:
s1: fluid coverage area search calibration phase
S1-1: clustering sensor nodes according to geographical positions, wherein each cluster at least comprises 4 nodes, and forming different monitoring areas according to node clustering; the sensor nodes covered by the fluid targets can sense the existence of the fluid, the sensor nodes are called event nodes, and the clusters where the event nodes are located are called event clusters;
s1-2: all event nodes send notification messages to a cluster head of an event cluster i, the cluster head determines the maximum and minimum values of horizontal and vertical coordinates in all event nodes in the cluster according to collected event node ID and coordinates, the maximum and minimum values are respectively marked as Xmin _ ia, Xmax _ ib, Ymin _ ic and Ymax _ ID, and local quadruplets of Quads _ i [ L _ ia, R _ ib, D _ ic and U _ ID ] are generated according to the node coordinates corresponding to the maximum and minimum values of the horizontal and vertical coordinates, and the quadruplets are used for representing the coverage ranges of fluid in four directions of up, down, left and right in the cluster;
s1-3: each event cluster in the range of the underwater acoustic sensor network exchanges respective local quadruple, and each event cluster updates the quadruple according to the coordinate value in the received quadruple until the local quadruple of all event clusters in the network does not change any more; at the moment, the quadruple maintained by the event cluster in the network represents the coverage of the fluid in the upper, lower, left and right directions of the whole network; a bounding rectangle that can contain the fluid coverage area is determined using the quadruple.
S2: elliptical envelope determination phase
S2-1: determining an elliptical envelope according to the length-width ratio of the circumscribed rectangle; if the length-width ratio of the circumscribed rectangle is greater than or equal to 2, determining that the circumscribed ellipse of the rectangle is used as an envelope; and if the aspect ratio of the circumscribed rectangle is less than 2, determining that the circumscribed circle of the rectangle is used as the envelope.
S3: annular zone division phase
S3-1: setting the number of expected annular area layers as H, and uniformly dividing H concentric ellipses by taking the elliptical envelope as the outermost layer to obtain H annular areas;
s3-2: and dividing the annular regions of each layer according to the organization mode of the full binary tree. The i-th layer annular region is uniformly divided into 2i-1A sub-region. Finally, the elliptical envelope will contain 2H1 sub-partitions of different sizes. All the above sub-divisions are collectively called cells, and are located at the outermost layer 2H-1Each unit cell is called a leaf cell.
S4 backbone cell selection
S4-1: selecting backbone cells from the cells formed in step S3-2, wherein the conditions satisfying the backbone cells are: a leaf cell containing an event node; event nodes are contained in the cell, and at least one cell of the left and right child cells does not contain any event node; if both of the above conditions are satisfied, the backbone cell is selected.
S5: fluid boundary extraction stage
S5-1: and all the sensor nodes in each backbone unit cell are distributed and organized according to geographical positions to form a learning community. Each learning community contains the same number of nodes, so that different backbone unit cells have the same size but different numbers of learning communities. All the learning communities still keep the organization form of a full binary tree in structure;
s5-2: and designing a longitudinal cascading learning strategy and a transverse cascading learning strategy by utilizing the parent-child and brother adjacency relations maintained by full binary trees in each learning community, enabling nodes in different learning communities to take self position information and whether fluid is monitored as samples, and performing cooperative training in a support vector machine to obtain a classification hyperplane maximizing the interval between event nodes and non-event nodes as an extraction result of fluid boundaries.
Preferably, the manner of updating the quadruple by each event cluster in the step S1-3 is as follows: and comparing the coordinates in the upper, lower, left and right directions of the local quadruple with the coordinates in the received quadruple, and updating the quadruple by using the node coordinates corresponding to the maximum and minimum horizontal and vertical coordinates obtained after comparison.
Preferably, in the above-mentioned elliptical envelope determining stage, since the ellipse changes into a circle when the eccentricity is 0, the circular envelope can be regarded as a specific example of the elliptical envelope.
Preferably, in the above step S3-2, the specific division manner of dividing the annular regions in each layer according to the organization manner of the full binary tree is to refer to the innermost annular region, i.e. the first layer annular region, as the root region. The second layer annular region is evenly divided into 2 regions of equal area as the left and right sub-partitions of the root region. The third layer of annular area is evenly divided into 4 sub-partitions with equal area, and the sub-partitions are respectively used as the left sub-partition and the right sub-partition of the two partitions in the second layer. The above process is equivalent to the process of generating left and right child nodes by parent nodes in a full binary tree, and is continuously executed until the H-th layer where the elliptic envelope is positioned is reached, and the layer is uniformly divided into 2H-1And (4) sub-partition. Finally, the elliptical envelope will contain 2H1 sub-partitions of different sizes.
Preferably, all the sensor nodes of the backbone unit cell described in step S5-1 are distributed uniformly according to the density of geographical location distribution, and the specific distribution mode is as follows:
the cell areas of the brother relations are the same, the distribution number of the nodes is the same, and the learning community number is the same. For cells with parent-child relationship, the parent cell has an area 2 times that of the child cell, and thus the number of sensor nodes in the parent cell is 2 times that of the child cell, and similarly, the number of learning communities in the parent cell is 2 times that of the child cell.
Preferably, the above step S5-2 is to make the nodes in different learning communitiesThe self-position information and whether the fluid is monitored as the sample are { (x)1,y1)，(x2,y2),…,(xm,ym) Is input { z }1,z2,…,zmAnd is the sample set formed by the output. Wherein (x)i，yi) Cartesian coordinates representing sensor node i; z is a radical ofiLabel, z, representing sensor node iiE { +1, -1}, when z isiA value of +1 indicates that node i is an event node, and when z isiAnd when the value is-1, the node i is a non-event node.
Preferably, the training of the classification hyperplane by using the support vector machine in the step S5-2 is to input the training samples into a classification decision function, where the classification decision function is:
Figure BDA0002633528390000031
wherein N issvIndicates the number of support vectors, αiFor the corresponding lagrange multiplier of the support vector, K (,) is a gaussian kernel function. Order sigmoid [ f (x)]And (5) the obtained classification decision function represents a classification hyperplane, namely the estimation result of the position of the fluid boundary.
Preferably, the longitudinal cascade learning strategy in step S5-2 is directed to the set of backbone unit cells having parent-child relationship. Starting from the top of the longitudinal cascade structure, the parent backbone cells send the local support vectors generated by training in their respective learning communities to the child backbone cells. The child backbone unit cell combines training samples of all school communities with the received local support vector to train the local support vector of the child backbone unit cell, and then the local support vector is continuously transmitted downwards. And repeating the process until the bottom end of the longitudinal cascade structure. And feeding back the local support vector at the bottom end of the cascade structure to the top end of the cascade structure. And iterating the process until the support vector at the top end of the longitudinal cascade structure converges.
Preferably, the transverse cascading learning strategy in step S4-3 is directed to the backbone cell sets with sibling relationships. From the beginning of the transverse cascade structure, the backbone cells send the local support vectors generated by training in their respective learning communities to their sibling backbone cells. The brother backbone unit cell combines the training samples of all school communities with the received local support vector to train the local support vector of the brother backbone unit cell, and then the local support vector is continuously transmitted backwards. The above process is repeated until the end of the transverse cascade structure. The local support vector at the end of the cascade structure is fed back to the beginning of the cascade structure. And iterating the process until the support vector at the top end of the transverse cascade structure converges.
Has the advantages that: the invention provides a fluid boundary extraction method based on classification hyperplane search, which has the following advantages:
the hyperplane is extracted by using the underwater acoustic sensor network, and the deployment area and the fluid movement area are in the same dimension, so that the monitoring task can be rapidly unfolded conveniently. And facing the fluid target which is formed suddenly and the spatial distribution of which is changed along with the height of the motion process, the underwater acoustic sensing network can compensate the technical shortages of observation means such as satellite remote sensing and underwater vehicle cruising and the like, and provides long-time, short-distance, seamless and real-time fluid target monitoring and tracking service.
The fluid boundary extraction method based on the classification hyperplane search can be used for extracting the fluid edges with high precision. Meanwhile, the learning community is used as a unit to train and classify the hyperplane, the limited resources of the sensor nodes are fully considered, and the life cycle of the underwater acoustic sensor network is prolonged.
Drawings
FIG. 1 is a schematic illustration of a fluid coverage area search calibration and elliptical envelope determination of the present invention;
FIG. 2 is a schematic diagram of backbone cell selection according to the present invention;
FIG. 3 is a schematic diagram of the longitudinal cascade learning strategy of the present invention;
FIG. 4 is a schematic diagram of the lateral cascade learning strategy 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 fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network comprises a fluid coverage area search calibration stage, an elliptical envelope determination stage, an annular area division stage, backbone unit cells selection and a fluid boundary extraction stage.
As shown in fig. 1, the fluid coverage area search calibration phase includes:
s1: fluid coverage area search calibration phase
S1-1: clustering sensor nodes according to geographical positions, wherein each cluster at least comprises 4 nodes, and forming different monitoring areas according to node clustering; the sensor nodes covered by the fluid targets can sense the existence of the fluid, the sensor nodes are called event nodes, and the clusters where the event nodes are located are called event clusters; as shown in fig. 1, there are three event clusters, and the cluster head nodes are a0, B0, and C0, respectively.
S1-2: all event nodes send notification messages to a cluster head of an event cluster i, the cluster head determines the maximum and minimum values of horizontal and vertical coordinates in all event nodes in the cluster according to collected event node ID and coordinates, the maximum and minimum values are respectively marked as Xmin _ ia, Xmax _ ib, Ymin _ ic and Ymax _ ID, and local quadruplets of Quads _ i [ L _ ia, R _ ib, D _ ic and U _ ID ] are generated according to the node coordinates corresponding to the maximum and minimum values of the horizontal and vertical coordinates, and the quadruplets are used for representing the coverage ranges of fluid in four directions of up, down, left and right in the cluster; as shown in fig. 1, taking an event cluster in which a cluster head node a0 is located as an example, event nodes a0 to a4 exist in the cluster, the maximum and minimum values of the horizontal and vertical coordinates of all event nodes in the cluster are Xmin _ a2, Xmax _ a4, Ymin _ A3 and Ymax _ a1, respectively, and quadruplets _ a ═ L _ a2, R _ a4, D _ A3 and U _ a1, which are used for describing the distribution of fluid in four directions, namely, up, down, left and right;
s1-3: and exchanging respective local quadruples for each event cluster in the whole network range, and updating the local quadruples for each event cluster according to the coordinate values in the received quadruples until the local quadruples of all event clusters in the network are not changed any more. At the moment, the local quadruple maintained by the event cluster in the network represents the coverage of the fluid in four directions of the upper direction, the lower direction, the left direction and the right direction of the whole network. After the three event clusters in the graph exchange quadruples with each other, the quadruples finally describing the distribution range of the fluid in four directions of upper, lower, left and right on the full-grid range are [ L _ C2, R _ B5, D _ C3 and U _ A1], as shown in FIG. 1. A bounding rectangle that can contain the fluid coverage area is determined using the quadruple.
As shown in fig. 1, the elliptical envelope determination stage comprises:
s2-1: a circumscribed rectangle is determined that can contain the fluid coverage area.
S2-2: the elliptical envelope is determined from the aspect ratio of the circumscribed rectangle. If the length-width ratio of the circumscribed rectangle is greater than or equal to 2, determining that the circumscribed ellipse of the rectangle is used as an envelope; and if the aspect ratio of the circumscribed rectangle is less than 2, determining that the circumscribed circle of the rectangle is used as the envelope.
The annular region dividing stage comprises:
s3-1: setting the number of expected annular area layers as H, and uniformly dividing H concentric ellipses by taking the elliptical envelope as the outermost layer to obtain H annular areas;
s3-2: and dividing the annular regions of each layer according to the organization mode of the full binary tree. The i-th layer annular region is uniformly divided into 2i-1A sub-region. Finally, the elliptical envelope will contain 2H1 sub-partitions of different sizes. All the above sub-divisions are collectively called cells, and are located at the outermost layer 2H-1Each unit cell is called a leaf cell. As shown in fig. 2, the elliptical envelope is divided into 6 layers, and after further division of each layer, 63 sub-regions are obtained, numbered 1-63, wherein 32 leaf cells are provided, and the number is 32-63.
The selection stage of the backbone unit cell comprises the following steps:
s4-1: selecting backbone cells from the cells formed in step S3-2, wherein the conditions satisfying the backbone cells are: a leaf cell containing an event node; event nodes are contained in the cell, and at least one cell of the left and right child cells does not contain any event node; if both of the above conditions are satisfied, the backbone cell is selected. As shown in fig. 2, cells 9, 12-14, 16-18, 20-24, 29-31, 35-36, 47-48 are selected as backbone cells.
S5: fluid boundary extraction stage
S5-1: and all the sensor nodes in each backbone unit cell are distributed and organized according to geographical positions to form a learning community. Each learning community contains the same number of nodes, so that different backbone unit cells have the same size but different numbers of learning communities. All the learning communities still keep the organization form of a full binary tree in structure;
s5-2: and designing a longitudinal cascading learning strategy and a transverse cascading learning strategy by utilizing the parent-child and brother adjacency relations maintained by full binary trees in each learning community, enabling nodes in different learning communities to take self position information and whether fluid is monitored as samples, and performing cooperative training in a support vector machine to obtain a classification hyperplane maximizing the interval between event nodes and non-event nodes as an extraction result of fluid boundaries.
The event clusters in step S1-3 update the quadruple by comparing the local quadruple with the received quadruple in terms of coordinates in the four directions, and updating the local quadruple with the coordinates of the node corresponding to the maximum and minimum horizontal and vertical coordinates obtained after the comparison.
In the ellipse envelope determining stage, since the ellipse changes into a circle when the eccentricity is 0, the circular envelope can be regarded as a special case of the ellipse envelope.
In step S3-2, according to the organization manner of the full binary tree, the specific division manner for dividing the annular regions of the respective layers is to refer to the innermost annular region, i.e., the first layer annular region, as the root region. The second layer annular region is evenly divided into 2 regions of equal area as the left and right sub-partitions of the root region. The third layer of annular area is uniformly divided into 4 sub-partitions with equal area, and the sub-partitions are respectively used asLeft and right subdivisions of the two partitions in the second layer. The above process is equivalent to the process of generating left and right child nodes by parent nodes in a full binary tree, and is continuously executed until the H-th layer where the elliptic envelope is positioned is reached, and the layer is uniformly divided into 2H -1And (4) sub-partition. Finally, the elliptical envelope will contain 2H1 sub-partitions of different sizes.
All the sensor nodes of the backbone unit cell in the step S5-1 are uniformly distributed according to the density of geographical position distribution, and the specific distribution mode is as follows:
the cell areas of the brother relations are the same, the distribution number of the nodes is the same, and the learning community number is the same. In the cells of father-son relationship, the area of the father cell is 2 times that of the child cells, the node distribution number of the father cell is 2 times that of the child cells, and the learning community number of the father cell is 2 times that of the child cells.
In step S5-2, the nodes in different learning communities use their own location information and whether they monitor fluid as samples, so as to { (x)1，y1)，(x2，y2)，……，(xm，ym) And { z }1，z2，……，zmTogether constitute a sample set. Wherein (x)i，yi) Cartesian coordinates representing sensor node i; z is a radical ofiLabel, z, representing sensor node iiE { +1, -1}, when z isiA value of +1 indicates that node i is an event node, and when z isiAnd when the value is-1, the node i is a non-event node.
The training of the classification hyperplane by using the support vector machine described in step S5-2 is to input the training sample into a classification decision function, where the classification decision function is:
Figure BDA0002633528390000071
wherein N issvIndicates the number of support vectors, αiFor the corresponding lagrange multiplier of the support vector, K (,) is a gaussian kernel function. Order sigmoid [ f (x)]0.5, the obtained classification decision function represents a classification hyperplane, namely the fluid edgeAnd estimating the position of the boundary.
The longitudinal cascade learning strategy described in step S5-2 is directed to the set of backbone unit cells having parent-child relationship. As shown in fig. 2, the backbone cells 9, 18, 36 are in a parent-child relationship with each other, forming a set of three layers of longitudinal cascade structure. As shown in fig. 3, the backbone cell 9 is divided into 4 learning communities, LC i _1, LC i _2, LC i _3, and LC i _4, respectively. Since the backbone cell 18 is a child cell of the backbone cell 9, the backbone cell 18 can be divided into 2 learning communities LC i +1_1 and LC i +1_ 2. Since the backbone cell 36 is a child cell of the backbone cell 18, the backbone cell 36 can be divided into 1 learning community LC i +2_ 1. After the local support vectors SV i _1, SV i _2, SV i _3, SV i _4 are trained by the four learning communities of the backbone cell 9, the four local support vectors are sent to the learning communities LC i +1_1, LC i +1_2 of its child cell 18. The two learning communities combine the local training data set with the received local support vectors, train local support vectors SV i +1_1 and SV i +1_2 of the two learning communities, and then send SV i +1_1 and SV i +1_2 to the learning communities LC i +2_1 of the child unit cells 36. The learning community LC i +2_1 combines the local training data set with the received local support vector, trains out the local support vector SV i +2_1 of the local training data set, and feeds back SV i +2_1 to the four learning communities of the backbone unit cell 9. And iterating the process until the support vector at the top end of the longitudinal cascade structure converges.
The transverse cascading learning strategy described in step S5-2 is directed to the backbone cell set having sibling relationships. As shown in fig. 2, the backbone cells 20, 21, 22 are in a sibling relationship with each other, forming a set of lateral cascade structures. The areas of the backbone cells 20, 21, 22 and the number of learning communities are the same. As shown in fig. 4, each backbone cell is divided into a learning community, i.e., the backbone cell 20 is divided into a learning community LC i _1, the backbone cell 21 is divided into a learning community LC i _2, and the backbone cell 22 is divided into a learning community LC i _ 3. The learning community LC i _1 of the backbone cell 20 trains out the local support vector SV i _1 and sends SV i _1 to the learning community LC i _2 in its sibling cell 21. The learning community LC i _2 combines the local training data set with the received local support vector, trains its own local support vector SV i _2, and then sends SV i _2 to the learning community LC i _3 of its sibling cell 22. The learning community LC i _3 combines the local training data set with the received local support vector, trains the local support vector SV i _3 of the local training data set, and feeds back the SV i _3 to the learning community LC i _1 of the backbone cell 20. And iterating the process until the support vector at the top end of the transverse cascade structure converges.
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 (9)

1. A fluid boundary extraction method based on classification hyperplane search in an underwater acoustic sensor network is characterized by comprising the following steps:
s1: fluid coverage area search calibration phase
S1-1: clustering sensor nodes according to geographical positions, wherein each cluster at least comprises 4 nodes, and forming different monitoring areas according to node clustering; the sensor nodes covered by the fluid targets can sense the existence of the fluid, the sensor nodes are called event nodes, and the clusters where the event nodes are located are called event clusters;
s1-2: all event nodes send notification messages to a cluster head of an event cluster i, the cluster head determines the maximum and minimum values of horizontal and vertical coordinates in all event nodes in the cluster according to collected event node ID and coordinates, the maximum and minimum values are respectively marked as Xmin _ ia, Xmax _ ib, Ymin _ ic and Ymax _ ID, and local quadruplets of Quads _ i [ L _ ia, R _ ib, D _ ic and U _ ID ] are generated according to the node coordinates corresponding to the maximum and minimum values of the horizontal and vertical coordinates, and the quadruplets are used for representing the coverage ranges of fluid in four directions of up, down, left and right in the cluster;
s1-3: each event cluster in the range of the underwater acoustic sensor network exchanges respective local quadruple, and each event cluster updates the quadruple according to the coordinate value in the received quadruple until the local quadruple of all event clusters in the network does not change any more; at the moment, the quadruple maintained by the event cluster in the network represents the coverage of the fluid in the upper, lower, left and right directions of the whole network; determining a circumscribed rectangle that can contain a fluid coverage area using the quadruple;
s2: elliptical envelope determination phase
S2-1: determining an elliptical envelope according to the length-width ratio of the circumscribed rectangle; if the length-width ratio of the circumscribed rectangle is greater than or equal to 2, determining that the circumscribed ellipse of the rectangle is used as an envelope; if the length-width ratio of the circumscribed rectangle is less than 2, determining that the circumscribed circle of the rectangle is used as an envelope;
s3: annular zone division phase
S3-1: setting the number of expected annular area layers as H, and uniformly dividing H concentric ellipses by taking the elliptical envelope as the outermost layer to obtain H annular areas;
s3-2: dividing each layer of annular area according to the organization mode of the full binary tree; the i-th layer annular region is uniformly divided into 2i-1A sub-region; finally, the elliptical envelope will contain 2H1 sub-partitions of different sizes. All the above sub-divisions are collectively called cells, and are located at the outermost layer 2H-1Individual cells are called leaf cells;
s4: backbone cell selection
S4-1: selecting backbone cells from the cells formed in step S3-2, wherein the conditions satisfying the backbone cells are: a leaf cell containing an event node; event nodes are contained in the cell, and at least one cell of the left and right child cells does not contain any event node; selecting backbone crystal cells when one of the above conditions is satisfied;
s5: fluid boundary extraction stage
S5-1: all the sensor nodes in each backbone unit cell are distributed and organized according to geographical positions to form a learning community; the number of nodes contained in each learning community is the same, so that different backbone unit cells have the same scale and different numbers of learning communities; all the learning communities still keep the organization form of a full binary tree in structure;
s5-2: and designing a longitudinal cascading learning strategy and a transverse cascading learning strategy by utilizing the parent-child and brother adjacency relations maintained by full binary trees in each learning community, enabling nodes in different learning communities to take self position information and whether fluid is monitored as samples, and performing cooperative training in a support vector machine to obtain a classification hyperplane maximizing the interval between event nodes and non-event nodes as an extraction result of fluid boundaries.
2. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: the method for updating the quadruple of each event cluster comprises the steps of comparing the coordinates of the local quadruple with the coordinates of the received quadruple in the upper, lower, left and right directions, and updating the quadruple by using the node coordinates corresponding to the maximum and minimum horizontal and vertical coordinates obtained after comparison.
3. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: in the ellipse envelope determining stage, since the ellipse changes into a circle when the eccentricity is 0, the circular envelope can be regarded as a special case of the ellipse envelope.
4. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: in step S3-2, according to the organization manner of the full binary tree, the specific division manner for dividing the annular regions of each layer is as follows:
the innermost annular region, i.e. the first layer annular region, is called the root region;
the second layer of annular area is evenly divided into 2 areas with equal area as the left sub-area and the right sub-area of the root area;
the third layer of annular area is evenly divided into 4 sub-subareas with equal area, and the sub-subareas are respectively used as the left sub-subarea and the right sub-subarea of the two subareas in the second layer;
the division process is equivalent to the process of generating left and right child nodes by parent nodes in a full binary tree, and is continuously executed until the H-th layer where the elliptic envelope is positioned is obtained, and the layer is uniformly divided into 2H-1A sub-partition; finally, an ellipseWill contain 2 within the envelopeH1 sub-partitions of different sizes.
5. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: all the sensor nodes of the backbone unit cell in the step S5-1 are uniformly distributed according to the density of geographical location distribution, and the specific distribution mode is as follows:
the cell areas of brother relations are the same, the distribution number of nodes is the same, and the number of learning communities is the same;
in the cells of father-son relationship, the area of the father cell is 2 times that of the child cells, the node distribution number of the father cell is 2 times that of the child cells, and the learning community number of the father cell is 2 times that of the child cells.
6. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: in the step S5-2, the nodes in the learning community use their own location information and whether the fluid is monitored as a sample, which means that the sensor nodes in each learning community use their own cartesian coordinates (x)i，yi) For input, sensing label ziPackaging the output into a training sample; wherein z isiE { +1, -1}, when z isiA value of +1 indicates that node i is an event node, and when z isiAnd when the value is-1, the node i is a non-event node.
7. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: in the step S5-2, the support vector machine is used to train the classification hyperplane, and the training samples are input into a classification decision function, where the classification decision function is:
Figure FDA0002633528380000031
wherein N issvIndicates the number of support vectors, αiThe corresponding Lagrange multiplier of the support vector is adopted, and K (,) is a Gaussian kernel function; order sigmoid [ f (x)]And (5) the obtained classification decision function represents a classification hyperplane, namely the estimation result of the position of the fluid boundary.
8. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: the longitudinal cascading learning strategy in the step S5-2 is directed to the set of backbone unit cells having a parent-child relationship; starting from the top end of the longitudinal cascade structure, the father backbone unit cell sends local support vectors generated by training in each learning community to the child backbone unit cells; the child backbone unit cell combines training samples of all school communities with the received local support vector to train the local support vector of the child backbone unit cell, and then the local support vector is continuously transmitted downwards; repeating the above process until the bottom end of the longitudinal cascade structure; feeding back the local support vector at the bottom end of the cascade structure to the top end of the cascade structure; and iterating the process until the support vector at the top end of the longitudinal cascade structure converges.
9. The fluid boundary extraction method based on classification hyperplane search in the underwater acoustic sensor network according to claim 1, characterized in that: the transverse cascading learning strategy in the step S5-2 is directed to the backbone unit cell set having a brother relationship; from the beginning of the transverse cascade structure, the backbone unit cells send local support vectors generated by training in each learning community to brother backbone unit cells; the brother backbone unit cell combines the training samples of all school communities with the received local support vector to train the local support vector of the brother backbone unit cell, and then the local support vector is continuously transmitted backwards; repeating the above process until the end of the transverse cascade structure; feeding back the local support vector at the tail end of the cascade structure to the start of the cascade structure; and iterating the process until the support vector at the top end of the transverse cascade structure converges.
CN202010818276.6A 2020-08-14 2020-08-14 Fluid boundary extraction method based on classified hyperplane search in underwater acoustic sensor network Active CN111970716B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010818276.6A CN111970716B (en) 2020-08-14 2020-08-14 Fluid boundary extraction method based on classified hyperplane search in underwater acoustic sensor network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010818276.6A CN111970716B (en) 2020-08-14 2020-08-14 Fluid boundary extraction method based on classified hyperplane search in underwater acoustic sensor network

Publications (2)

Publication Number Publication Date
CN111970716A true CN111970716A (en) 2020-11-20
CN111970716B CN111970716B (en) 2023-06-20

Family

ID=73365761

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010818276.6A Active CN111970716B (en) 2020-08-14 2020-08-14 Fluid boundary extraction method based on classified hyperplane search in underwater acoustic sensor network

Country Status (1)

Country Link
CN (1) CN111970716B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113179136A (en) * 2021-04-23 2021-07-27 河海大学 Fluid boundary tracking method based on virtual node migration in underwater acoustic sensor network
CN113179135A (en) * 2021-04-23 2021-07-27 河海大学 Fluid boundary extraction method based on information entropy maximization in underwater acoustic sensor network

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101765095A (en) * 2009-12-14 2010-06-30 浙江大学 Tracking method for wireless sensor network moving target based on mixed cluster
CN104219704A (en) * 2014-09-24 2014-12-17 河海大学常州校区 Toxic gas boundary monitoring and tracking method based on double-layer mesh model in wireless sensor network
CN108364031A (en) * 2018-03-20 2018-08-03 盘锦中录油气技术服务有限公司 A kind of logging explanation plate classification boundaries division methods based on support vector machines

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101765095A (en) * 2009-12-14 2010-06-30 浙江大学 Tracking method for wireless sensor network moving target based on mixed cluster
CN104219704A (en) * 2014-09-24 2014-12-17 河海大学常州校区 Toxic gas boundary monitoring and tracking method based on double-layer mesh model in wireless sensor network
CN108364031A (en) * 2018-03-20 2018-08-03 盘锦中录油气技术服务有限公司 A kind of logging explanation plate classification boundaries division methods based on support vector machines

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
苏萁等: "基于机器学习算法的舰面流场预警系统研究", 《测控技术》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113179136A (en) * 2021-04-23 2021-07-27 河海大学 Fluid boundary tracking method based on virtual node migration in underwater acoustic sensor network
CN113179135A (en) * 2021-04-23 2021-07-27 河海大学 Fluid boundary extraction method based on information entropy maximization in underwater acoustic sensor network
CN113179135B (en) * 2021-04-23 2023-09-26 河海大学 Fluid boundary extraction method based on information entropy maximization in underwater acoustic sensor network
CN113179136B (en) * 2021-04-23 2023-11-28 河海大学 Fluid boundary tracking method based on virtual node migration in underwater acoustic sensor network

Also Published As

Publication number Publication date
CN111970716B (en) 2023-06-20

Similar Documents

Publication Publication Date Title
Rizzi et al. The production of scientific knowledge on renewable energies: Worldwide trends, dynamics and challenges and implications for management
CN105775054B (en) Ant colony formula dynamic oil spilling recovery method
CN111427100B (en) Typhoon center positioning method and device and typhoon path generation method
CN111970716A (en) Fluid boundary extraction method based on classification hyperplane search in underwater acoustic sensor network
CN110502006A (en) A kind of Mine Abandoned Land mobile robot complete coverage path planning method
CN106202744A (en) Wind power collection system of offshore wind power plant avoidance method for optimizing route and system
CN110263111A (en) Land use based on priori knowledge/cover information Space monitoring method
CN102496187B (en) Method for tracking contour line to boundary and fault based on triangular mesh
CN110889198B (en) Space-time probability distribution prediction method and system based on multi-factor joint learning
CN106959700A (en) A kind of unmanned aerial vehicle group collaboration patrol tracing path planing method based on upper limit confidential interval algorithm
CN108416392A (en) Building clustering method based on SOM neural networks
CN112926468B (en) Tidal flat elevation automatic extraction method
CN115294452B (en) Rotary SAR ship target detection method based on bidirectional characteristic pyramid network
CN111970715B (en) Fluid boundary extraction method based on virtual backbone node connection in underwater acoustic sensor network
CN106714193A (en) Settling deploying method for underwater sensor nodes based on ideal pattern model
CN113179135B (en) Fluid boundary extraction method based on information entropy maximization in underwater acoustic sensor network
CN115525769A (en) Battlefield environment data organization method and related device for the whole world
CN106872970A (en) A kind of multiple target based on differential evolution becomes data transfer rate tracks of device and its method
CN106780586A (en) A kind of solar energy potential evaluation method based on ground laser point cloud
CN116467565A (en) Enteromorpha green tide plaque optimal search area forecasting method
Wu et al. Multi-source and heterogeneous marine hydrometeorology spatio-temporal data analysis with machine learning: a survey
CN116562466B (en) Planning method for high orbit satellite observation task
Anas et al. Novel climate classification based on the information of solar radiation intensity: An application to the climatic zoning of Morocco
Huang A mathematical modeling and an optimization algorithm for marine ship route planning
CN115049825B (en) Water surface cleaning method, device, equipment and computer readable storage medium

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