CN110717507A - Soil moisture content sensor optimization layout method based on APDJ algorithm - Google Patents

Soil moisture content sensor optimization layout method based on APDJ algorithm Download PDF

Info

Publication number
CN110717507A
CN110717507A CN201910808140.4A CN201910808140A CN110717507A CN 110717507 A CN110717507 A CN 110717507A CN 201910808140 A CN201910808140 A CN 201910808140A CN 110717507 A CN110717507 A CN 110717507A
Authority
CN
China
Prior art keywords
soil moisture
moisture content
algorithm
nodes
sensor
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
CN201910808140.4A
Other languages
Chinese (zh)
Other versions
CN110717507B (en
Inventor
张武
张嫚嫚
韩华威
张森林
饶元
金�秀
苗犇犇
冯金磊
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.)
Hefei Hanya Information Technology Co Ltd
Anhui Agricultural University AHAU
Original Assignee
Hefei Hanya Information Technology Co Ltd
Anhui Agricultural University AHAU
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 Hefei Hanya Information Technology Co Ltd, Anhui Agricultural University AHAU filed Critical Hefei Hanya Information Technology Co Ltd
Priority to CN201910808140.4A priority Critical patent/CN110717507B/en
Publication of CN110717507A publication Critical patent/CN110717507A/en
Application granted granted Critical
Publication of CN110717507B publication Critical patent/CN110717507B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Forestry; Mining
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Mining & Mineral Resources (AREA)
  • General Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Animal Husbandry (AREA)
  • Marine Sciences & Fisheries (AREA)
  • Agronomy & Crop Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Investigating Or Analyzing Materials By The Use Of Electric Means (AREA)
  • Arrangements For Transmission Of Measured Signals (AREA)

Abstract

The invention discloses an optimized layout method of soil moisture sensors based on an APDJ algorithm, provides an improved Dijkstra algorithm (APDJ algorithm), and develops the application of the optimized layout of the soil moisture sensors. On the basis of ensuring the full coverage of a sensing network, soil moisture content data of each sensing node is collected in real time, an AP clustering algorithm is used for obtaining a clustering center, the clustering center is used as a starting point of a Dijkstra algorithm for searching a shortest path, a convergence point of a wireless sensing network is used as an end point, an improved Dijkstra algorithm is used for searching an optimal path by taking the maximum data dissimilarity and the shortest distance as targets, and finally an optimal point distribution path consisting of 6 sensor nodes is obtained.

Description

Soil moisture content sensor optimization layout method based on APDJ algorithm
Technical Field
The invention relates to the technical field of optimal layout of soil moisture sensors, in particular to optimal layout of soil moisture sensors by adopting clustering and Dijkstra algorithm.
Background
The agricultural water-saving irrigation system collects soil moisture content information through the sensors distributed in the farmland, and reasonable sensor selection and optimized layout play an important role in accurate acquisition of soil moisture content. With respect to the problem of optimal layout of sensors, a great deal of research is carried out on the aspects of set coverage, multi-objective combination optimization and the like, and the method is widely applied to multiple fields. Considering the fault detection capability of the sensor, Zhu xi Hua et al establishes a sensor layout optimization model, and realizes the layout optimization of the sensor by using the improved discrete particle swarm algorithm. Yang sail and the like research the optimal distribution of the sensors and the algorithm thereof under the condition of limited resources, establish a probability SDG model and apply the probability SDG model to a boiler system. Liulimna et al have studied the problem of sensor node distribution with the greatest coverage efficiency of wireless sensor networks and have proposed a nested grid coverage method that adapts to physical quantity distribution. At present, the research on the optimization layout problem of farmland soil moisture content sensors is rare. Wu Zhengyu et al combine the genetic algorithm with the weighted round set layout theory, and obtain good effect through the optimization calculation of the soil moisture content sensor under the constraint conditions of considering the coverage accuracy of the sensor, the overlap limit and the like.
At present, people focus on a stationing model and a coverage algorithm for sensor layout research, and optimization research is developed from the aspects of base station quantity, network connectivity, minimum coverage in an environment with obstacles and the like through different algorithm models. However, relevant researches on the redundancy of soil moisture content data and a sensor distribution path are still lacked. In the optimization layout process of the farmland soil moisture content sensor, not only the energy loss, the coverage precision, the signal completeness and the like of a system are considered, but also the problems of similarity and difference of soil moisture content data are considered, so that the layout of the sensor is reasonably optimized with the aims of reducing the redundancy of the data and reducing the input cost of the sensor.
Disclosure of Invention
The invention aims to make up for the defects of the prior art, combine the AP clustering with the Dijkstra algorithm, provide the improved Dijkstra algorithm and apply the improved Dijkstra algorithm to the optimal layout of the soil moisture content sensor, and the purpose of optimizing the layout of the sensor is to reduce the redundancy of data and the system cost.
The invention is realized by the following technical scheme:
the method comprises the following steps of (1) carrying out optimized layout on soil moisture content sensors by using an AP clustering algorithm and a Dijkstra algorithm:
(1) firstly, collecting soil moisture content data: arranging a plurality of soil moisture content sensors in an evenly distributed mode;
(2) and clustering soil moisture content data: collecting soil moisture content data on the basis of full coverage of a sensing network, and clustering the collected data by using an AP (access point) clustering algorithm to obtain a clustering center;
(3) constructing an adjacency matrix of the Dijkstra algorithm: determining an edge weight value by the weighted combination of the dissimilarity value of the soil moisture content data and the network node path value, and constructing an adjacency matrix of the Dijkstra algorithm;
(4) and searching the optimal path by the improved Dijkstra algorithm: and (3) searching the optimal path of sensor point distribution by using the Dijkstra algorithm with the AP clustering center in the step (2) as a starting point, the convergent point of the wireless sensor network as an end point and the maximum data dissimilarity and the shortest distance as targets.
The method comprises the following steps that in the step (1), a plurality of soil moisture content sensors are arranged in an evenly distributed mode, and the method specifically comprises the following steps: 5 sensor detection points are uniformly arranged in the transverse direction and the longitudinal direction respectively, 25 detection points are arranged in total, and each detection point acquires the mass water content parameter of soil 25cm below the surface of the soil.
And (3) the AP clustering algorithm in the step (2) realizes clustering by calculating a similarity matrix of the water content of the soil nodes to obtain a clustering center. The method comprises the following specific steps:
suppose { m }1,m2,...,mnAnd (4) a data sample set, wherein the clustering algorithm uses negative Euclidean distance to express the similarity among n data points, and the similarity is recorded as:
si,j＝-||mi-mj||2,i≠j (1)
i. j respectively represents the ith and the jth data points in the data sample set, n data points form an n multiplied by n similarity matrix S, and the clustering center is formed by the similarity matrix S [ si, j ]]n×nAnd determining the numerical value of an element s (k, k) on the diagonal, wherein the s (k, k) is called a bias parameter p, the size of the p value influences the clustering result, the larger the p is, the more the clustering number is, and the smaller the p value is, the less the clustering number is.
Constructing an adjacency matrix of the Dijkstra algorithm in the step (3), which is specifically as follows:
1) calculating the dissimilarity degree of soil moisture content of the sensor nodes:
and (3) calculating the sum of the dissimilarity values between the data of each sensor node and the data of the rest other nodes, and reflecting the dissimilarity degree of soil moisture content of each node in the whole test area from the global perspective, wherein the dissimilarity degree calculation formula is shown as a formula (2).
Figure BDA0002184273410000031
Wherein n is the number of sensing nodes, riAnd rjRepresenting the soil mass water content, R, of the ith and jth sensor nodesiThe larger the value is, the larger the difference between the soil moisture content of the point and the soil moisture content of the whole test area is.
2) Determining the weight value of the edge between the sensor nodes:
obtaining the dissimilarity value of each sensor node according to the formula (2) and calculating Tij，TijIs equal to the sum of the reciprocal of the dissimilarity value of the sensor node i and the reciprocal of the dissimilarity value of the sensor node j; dijDenotes the planar distance, x, between 2 sensor nodesi,yiRepresenting plane coordinates. Edge weight W between two sensor nodesijFrom TijAnd DijAnd obtaining the result by weighted combination. The calculation formula is (3), (4) and (5).
Figure BDA0002184273410000033
Wij＝αTij+βDij(i,j＝1,2,...,n) (5)
Alpha and beta are weight factors, which represent T in the edge weight value WijAnd DijThe proportion of the components is calculated; t isijEqual to the sum of reciprocal of dissimilarity value of two nodes, the greater the dissimilarity of two nodes, the greater TijThe smaller; dijRepresenting the planar distance between two adjacent nodes.
3) Constructing an adjacent matrix:
and (3) calculating the edge weights of 25 sensor nodes, wherein the edge weight formed by two adjacent sensors is calculated by a formula (5), the edge weight between two non-adjacent nodes is set to be infinite, which indicates that the two nodes cannot be reached, and finally the edge weight formed by 25 sensor nodes is stored by using an adjacency matrix.
Searching the best path by the improved Dijkstra algorithm in the step (4): the method comprises the following steps of taking an AP clustering center as a starting point, a convergence point of a wireless sensor network as an end point, and a Dijkstra algorithm to search an optimal path for sensor point distribution by taking the maximum data dissimilarity and the shortest distance as targets, wherein the optimal path is as follows:
applying the adjacency matrix calculated in the step (3) to a Dijkstra algorithm, and taking F as the reference value according to the starting point and the end point determined in the step (2)min(ri,rj,xi,yi,xj,yj)＝αTij(ri,rj)+βDij(xi,yi,xj,yj) And searching a path for the objective function to finally obtain the optimal path.
On the basis of researching the layout problem of the soil moisture sensors for water-saving irrigation, the invention combines an Affinity Propagation Clustering (AP Clustering for short) algorithm with a Dijkstra algorithm, provides an improved Dijkstra algorithm (APDJ algorithm) and applies the improved Dijkstra algorithm to the optimized layout of the soil moisture sensors. The method comprises the steps of arranging soil moisture content sensors on the basis of full coverage of a wireless sensor network, carrying out AP clustering on soil moisture content data of all nodes, determining edge weights by taking a clustering center as a starting point and a wireless sensor network convergence point as an end point and taking weighted combination of network node path values and data dissimilarity values, constructing an adjacency matrix, searching an optimal path for soil moisture content data transmission by using a Dijkstra algorithm, optimizing sensor layout, reducing the number of sensors and saving cost.
The invention has the advantages that:
(1) in the past, researches on the layout of farmland soil moisture sensors mainly focus on the aspect of optimization algorithms of network coverage rate, and related researches on redundancy problems of soil moisture data are lacked, so that an affinity propagation Clustering (AP Clustering for short) algorithm is combined with a Dijkstra algorithm, an improved Dijkstra algorithm is provided, and the improved Dijkstra algorithm is applied to the optimization layout of the soil moisture sensors. Arranging soil moisture content sensors on the basis of full coverage of a wireless sensor network, carrying out AP clustering on soil moisture content data of each node, and taking a clustering center as a starting point and a wireless sensor network convergence point as an end point; determining edge weight values by weighted combination of network node path values and data dissimilarity values, constructing an adjacency matrix, searching an optimal path for soil moisture content data transmission by using a Dijkstra algorithm, optimizing sensor layout, reducing data redundancy and sensor quantity, and saving cost.
(2) The clustering algorithm used in the method is a deterministic algorithm, clustering results of multiple independent operations are generally stable, the number and arrangement positions of the obtained sensors are the same after the multiple independent operations, the reliability and the stability are realized, and the Dijkstra algorithm is simple and is convenient to realize.
Drawings
FIG. 1 is a schematic diagram of the implementation of the method of the present invention.
FIG. 2 is a distribution plot of experimental data acquisition points for the method of the invention.
FIG. 3 is a plot of the spatial distribution of field capacity for the experimental area of the method of the present invention.
Fig. 4 is a diagram of the AP clustering result of the method of the present invention.
Detailed Description
As shown in fig. 1 and 2, firstly, on the basis of ensuring the full coverage of the sensing network, collecting soil moisture content data of each node in a test area in real time, clustering the collected data by using an AP clustering algorithm to obtain a clustering center, secondly, determining an edge weight value by using a weighted combination of a dissimilarity value of the soil moisture content data and a network node path value, and constructing an adjacency matrix of a Dijkstra algorithm; and finally, with the class center as a starting point, a convergence point of the wireless sensor network as an end point, and with the maximum data dissimilarity and the shortest distance as targets, searching for an optimal path by using a Dijkstra algorithm, and arranging the soil moisture content data acquired by the sensors on the path to achieve low redundancy, so that the system cost is saved, and the problem of optimal layout of the sensors is well solved. The method specifically comprises the following steps:
step (1), collecting soil moisture content data, wherein the area of a test site is about 16000m2About 190m in the transverse direction and about 85m in the longitudinal direction. The sensors are uniformly distributed. In order to ensure the full coverage of the sensing network of the test area, 5 sensor detection points are uniformly arranged in the transverse direction, 5 sensor detection points are also uniformly arranged in the longitudinal direction, and 25 detection points are arranged in total. The mass water content of the soil 25cm below the soil surface was collected at each test point.
Step (2), clustering soil moisture content data: and collecting soil moisture content data on the basis of full coverage of the sensing network, and clustering the collected data by using an AP (access point) clustering algorithm to obtain a clustering center.
Step (3), constructing an adjacency matrix of the Dijkstra algorithm: and determining an edge weight value by the weighted combination of the dissimilarity value of the soil moisture content data and the network node path value, and constructing an adjacency matrix of the Dijkstra algorithm. The specific process is as follows:
(1) calculating the dissimilarity degree of soil moisture content of the sensor nodes:
and (3) calculating the sum of the dissimilarity values between the data of each sensor node and the data of the rest other nodes, and reflecting the dissimilarity degree of soil moisture content of each node in the whole test area from the global perspective, wherein the dissimilarity degree calculation formula is shown as a formula (2).
Wherein n is the number of sensing nodes, riAnd rjRepresenting the soil mass water content, R, of the ith and jth sensor nodesiThe larger the value is, the larger the difference between the soil moisture content of the point and the soil moisture content of the whole test area is.
(2) Determining the weight value of the edge between the sensor nodes:
the AP clustering and Dijkstra algorithm soil moisture sensor optimal layout method according to claim 4, wherein the dissimilarity value of each sensing node is obtained by formula (2) in step 1, and T is calculatedij，TijIs equal to the sum of the reciprocal of the dissimilarity value of the sensor node i and the reciprocal of the dissimilarity value of the sensor node j; dijRepresents the planar distance between two sensor nodes, (x)i,yi) Representing plane coordinates. Edge weight W between two sensor nodesijFrom TijAnd DijAnd obtaining the result by weighted combination. The calculation formula is (3), (4) and (5).
Figure BDA0002184273410000061
Figure BDA0002184273410000062
Wij＝αTij+βDij(i,j＝1,2,...,n) (5)
Alpha and beta are weight factors, which represent T in the edge weight value WijAnd DijThe proportion of the components is calculated; t isijEqual to the sum of reciprocal of dissimilarity value of two nodes, the greater the dissimilarity of two nodes, the greater TijThe smaller; dijRepresenting the planar distance between two adjacent nodes.
(3) Constructing an adjacent matrix:
and (3) calculating the edge weights of 25 sensor nodes, wherein the edge weight formed by two adjacent sensors is calculated by a formula (5), the edge weight between two non-adjacent nodes is set to be infinite, which indicates that the two nodes cannot be reached, and finally the edge weight formed by 25 sensor nodes is stored by using an adjacency matrix.
Step (4), searching the optimal path by the improved Dijkstra algorithm: and (3) searching the optimal path of sensor point distribution by using the Dijkstra algorithm with the AP clustering center in the step (2) as a starting point, the convergent point of the wireless sensor network as an end point and the maximum data dissimilarity and the shortest distance as targets.
The method comprises the following specific implementation steps:
the method comprises the following steps: test environment and data acquisition
The area of the test site is about 16000m2About 190m in the transverse direction and about 85m in the longitudinal direction. The sensors are uniformly distributed. In order to ensure the full coverage of the sensing network of the test area, 5 sensor detection points are uniformly arranged in the transverse direction, 5 sensor detection points are also uniformly arranged in the longitudinal direction, and 25 detection points are arranged in total. The distribution of the sensor detection points is shown in fig. 2. The mass water content parameters of the soil 25cm below the soil surface are collected at each detection point.
Step two: and (4) according to the collected real-time data, making the spatial distribution of the soil mass water content by using a Kiring algorithm.
Step three: AP clustering algorithm determination starting point
The AP clustering algorithm is a semi-supervised clustering algorithm, and the number of classes does not need to be specified in advance. The algorithm is to look each data as a node in the graph, iteratively update according to the similarity between the nodes to find a clustering center, and the similarity of two data points is calculated by using the negative number of the distance, namely the closer the distance is, the greater the similarity is. The similarity is calculated according to the formula (1).
s(i,j)＝-||ri-rj||2,i≠j (1)
riAnd rjThe soil mass water content of the ith sensor node and the jth sensor node respectively. And taking the AP clustering center as a starting point of a Dijkstra algorithm searching path, and determining a convergent point of the wireless sensor network as an end point of the searching path.
Step four: calculating the dissimilarity degree of soil moisture content of sensor nodes
And (3) calculating the sum of the dissimilarity values between the data of each sensor node and the data of the rest other nodes, and reflecting the dissimilarity degree of soil moisture content of each node in the whole test area from the global perspective, wherein the dissimilarity degree calculation formula is shown as a formula (2).
Figure BDA0002184273410000071
Wherein n is the number of sensing nodes, riAnd rjRepresenting the soil mass water content, R, of the ith and jth sensor nodesiThe larger the value is, the larger the difference between the soil moisture content of the point and the soil moisture content of the whole test area is.
Step five: determining weight values of edges between sensor nodes
Calculating T according to the dissimilarity value of each sensing nodeij，TijIs equal to the sum of the reciprocal of the dissimilarity value of the sensor node i and the reciprocal of the dissimilarity value of the sensor node j; dijRepresents the planar distance between two sensor nodes, (x)i,yi) Representing plane coordinates. Edge weight W between two sensor nodesijFrom TijAnd DijAnd obtaining the result by weighted combination. The calculation formula is (3), (4) and (5).
Figure BDA0002184273410000073
Wij＝αTij+βDij(i,j＝1,2,...,n) (5)
Alpha and beta are weight factors, which represent T in the edge weight value WijAnd DijThe proportion of the components is calculated; t isijEqual to the sum of reciprocal of dissimilarity value of two nodes, the greater the dissimilarity of two nodes, the greater TijThe smaller; dijRepresenting the planar distance between two adjacent nodes.
Step six: constructing a contiguous matrix
And (3) calculating the edge weight values of 25 sensor nodes, wherein the edge weight value formed by two adjacent sensors is calculated by a formula (5), and the edge weight value between nonadjacent nodes is set to be infinite, which indicates that the two nodes cannot be reached. And finally, storing the edge weight value formed by 25 sensor nodes by using an adjacency matrix.
Step seven: searching shortest path using Dijkstra algorithm
Applying the adjacency matrix calculated in the step six to a Dijkstra algorithm, and taking F as the reference value according to the determined starting point and the determined end pointmin(ri,rj,xi,yi,xj,yj)＝αTij(ri,rj)+βDij(xi,yi,xj,yj) And searching a path for the objective function to finally obtain the optimal path.
Path optimization
On an MATLAB platform, firstly using an AP clustering algorithm to obtain a clustering center, secondly, determining an edge weight value by the weighted combination of the dissimilarity value of soil moisture content data and a network node path value, and constructing an adjacency matrix of a Dijkstra algorithm; and finally, searching the optimal path by using Dijkstra algorithm by taking the class center as a starting point, the convergent point of the wireless sensor network as an end point and the maximum data dissimilarity and the shortest distance as targets.
Analysis of Experimental results
The spatial distribution of the soil moisture obtained from the data collected in real time is shown in fig. 3, fig. 4 shows the AP clustering results of the soil moisture data, which are clustered into two classes, with the clustering centers being the sensor nodes S13 and S14, which serve as the starting points of Dijkstra algorithm search paths.
TABLE 1 Dijkstra Algorithm Path Total cost s
Figure BDA0002184273410000081
Figure BDA0002184273410000091
Table 1 shows the test results of data collected 6 times in two days, and it can be known from the table that the number of AP clustering centers of 6 times is 2, i.e., S13 and S14 sensing nodes, indicating that the test data has good consistency. And (3) taking the AP clustering center as a starting point and the convergence point S1 as an end point, and performing path search by using the improved Dijkstra algorithm to obtain 6 different path results. The results of the 6 tests can realize reliable data transmission only through 6 sensor nodes, namely after the path optimization, the number of the sensors can be reduced from 25 to 6, and the system cost is effectively reduced.
In order to obtain the optimal path with the maximum data dissimilarity and the shortest distance, the average dissimilarity and the total path cost are used for carrying out quantitative evaluation on the test result. As can be seen from table 1, the path calculated in test No. 1 is S14-S13-S12-S11-S6-S1, the total cost of the path is 0.3613, which is the smallest among 6 test results, and it indicates that the path is the shortest distance path, and the average degree of dissimilarity of the path is 25.0652, which is the largest among 6 test results, so that the data dissimilarity of each sensing node in the path is the largest, and the redundancy of the data is the lowest.
In addition, the path calculated by the test No. 5 is S14-S13-S12-S7-S2-S1, the total path cost is 0.4274, the maximum path is the maximum path in 6 test results, the average dissimilarity degree is 20.9017, the minimum path is the minimum path in 6 test results, and the path is the longest path, the data dissimilarity of each sensing node is the minimum path, and the data redundancy is large.
The total cost of the paths for trials nos. 2, 3, 4 and 6 and the average dissimilarity value are between the two paths for trials nos. 1 and 5. Therefore, the results of 6 tests are compared, and the paths S14-S13-S12-S11-S6-S1 calculated by the test No. 1 are the best paths.
The soil temperature and humidity sensor has the model of SWR-100W, integrates the soil moisture and soil temperature sensors together, is convenient for measurement and research of soil moisture content and soil temperature, and has the advantages of convenience in carrying, sealing, high precision and the like.

Claims (5)

1. An optimized layout method of soil moisture content sensors based on an APDJ algorithm is characterized by comprising the following steps: the method comprises the following specific steps:
(1) firstly, collecting soil moisture content data: arranging a plurality of soil moisture content sensors in an evenly distributed mode;
(2) and clustering soil moisture content data: collecting soil moisture content data through a plurality of soil moisture content sensors on the basis of full coverage of a sensing network, and clustering the collected data by using an AP (access point) clustering algorithm to obtain a clustering center;
(3) constructing an adjacency matrix of the Dijkstra algorithm: determining an edge weight value by the weighted combination of the dissimilarity value of the soil moisture content data and the network node path value, and constructing an adjacency matrix of the Dijkstra algorithm;
(4) and searching the optimal path by the improved Dijkstra algorithm: and (3) searching the optimal path of the soil moisture sensor distribution point by using the Dijkstra algorithm with the AP clustering center obtained in the step (2) as a starting point, the convergence point of the wireless sensor network as an end point and the maximum data dissimilarity and the shortest distance as targets.
2. The method for optimizing the layout of the soil moisture sensors based on the APDJ algorithm as claimed in claim 1, wherein: the method comprises the following steps that in the step (1), a plurality of soil moisture content sensors are arranged in an evenly distributed mode, and the method specifically comprises the following steps: 5 sensor detection points are uniformly arranged in the transverse direction and the longitudinal direction respectively, 25 detection points are arranged in total, and each detection point acquires the mass water content parameter of soil 25cm below the surface of the soil.
3. The method for optimizing the layout of the soil moisture sensors based on the APDJ algorithm as claimed in claim 2, wherein: and (2) clustering is realized by calculating a similarity matrix of the water content of the soil nodes to obtain a clustering center, which specifically comprises the following steps:
suppose { m }1，m2，...，mnAnd (4) a data sample set, wherein the clustering algorithm uses negative Euclidean distance to express the similarity among n data points, and the similarity is recorded as:
s(i，j)＝-||mi-mj||2，i≠j(1)
i. j respectively represents the ith and jth data points in the data sample set, n data points form an n multiplied by n similarity matrix S, and the clustering center is formed by the similarity matrix S [ S (i, j) ]]n×nAnd determining the numerical value of an element s (k, k) on the diagonal, wherein the s (k, k) is called a bias parameter p, the size of the p value influences the clustering result, the larger the p is, the more the clustering number is, and the smaller the p value is, the less the clustering number is.
4. The APDJ algorithm-based soil moisture content sensor optimization layout method according to claim 3, characterized in that: determining an edge weight value by the weighted combination of the dissimilarity value of the soil moisture content data and the network node path value, and constructing an adjacency matrix of the Dijkstra algorithm, wherein the method comprises the following specific steps:
1) calculating the dissimilarity degree of soil moisture content of the sensor nodes:
calculating the sum of dissimilarity values between the data of each sensor node and the data of the rest other nodes, reflecting the dissimilarity degree of soil moisture content of each node in the whole test area from the global perspective, wherein the dissimilarity degree calculation formula is as the formula (2):
Figure FDA0002184273400000021
wherein n is the number of sensing nodes, riAnd rjRepresenting the soil mass water content, R, of the ith and jth sensor nodesiThe larger the value is, the larger the difference between the soil moisture content of the point and the soil moisture content of the whole test area is;
2) determining the weight value of the edge between the sensor nodes:
each obtained according to the formula (2)The dissimilarity values of the sensing nodes and calculating Tij，TijIs equal to the sum of the reciprocal of the dissimilarity value of the sensor node i and the reciprocal of the dissimilarity value of the sensor node j; dijRepresents the planar distance between two sensor nodes, (x)i，yi) Representing plane coordinates, the edge weight W between two sensor nodesijFrom TijAnd DijThe weighted combination is obtained, and the calculation formulas are (3), (4) and (5):
Figure FDA0002184273400000023
Wij＝αTij+βDij(i，j＝1，2，...，n) (5)
alpha and beta are weight factors, which represent T in the edge weight value WijAnd DijThe proportion of the components is calculated; t isijEqual to the sum of reciprocal of dissimilarity value of two nodes, the greater the dissimilarity of two nodes, the greater TijThe smaller; dijRepresenting the plane distance between two adjacent nodes;
3) constructing an adjacent matrix:
and (3) calculating the edge weights of 25 sensor nodes, wherein the edge weight formed by two adjacent sensors is calculated by a formula (5), the edge weight between two non-adjacent nodes is set to be infinite, which indicates that the two nodes cannot be reached, and finally the edge weight formed by 25 sensor nodes is stored by using an adjacency matrix.
5. The APDJ algorithm-based soil moisture content sensor optimization layout method according to claim 4, characterized in that: searching an optimal path by the improved Dijkstra algorithm in the step (4), wherein the optimal path is as follows: applying the adjacency matrix calculated in the step (3) to a Dijkstra algorithm, and taking F as the reference value according to the starting point and the end point determined in the step (2)min(ri，rj，xi，yi，xj，yj)＝αTij(ri，rj)+βDij(xi，yi，xj，yj) And searching a path for the target function to obtain the optimal path.
CN201910808140.4A 2019-08-29 2019-08-29 Soil moisture content sensor optimization layout method based on APDJ algorithm Active CN110717507B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910808140.4A CN110717507B (en) 2019-08-29 2019-08-29 Soil moisture content sensor optimization layout method based on APDJ algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910808140.4A CN110717507B (en) 2019-08-29 2019-08-29 Soil moisture content sensor optimization layout method based on APDJ algorithm

Publications (2)

Publication Number Publication Date
CN110717507A true CN110717507A (en) 2020-01-21
CN110717507B CN110717507B (en) 2023-03-14

Family

ID=69209493

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910808140.4A Active CN110717507B (en) 2019-08-29 2019-08-29 Soil moisture content sensor optimization layout method based on APDJ algorithm

Country Status (1)

Country Link
CN (1) CN110717507B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112291760A (en) * 2020-11-18 2021-01-29 同济大学 Underground wireless sensor network node energy consumption optimization method based on LoRa

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2418756A1 (en) * 2002-02-12 2003-08-12 Yusuke Fujita Dual dijkstra search for planning multiple paths
CN105825433A (en) * 2016-04-01 2016-08-03 北京邮电大学 Soil moisture status determining method and apparatus
CN110059881A (en) * 2019-04-19 2019-07-26 安徽农业大学 The method of novel tea garden soil soil moisture content sensor optimization layout based on clustering algorithm

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2418756A1 (en) * 2002-02-12 2003-08-12 Yusuke Fujita Dual dijkstra search for planning multiple paths
CN105825433A (en) * 2016-04-01 2016-08-03 北京邮电大学 Soil moisture status determining method and apparatus
CN110059881A (en) * 2019-04-19 2019-07-26 安徽农业大学 The method of novel tea garden soil soil moisture content sensor optimization layout based on clustering algorithm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张武等: "基于近邻传播算法的茶园土壤墒情传感器布局优化", 《农业工程学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112291760A (en) * 2020-11-18 2021-01-29 同济大学 Underground wireless sensor network node energy consumption optimization method based on LoRa
CN112291760B (en) * 2020-11-18 2021-11-09 同济大学 Underground wireless sensor network node energy consumption optimization method based on LoRa

Also Published As

Publication number Publication date
CN110717507B (en) 2023-03-14

Similar Documents

Publication Publication Date Title
CN108696932B (en) Outdoor fingerprint positioning method using CSI multipath and machine learning
CN103476118B (en) A kind of WLAN indoor location fingerprint positioning method for monitoring in real time
CN110334391B (en) Automatic planning method for collecting circuit of multi-dimensional constraint wind power plant
CN103747419B (en) A kind of indoor orientation method based on signal strength difference and dynamic linear interpolation
CN106102163B (en) WLAN fingerprint positioning method based on RSS linear correlation Yu secondary weighted centroid algorithm
CN109813273A (en) A kind of agricultural machinery repetition working area determination method based on spatial analysis
CN102638889A (en) Indoor wireless terminal positioning method based on Bayes compression sensing
CN103763769B (en) Based on the indoor fingerprint positioning method that access point reselection procedure and self-adapting cluster divide
JP2009065394A (en) Attenuation constant estimation system for sensor network, node position estimation system, estimating method and program thereof
CN109547929B (en) Distributed sensor node positioning method based on conjugate gradient method
CN110059881A (en) The method of novel tea garden soil soil moisture content sensor optimization layout based on clustering algorithm
CN108876136A (en) Recommend the attack of terrorism methods of risk assessment of innovatory algorithm based on position
CN107179525A (en) A kind of location fingerprint construction method of the Kriging regression based on Thiessen polygon
CN112954594A (en) Wireless sensor network node positioning algorithm based on artificial bee colony
CN110213710A (en) A kind of high-performance indoor orientation method, indoor locating system based on random forest
CN103249144A (en) C-type-based wireless sensor network node location method
CN110717507B (en) Soil moisture content sensor optimization layout method based on APDJ algorithm
CN117241215A (en) Wireless sensor network distributed node cooperative positioning method based on graph neural network
Rabhi et al. An improved method for distributed localization in WSNs based on fruit fly optimization algorithm
Hu et al. Phenotyping of poplar seedling leaves based on a 3D visualization method
CN113919185A (en) Method and device for measuring landform and landform conditions
CN108287940B (en) Method and device for arranging sampling points, electronic equipment and storage medium
CN117627001A (en) Vegetation slope protection establishment method for preventing water and soil loss and fixing sand
CN110297212B (en) Outdoor grouping test positioning method and system based on Voronoi diagram
CN109859323B (en) Method for weighting forest stand space pattern based on triangular network model

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