CN109600710B - Multi-moving-target monitoring method based on difference algorithm in video sensor network - Google Patents

Multi-moving-target monitoring method based on difference algorithm in video sensor network Download PDF

Info

Publication number
CN109600710B
CN109600710B CN201811501729.1A CN201811501729A CN109600710B CN 109600710 B CN109600710 B CN 109600710B CN 201811501729 A CN201811501729 A CN 201811501729A CN 109600710 B CN109600710 B CN 109600710B
Authority
CN
China
Prior art keywords
video sensor
target
moving
node
sensor node
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
CN201811501729.1A
Other languages
Chinese (zh)
Other versions
CN109600710A (en
Inventor
蒋一波
王伟
郑旭峥
何成龙
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.)
Hangzhou Incandescent Orange Digital Technology Co ltd
Original Assignee
Zhejiang University of Technology ZJUT
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 Zhejiang University of Technology ZJUT filed Critical Zhejiang University of Technology ZJUT
Priority to CN201811501729.1A priority Critical patent/CN109600710B/en
Publication of CN109600710A publication Critical patent/CN109600710A/en
Application granted granted Critical
Publication of CN109600710B publication Critical patent/CN109600710B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/023Services making use of location information using mutual or relative location information between multiple location based services [LBS] targets or of distance thresholds
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/025Services making use of location information using location based information parameters
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W4/00Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02Services making use of location information
    • H04W4/029Location-based management or tracking services
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W64/00Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/003Locating users or terminals or network equipment for network management purposes, e.g. mobility management locating network equipment
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W84/00Network topologies
    • H04W84/18Self-organising networks, e.g. ad-hoc networks or sensor networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Image Analysis (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

A multi-moving-target monitoring method based on a difference algorithm in a video sensor network comprises the following steps: (1) randomly deploying a plurality of video sensor nodes in a monitoring area; (2) monitoring a moving target; (3) predicting the position of the moving target at the next moment; (4) the monitorable nodes establish and maintain a coverage rotation information table; (5) a monitoring node rotation decision based on a differential optimization algorithm; (6) the monitorable nodes rotate according to the final better solution. The invention can predict the position of the moving target at the next moment, and accordingly, the difference algorithm is utilized to select the coverable node to rotate in the sensing direction. The invention carries out global optimization at each moment, can cover the position where the target possibly appears at the maximum probability, and improves the coverage quality of multiple moving targets.

Description

Multi-moving-target monitoring method based on difference algorithm in video sensor network
Technical Field
The invention relates to a video sensor network, in particular to a method for monitoring and covering optimization of multiple moving targets in the video sensor network.
Technical Field
The video sensor is a brand-new information acquisition and target monitoring sensing device, so that a more convenient and efficient solution is provided for state sensing, information processing and tracking detection of a moving target. The video sensor network formed by the sensors has the characteristics of convenient and visual acquisition, transmission and processing capability of multimedia information such as data, images, videos and the like, and abundant information quantity of visual monitoring of the video sensor network. The high-efficiency capability of the method makes the method have good development prospect in the aspects of industry, agriculture, military, monitoring security and environmental monitoring.
Target coverage control is an important research hotspot in a video sensor network, however, a video sensor is different from a traditional sensor, and is limited by the visual field of the device, the sensing range of the video sensor is a sector area which takes a node as the center of a circle and the radius as the sensing distance, so that the traditional target coverage optimization algorithm is not suitable for the video sensor, in addition, the nodes in a complex monitoring area often adopt a random throwing strategy, the deployment randomness also brings great challenges to the research of target coverage, in addition, most of target coverage is directed to static targets, the monitoring research of moving targets with more random characteristics is very few, and the real-time monitoring quality of the video sensor network is still to be improved. Therefore, how to utilize the video sensor network to realize sufficiently and efficiently monitoring multiple moving targets in an area becomes a problem which needs to be solved urgently.
The method is capable of fundamentally solving the coverage problem, but the implementation cost is high, and partial areas may not be capable of deploying more sensors to meet the coverage requirement due to topographic features. In another method, an optimization algorithm strategy is designed according to the positions of randomly deployed video sensor nodes and the prediction of a target moving path, so that the sensing direction of the video sensor is rotated in real time, the moving target is accurately sensed and monitored, the track of the moving target is covered, and the effect of remote safety monitoring is achieved. Therefore, the accurate prediction of the path and the reasonable scheduling of the node perception direction become the key of the problem, the real-time performance and the reliability of the multi-moving-target monitoring are directly influenced by the path prediction and the node perception direction, and a good optimization monitoring algorithm has great significance.
Disclosure of Invention
In order to overcome the defects of small effect, low track coverage quality and poor real-time performance of the conventional video sensor network in sensing and monitoring of multiple moving targets, the invention provides a differential algorithm-based multiple moving target monitoring method in the video sensor network, which can predict the position of a moving target at the next moment and select a coverable node to rotate in a sensing direction by utilizing a differential algorithm. The invention carries out global optimization at each moment, can cover the position where the target possibly appears at the maximum probability, and improves the coverage quality of multiple moving targets.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a multi-moving-target monitoring method based on a difference algorithm in a video sensor network comprises the following steps:
(1) randomly deploying a plurality of video sensor nodes in a monitoring area, and numbering the video sensor nodes with the number S ═ Si1,2, …, n, each video sensor node is represented by a quintuple representing Si=<Pi,R,α,θ,ω>Respectively representing a node position, a perception radius, a perception visual angle, a perception direction angle and a rotation angular velocity;
(2) monitoring a moving target, wherein the process is as follows:
(2.1) numbering the moving targets newly entering the monitoring area, and adding the moving targets into a moving target set T ═ { T ═ Tj1,2, …, m };
(2.2) determining a target T at a certain timejNode S of video sensoriThe method for monitoring coverage is as follows:
Figure BDA0001898345930000021
and vector
Figure BDA0001898345930000022
Is located in [ theta ]i-α/2,θi+α/2]Wherein
Figure BDA0001898345930000023
Indicating a monitoring node SiPointing to a moving target TjThe modulus of the vector of (a), i.e. the vector absolute distance;
(3) predicting the position of the moving target at the next moment, and the process is as follows:
(3.1) according to the fact that the speed V of the moving target is kept unchanged in the whole moving process, the direction changes randomly, and the position to which the target can move at the time t +1 is predicted to be on a circle with the position at the time t as the center and the radius V delta t, namely:
Tjt+1={(Xjt+1,Yjt+1)|(Xjt+1-Xjt)2+(Yjt+1-Yjt(2=(VΔt)2},=0,1,2,…
taking into account TjDirection of movement V at time tjtThen T +1 time TjThe most likely direction of movement is VjtThe least probable direction of movement being-VjtI.e. the reverse direction;
(3.2) is target TjThe moving direction deviates from V at time t +1jtAngle of (2) introducing a probability density function
Figure BDA0001898345930000031
The direction of movement of the target at the next instant of time is random, the target direction of movement offset angle is a random variable, denoted X,
Figure BDA0001898345930000032
represents the probability density function of X, phi (X) represents the distribution function of X, P (a ≦ X ≦ b) represents that X is in the interval [ a, b ≦ b ]]A probability value of (d); the object can rotate the direction of movement clockwise or counterclockwise, so X ∈ [ - π, π]And satisfies the following conditions:
Figure BDA0001898345930000033
and the probability value of X in the interval [ a, b ] is:
Figure BDA0001898345930000034
that is, the predicted position of the target at time t +1 is located on a circle having a radius V Δ t and centered at the position at time t, and shifted by positive direction angles [ a, b ] with probabilities Φ (b) - Φ (a)]Or target next timeIs inscribed at the deviation VjtPositive direction [ a, b ]]The probability on the circular arc is phi (b) -phi (a);
therefore, the larger the arc coverage range of the node in the positive direction of the target is, the higher the coverage possibility of the node at the next moment of the target is;
(3.3) determining a set of monitorable nodes Z ═ Z according to whether the predicted circumferential positions of the plurality of moving objects enter into the coverage area of the monitoring node or noti1,2 … n, and a circular area with the node position as the center and the sensing radius of R is called as the coverage area of the node;
(4) the monitorable nodes create and maintain an overlay rotation information table, and the process is as follows:
(4.1) each monitorable node updates and maintains an overlay rotation information table, wherein the content of the overlay rotation information table comprises: four items of 'node S _ id', 'moving target T _ id', 'required rotation angle A' and 'activated or not';
(4.2) each node in the monitorable node set can monitor T each moving target according to all moving targets appearing in the coverage area of the nodejCalculating the covering rotation angle and forming a rotation set delta A ═ delta Aij1,2 … n, j 1,2 … m, each monitorable node according to its own perception direction angle thetaiRespectively calculating the rotation angle delta A covered on the predicted position of each moving target without deviation in the moving direction at the moment of t +1, and adding the information into a covering information table; according to node position PiPredicted position p without deviation from moving direction of target at next momentvCalculating a vector
Figure BDA0001898345930000041
Direction angle theta ofkCalculating the perceived direction angle theta of the nodeiAnd thetakIf the angle is less than pi, the node SiThe rotation angle is delta A-angle + alpha/2, otherwise, the delta A is 2 pi-angle-alpha/2;
(4.3) each monitorable node sends a coverage rotation information table of the node to the server;
(5) a monitoring node rotation decision based on a differential optimization algorithm comprises the following processes:
(5.1) the server initializes a population individual a to { a ] according to the received each node coverage rotation information table1,A2,…,ANPIn which A isk=(ΔAk1,ΔAk2,…,ΔAkn)TFor the kth individual of the population, NP is the population size, Δ AkiRepresentative monitoring node SiAngle of rotation of (1), boundary conditions
Figure BDA0001898345930000042
Wherein
Figure BDA0001898345930000043
Figure BDA0001898345930000044
Respectively represents Delta AkiThe upper and lower bounds of the span are related to the angular velocity ω;
(5.2) maximizing the sum of probability values corresponding to the arc segments of the node coverage target prediction positions, and designing a fitness function F:
Figure BDA0001898345930000045
wherein the node is covered to the target T with the probability g (x)jPredicting the position p at the next momentjThe coverage probability g (x) is:
Figure BDA0001898345930000051
judging the node S at the time tiCoverage to target TjPredicted position p ofjThe method is that
Figure BDA0001898345930000052
And vector
Figure BDA0001898345930000053
Is located in [ theta ]i-α/2,θi+α/2];
(5.3) realizing individual variation through a difference strategy, randomly selecting two different individuals in a population, scaling the vector difference of the two different individuals, and then carrying out vector synthesis with the individual to be varied, namely
vk(t+1)=Ar1(t)+G·(Ar2(t)-Ar3(t))j≠r1≠r2≠r3
Wherein G is a scaling factor, Ak(t) representing the kth individual in the population of the t generation, wherein in the evolution process, in order to ensure the effectiveness of a solution, whether the intermediate meets a boundary condition needs to be judged;
(5.4) interleaving: for the t generation population { Ak(t) } and intermediates of variants thereof { vk(t +1) } Cross-manipulations between individuals were performed, yielding test individuals uk(t+1);
Figure BDA0001898345930000054
Where CR is the crossover probability, irandIs [1, 2, …, n ]]Random integers of (a) ensure variant intermediates vk(t +1) at least one "gene" is inherited by the next generation;
(5.5) selecting the test individuals and the original individuals, and selecting the individuals with high fitness to enter the next generation of population by adopting a greedy algorithm:
Figure BDA0001898345930000055
(5.6) looping the above steps from (5.3) until the number of iterations is reached;
(6) the monitorable nodes rotate according to the final better solution, and the process is as follows:
(6.1) after the differential optimization algorithm is finished, the server obtains a better individual solved by the algorithm, and the 'gene position' of the individual corresponds to the rotation angle delta A of the relevant nodeki
(6.2) the server returns the better individual solution to each monitorable node, and the node finds the gene position of the node and rotates by a corresponding angle;
and (6.3) after waiting for a time step, transferring to the step (2) for recalculation until the monitoring is finished.
The technical conception of the invention is as follows: some video sensor nodes are randomly deployed in a target monitoring area, a probability density function is adopted in the offset direction according to the constant moving speed of the target, the possible position of the target at the next moment is deduced on the reachable circumferential position within the delta t time, and the process of monitoring the coverage of the moving target by selecting part of the video sensor nodes by using a differential algorithm strategy is given. The method can calculate the rotation angle required by the node capable of being monitored to cover the maximum possible reaching position of the moving target, calculates a better solution through the server overall situation, and finally sends the better solution to the node capable of being monitored for rotation decision.
The beneficial effects of the invention are mainly as follows: 1. the target moving track is predicted more accurately; 2. the target path coverage quality is improved; 3. the optimizing speed is high, and the monitoring real-time performance is high.
Drawings
Fig. 1 is a flow chart of a differential algorithm-based multi-moving-target monitoring method in a video sensor network.
FIG. 2 is a block diagram of a multi-moving-target monitoring optimization system based on a difference algorithm in a video sensor network.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 and 2, a method for monitoring multiple moving targets based on a difference algorithm in a video sensor network includes the following steps:
the first step is as follows: randomly deploying video sensor nodes in a target monitoring area, and numbering the video sensor nodes with the number S ═ Si|i=1,2,…,n};
The second step is that: judging whether a new moving target is monitored in the monitoring area, and if not, jumping to the fourth step;
the third step: numbering the newly monitored moving target, and adding the numbered moving target into a moving target set T ═ Tj1,2, …, m };
the fourth step: predicting the position of the moving target at the next moment; according to the fact that the moving target keeps the speed V unchanged in the whole moving process and the direction changes randomly, the position to which the target may move at the time t +1 is on a circle with the position at the time t as the center and the radius V delta t, namely:
Tjt+1={(Xjt+1,Yjt+1)|(Xjt+1-Xjt)2+(Yjt+1-Yjt)2=(VΔt)2},=0,1,2,…
taking into account TjDirection of movement V at time tjtThen T +1 time TjThe most likely direction of movement is VjtThe least probable direction of movement being-VjtI.e. the reverse direction;
according to the actual situation, is the target TjDeviating from the direction of movement V at time t +1jtAngle of (2) introducing a probability density function
Figure BDA0001898345930000071
The direction of movement at the next instant of the target is random, the target direction offset angle is a random variable, denoted by X,
Figure BDA0001898345930000072
represents the probability density function of X, phi (X) represents the distribution function of X, P (a ≦ X ≦ b) represents that X is in the interval [ a, b ≦ b ]]The object can rotate the direction of movement clockwise or counterclockwise, so that X ∈ [ - π, π]And satisfies the following conditions:
Figure BDA0001898345930000073
and the probability value of X in the interval [ a, b ] is:
Figure BDA0001898345930000074
i.e. the advance of the target at time t +1The measured position is on a circle with the position at the time t as the center and the radius V delta t, and the positive direction angle [ a, b ] is shifted by the probability phi (b) -phi (a)]Or the target falling off at the next moment in time from VjtPositive direction [ a, b ]]The probability on the circular arc is phi (b) -phi (a);
the larger the coverage range of the node on the arc segment in the positive direction of the target is, the higher the coverage possibility of the node on the target at the next moment is;
the fifth step: determining a set of monitorable nodes Z ═ Z according to whether the predicted circumferential positions of the plurality of moving objects enter into the coverage area of the monitoring nodesi1,2 … n, and a circular area with the node position as the center and the sensing radius of R is called as the coverage area of the node;
and a sixth step: each monitorable node updates and maintains an overlay rotation information table, and the content of the overlay rotation information table comprises: four items of 'node S _ id', 'moving target T _ id', 'required rotation angle A' and 'activated or not';
the seventh step: each node in the monitorable node set targets each moving target T according to all moving targets appearing in the coverage area of the nodejCalculating the covering rotation angle and forming a rotation set delta A ═ delta Aij1,2 … n, j 1,2 … m, each monitorable node according to its own perception direction angle thetaiRespectively calculating the rotation angle delta A covered on the predicted position of each moving target without deviation in the moving direction at the moment of t +1, and adding the information into a covering information table; according to node position PiPredicted position p without deviation from moving direction of target at next momentvCalculating a vector
Figure BDA0001898345930000081
Direction angle theta ofkCalculating the perceived direction angle theta of the nodeiAnd thetakIf the angle is less than pi, the node SiThe rotation angle is delta A-angle + alpha/2, otherwise, the delta A is 2 pi-angle-alpha/2;
eighth step: each monitorable node sends a coverage rotation information table of the node to the server;
the ninth step: monitoring node rotation decision based on difference optimization algorithm, and initializing a population individual A ═ A by a server according to a received node coverage rotation information table1,A2,…,ANPIn which A isk=(ΔAk1,ΔAk2,…,ΔAkn)TFor the kth individual of the population, NP is the population size, Δ AkiRepresentative monitoring node SiAngle of rotation of (1), boundary conditions
Figure BDA0001898345930000082
Wherein
Figure BDA0001898345930000083
Figure BDA0001898345930000084
Respectively represents Delta AkiThe upper and lower bounds of the span are related to the angular velocity ω;
the tenth step: and calculating the fitness value of the individual according to the fitness function F. The sum of probability values corresponding to the arc segments of the node coverage target prediction positions is maximized, so that:
Figure BDA0001898345930000091
wherein the node is covered to the target T with the probability g (x)jPredicting the position p at the next momentjThe coverage probability g (x) is:
Figure BDA0001898345930000092
judging the node S at the time tiCoverage to target TjPredicted position p ofjThe method is that
Figure BDA0001898345930000093
And vector
Figure BDA0001898345930000094
Is located in [ theta ]i-α/2,θi+α/2];
The eleventh step: realizing individual variation through a difference strategy, randomly selecting two different individuals in a population, scaling the vector difference of the two different individuals, and then carrying out vector synthesis with the individual to be varied, namely
vk(t+1)=Ar1(t)+G·(Ar2(t)-Ar3(t))j≠r1≠r2≠r3
Wherein G is a scaling factor, Ak(t) representing the kth individual in the population of the t generation, wherein in the evolution process, in order to ensure the effectiveness of a solution, whether the intermediate meets a boundary condition needs to be judged;
the twelfth step: and (3) cross operation: for the t generation population { Ak(t) } and intermediates of variants thereof { vk(t +1) } Cross-manipulations between individuals were performed, yielding test individuals uk(t+1),
Figure BDA0001898345930000095
Where CR is the crossover probability, irandIs [1, 2, …, n ]]Random integers of (a) ensure variant intermediates vk(t +1) at least one "gene" is inherited by the next generation;
the thirteenth step: selecting the test individuals and the original individuals, and selecting the individuals with high fitness to enter the next generation of population by adopting a greedy algorithm:
Figure BDA0001898345930000096
the fourteenth step is that: circulating the above steps from ten steps until the iteration times are reached;
the fifteenth step: after the differential optimization algorithm is finished, the server obtains a better individual according to the individual fitness value, and the monitorable node rotates according to the final better individual solution;
sixteenth, step: the server returns the solution of the better individual to each monitorable node, the node finds the gene position of the node and rotates the corresponding angle delta Aji
Seventeenth step: and after waiting for a time step, jumping to the second step to continue execution until the monitoring is finished.
Referring to fig. 2, the multi-moving-target monitoring and optimizing system based on the difference algorithm in the video sensor network implemented by the method mainly includes: the system comprises a moving target track prediction module, a differential algorithm node decision module, a monitorable node rotation module and a user interaction interface module.
(1) And the moving target track prediction module is used for predicting the target position at the next moment according to the constant moving target speed and the probability density function.
(2) A differential algorithm node decision module: and establishing a monitorable node set, maintaining and sending respective coverage rotation information tables to a server, and carrying out global optimization on the server based on a difference algorithm.
(3) The monitorable node rotation module: the monitoring node can accurately seek the position according to the better individual solution returned by the server and rotate according to the corresponding angle.
(4) A user interaction interface module: and configuring relevant monitoring environment parameters, such as the size of a monitoring area, the number of nodes of the video sensor, the positions of the nodes, the sensing radius, the initial sensing direction angle, the sensing visual angle, the time interval and the like, changing the environment parameters, and performing multi-experiment comparison.

Claims (1)

1. A multi-moving-target monitoring method based on a difference algorithm in a video sensor network is characterized by comprising the following steps:
(1) randomly deploying a plurality of video sensor nodes in a monitoring area, and numbering the video sensor nodes with the number S ═ Si1,2, …, n, each video sensor node is represented by a quintuple representing Si=<Pi,R,α,θ,ω>Respectively representing a node position, a perception radius, a perception visual angle, a perception direction angle and a rotation angular velocity;
(2) monitoring a moving target, wherein the process is as follows:
(2.1) numbering the moving targets newly entering the monitoring area, and adding the moving targets into a moving target set T ═ { T ═ Tj1,2, …, m };
(2.2) determining a target T at a certain timejNode S of video sensoriThe method for monitoring coverage is as follows:
Figure FDA0002480802270000011
and vector
Figure FDA0002480802270000012
Is located in [ theta ]i-α/2,θi+α/2]Wherein
Figure FDA0002480802270000013
Representing video sensor node SiPointing to a moving target TjThe modulus of the vector of (a), i.e. the vector absolute distance;
(3) predicting the position of the moving target at the next moment, and the process is as follows:
(3.1) according to the fact that the speed V of the moving target is kept unchanged in the whole moving process, the direction changes randomly, and the position to which the target can move at the time t +1 is predicted to be on a circle with the position at the time t as the center and the radius V delta t, namely:
Tjt+1={(Xjt+1,Yjt+1)|(Xjt+1-Xjt)2+(Yjt+1-Yjt)2=(VΔt)2},t=0,1,2,…
taking into account TjDirection of movement V at time tjtThen T +1 time TjThe most likely direction of movement is VjtThe least probable direction of movement being-VjtI.e. the reverse direction;
(3.2) is target TjThe moving direction deviates from V at time t +1jtAngle of (2) introducing a probability density function
Figure FDA0002480802270000017
The direction of movement of the target at the next instant of time is random, the target direction of movement offset angle is a random variable, denoted X,
Figure FDA0002480802270000014
represents the probability density function of X, phi (X) represents the distribution function of X, P (a ≦ X ≦ b) represents that X is in the interval [ a, b ≦ b ]]A probability value of (d); the object can rotate the direction of movement clockwise or counterclockwise, so X ∈ [ - π, π]And satisfies the following conditions:
Figure FDA0002480802270000015
and the probability value of X in the interval [ a, b ] is:
Figure FDA0002480802270000016
that is, the predicted position of the target at time t +1 is located on a circle having a radius V Δ t and centered at the position at time t, and shifted by positive direction angles [ a, b ] with probabilities Φ (b) - Φ (a)]Or the target falling off at the next moment in time from VjtPositive direction [ a, b ]]The probability on the circular arc is phi (b) -phi (a);
therefore, the larger the arc coverage range of the video sensor node in the positive direction of the target is, the higher the coverage possibility of the video sensor node on the target at the next moment is;
(3.3) determining a video sensor node set Z ═ Z according to whether the predicted circular positions of the plurality of moving objects enter the coverage area of the video sensor node or noti1,2 … n, taking the node position as the center of a circle, and regarding a circular area with the sensing radius of R as the coverable area of the video sensor node;
(4) the nodes of the visual sensor establish and maintain an overlay rotation information table, and the process is as follows:
(4.1) each video sensor node updates and maintains an overlay rotation information table, wherein the content of the overlay rotation information table comprises: four items of 'node S _ id', 'moving target T _ id', 'required rotation angle A' and 'activated or not';
(4.2) each video sensor node in the video sensor node set carries out T aiming at each moving target according to all moving targets appearing in the coverage area of the video sensor node setjCalculating the coverage rotation angleDegree, and constitutes a rotation set Δ a ═ Δ aij1,2 … n, j 1,2 … m, each video sensor node according to its own perception direction angle thetaiRespectively calculating the rotation angle delta A covered on the predicted position of each moving target without deviation in the moving direction at the moment of t +1, and adding the information into a covering information table; according to node position PiPredicted position p without deviation from moving direction of target at next momentvCalculating a vector
Figure FDA0002480802270000021
Direction angle theta ofkCalculating the sensing direction angle theta of the video sensor nodeiAnd thetakIf the angle is less than pi, the video sensor node SiThe rotation angle is delta A-angle + alpha/2, otherwise, the delta A is 2 pi-angle-alpha/2;
(4.3) each video sensor node sends a coverage rotation information table of the video sensor node to a server;
(5) the video sensor node rotation decision based on the difference optimization algorithm comprises the following processes:
(5.1) the server initializes a population individual a to { a ] according to the received coverage rotation information table of each video sensor node1,A2,…,ANPIn which A isk=(ΔAk1,ΔAk2,…,ΔAkn)TFor the kth individual of the population, NP is the population size, Δ AkiRepresenting a video sensor node SiAngle of rotation of (1), boundary conditions
Figure FDA0002480802270000022
Wherein
Figure FDA0002480802270000023
Respectively represents Delta AkiThe upper and lower bounds of the span are related to the angular velocity ω;
(5.2) maximizing the sum of probability values corresponding to the arc segments of the video sensor nodes covering the target prediction positions, and designing a fitness function F:
Figure FDA0002480802270000024
wherein the video sensor node is covered to the target T with the probability g (x)jPredicting the position p at the next momentjThe coverage probability g (x) is:
Figure FDA0002480802270000025
video sensor node S for judging time tiCoverage to target TjPredicted position p ofjThe method is that
Figure FDA0002480802270000031
And vector
Figure FDA0002480802270000032
Is located in [ theta ]i-α/2,θi+α/2];
(5.3) realizing individual variation through a difference strategy, randomly selecting two different individuals in a population, scaling the vector difference of the two different individuals, and then carrying out vector synthesis with the individual to be varied, namely
vk(t+1)=Ar1(t)+G·(Ar2(t)-Ar3(t))j≠r1≠r2≠r3
Wherein G is a scaling factor, Ak(t) representing the kth individual in the population of the t generation, wherein in the evolution process, in order to ensure the effectiveness of a solution, whether the intermediate meets a boundary condition needs to be judged;
(5.4) interleaving: for the t generation population { Ak(t) } and intermediates of variants thereof { vk(t +1) } Cross-manipulations between individuals were performed, yielding test individuals uk(t+1);
Figure FDA0002480802270000033
Where CR is the crossover probability, irandIs [1, 2, …, n ]]Is a random integer of (a) to (b),ensures variant intermediate vk(t +1) at least one "gene" is inherited by the next generation;
(5.5) selecting the test individuals and the original individuals, and selecting the individuals with high fitness to enter the next generation of population by adopting a greedy algorithm:
Figure FDA0002480802270000034
(5.6) looping the above steps from (5.3) until the number of iterations is reached;
(6) the video sensor node rotates according to the final better solution, and the process is as follows:
(6.1) after the differential optimization algorithm is finished, the server obtains a better individual solved by the algorithm, and the 'gene position' of the individual corresponds to the rotation angle delta A of the relevant nodeki
(6.2) the server returns the better individual to each video sensor node, and the video sensor nodes find the gene positions of the video sensor nodes and rotate by corresponding angles;
and (6.3) after waiting for a time step, transferring to the step (2) for recalculation until the monitoring is finished.
CN201811501729.1A 2018-12-10 2018-12-10 Multi-moving-target monitoring method based on difference algorithm in video sensor network Active CN109600710B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811501729.1A CN109600710B (en) 2018-12-10 2018-12-10 Multi-moving-target monitoring method based on difference algorithm in video sensor network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811501729.1A CN109600710B (en) 2018-12-10 2018-12-10 Multi-moving-target monitoring method based on difference algorithm in video sensor network

Publications (2)

Publication Number Publication Date
CN109600710A CN109600710A (en) 2019-04-09
CN109600710B true CN109600710B (en) 2020-10-30

Family

ID=65961488

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811501729.1A Active CN109600710B (en) 2018-12-10 2018-12-10 Multi-moving-target monitoring method based on difference algorithm in video sensor network

Country Status (1)

Country Link
CN (1) CN109600710B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113784361B (en) * 2021-05-13 2022-07-19 中国地质大学(武汉) Dynamic adjustment method and device for sensor node and storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184551A (en) * 2011-05-10 2011-09-14 东北大学 Automatic target tracking method and system by combining multi-characteristic matching and particle filtering
CN103188707A (en) * 2013-03-12 2013-07-03 南京邮电大学 Path coverage monitoring method for wireless multimedia sensor network
CN104038730A (en) * 2014-05-09 2014-09-10 浙江工业大学 Greedy algorithm-based video sensor network multi-mobile target monitoring and optimizing method
CN104349356A (en) * 2013-08-05 2015-02-11 江南大学 Video sensor network coverage enhancement realization method based on differential evolution
CN104602251A (en) * 2014-12-31 2015-05-06 浙江工业大学 Multi-moving-target dynamic monitoring optimization method based on diploid genetic algorithm in video sensor network
CN105974799A (en) * 2016-07-15 2016-09-28 东南大学 Fuzzy control system optimization method based on differential evolution-local unimodal sampling algorithm

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184551A (en) * 2011-05-10 2011-09-14 东北大学 Automatic target tracking method and system by combining multi-characteristic matching and particle filtering
CN103188707A (en) * 2013-03-12 2013-07-03 南京邮电大学 Path coverage monitoring method for wireless multimedia sensor network
CN104349356A (en) * 2013-08-05 2015-02-11 江南大学 Video sensor network coverage enhancement realization method based on differential evolution
CN104038730A (en) * 2014-05-09 2014-09-10 浙江工业大学 Greedy algorithm-based video sensor network multi-mobile target monitoring and optimizing method
CN104602251A (en) * 2014-12-31 2015-05-06 浙江工业大学 Multi-moving-target dynamic monitoring optimization method based on diploid genetic algorithm in video sensor network
CN105974799A (en) * 2016-07-15 2016-09-28 东南大学 Fuzzy control system optimization method based on differential evolution-local unimodal sampling algorithm

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
《k -Angle Object Coverage Problem in a Wireless Sensor Network》;Yu-Chee Tseng;《IEEE Sensors Journal》;20120504;全文 *
《New Method for Weighted Coverage Optimization of Occlusion-Free Surveillance in Wireless Multimedia Sensor Network》;Yibo Jiang;《2010 First International Conference on Networking and Distributed Computing》;20101024;全文 *
《视频传感器网络覆盖优化及目标追踪技术的研究》;翟天琦;《中国学位论文全文数据库》;20180530;全文 *

Also Published As

Publication number Publication date
CN109600710A (en) 2019-04-09

Similar Documents

Publication Publication Date Title
Akcan et al. GPS-Free node localization in mobile wireless sensor networks
Han et al. A mobile anchor assisted localization algorithm based on regular hexagon in wireless sensor networks
US9069055B2 (en) Wireless positioning method and apparatus using wireless sensor network
Chuang et al. An effective PSO-based node localization scheme for wireless sensor networks
Wang et al. Using rotatable and directional (R&D) sensors to achieve temporal coverage of objects and its surveillance application
Symington et al. Encounter based sensor tracking
CN107064912A (en) System and method for the sane and efficient location estimation based on TDOA in the case where there is various multipaths to postpone
CN112558076B (en) Volume scanning mode calculation method based on networking weather radar coverage area and application
CN109600710B (en) Multi-moving-target monitoring method based on difference algorithm in video sensor network
Song et al. Multiple target counting and tracking using binary proximity sensors: Bounds, coloring, and filter
Manjarres et al. On the design of a novel two-objective harmony search approach for distance-and connectivity-based localization in wireless sensor networks
Fang et al. Fingerprint localisation algorithm for noisy wireless sensor network based on multi‐objective evolutionary model
Waqar et al. Smartphone positioning in sparse Wi-Fi environments
Kotrotsios et al. Accurate gridless indoor localization based on multiple bluetooth beacons and machine learning
Mozamir et al. Improved GbLN-PSO Algorithm for Indoor Localization in Wireless Sensor Network.
Wang et al. Optimal target tracking based on dynamic fingerprint in indoor wireless network
Xu et al. Fortifying barrier-coverage of wireless sensor network with mobile sensor nodes
Chen et al. Graph-based radio fingerprint augmentation for deep-learning-based indoor localization
Brouwers et al. Detecting dwelling in urban environments using gps, wifi, and geolocation measurements
Liu et al. Barrier coverage in mobile camera sensor networks with grid-based deployment
KR20200132295A (en) Apparatus and method for indoor location estimation based on packet capture
CN109874149A (en) Localization method, device and the computer readable storage medium of mobile terminal
Zhang et al. Context-aware telco outdoor localization
Niranchana et al. Optimized tracking and detection of target nodes in wireless sensor networks
Priyadarshini et al. A distributed scalar controller selection scheme for redundant data elimination in sensor networks

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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20230717

Address after: 310000 rooms 702 and 703, Yuesheng International Center, ningwei street, Xiaoshan District, Hangzhou City, Zhejiang Province

Patentee after: Hangzhou Incandescent Orange Digital Technology Co.,Ltd.

Address before: The city Zhaohui six districts Chao Wang Road Hangzhou City, Zhejiang province 310014 18

Patentee before: JIANG University OF TECHNOLOGY