Wireless sensor network node deployment optimization system based on combinatorial mathematical coverage

Abstract

The invention discloses a wireless sensor network node deployment optimization system based on combinatorial mathematical coverage in the technical field of wireless sensing, and the system comprises a combination module which is used for deploying nodes in a forest two-dimensional plane graph based on a single network node detection range, and generating a basic node distribution diagram containing node positions and a detection range; the blind area judgment module is used for acquiring the position and diameter of the tree, filling interference points with interference round marks with corresponding diameters, connecting adjacent interference points to form a polygonal area, calculating the proportion of the interference round marks of each side line of the area, and judging whether the area is a detection blind area or not according to a preset strategy; and the optimization module is used for adding nodes in the blind area to realize coverage according to the position of the blind area in the basic node distribution map. According to the invention, the combination module distributes nodes and generates a graph, the blind area module judges a blind area and the optimization module adds nodes, so that the blind area formed by tree shielding in a forest is reduced, early warning is performed in time, and reasonable deployment and effective monitoring are ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless sensing technology, in particular to a wireless sensor network node deployment optimization system based on combinatorial mathematics coverage. BACKGROUND

[0002] The node deployment of a wireless sensor network (WSN) is a key link in the scientific planning of the spatial position, quantity ratio, and network topology of sensor nodes with sensing, communication, and data processing capabilities according to monitoring requirements and target area characteristics, and is also a basic prerequisite for the effective implementation of the monitoring function of WSN. This process is not simply a node placement, but needs to balance multiple dimensions around the core target: it needs to ensure that there is no coverage blind area in the monitoring area, and it needs to achieve balanced node energy consumption to prolong the overall network life, while also ensuring the communication connectivity between nodes, controlling the deployment cost, and adapting to complex environmental factors such as extreme temperature and humidity, electromagnetic interference, etc.

[0003] With the evolution of technology, existing deployment schemes often introduce combinatorial mathematical models to analyze node coverage problems, striving to reduce the number of nodes deployed to control costs while meeting coverage requirements, but the adaptability in complex scenarios is still insufficient. WSN node deployment is widely used in environmental monitoring, industrial automation, military reconnaissance, etc., among which in the field of environmental monitoring, it is often used for monitoring air quality, soil humidity, temperature and humidity changes, and forest fire scenarios.

[0004] However, in a dense forest scenario, although the traditional static combinatorial mathematical model can better utilize the detection range of a single node through the arrangement and combination of WSN nodes, due to the three-dimensional shielding of multiple trees, one or more detection blind areas will still be formed in some coverage areas. Once a fire occurs in these detection blind areas, the infrared radiation generated by the fire source will be difficult to be captured by adjacent nodes due to the shielding of tree trunks; even worse, the smoke will be intercepted by the tree canopy, delaying its diffusion to the effective monitoring range, ultimately causing a delay of several minutes to several tens of minutes in fire warning, directly missing the key window for initial extinguishing, and making it difficult to meet the core needs of "early detection and early disposal" of forest fires.

[0005] Therefore, the present application proposes a wireless sensor network node deployment optimization system based on combinatorial mathematics coverage to solve the above problems. SUMMARY

[0006] To solve the above problems, the present application provides a wireless sensor network node deployment optimization system based on combinatorial mathematics coverage, which identifies and judges the blind areas formed in the monitoring range due to the surrounding of trees, and then deploys network nodes in these blind areas, so as to timely warn and ensure reasonable deployment and effective monitoring.

[0007] In order to achieve the above object, the technical scheme of the present application is as follows: a wireless sensor network node deployment optimization system based on combinatorial mathematics coverage comprises: a combination module for optimizing network node deployment in a two-dimensional planar graph of a forest based on the range of a single network node and a combinatorial mathematics coverage subset optimization algorithm, and generating a basic node distribution map showing the position of each network node and the corresponding detection range; a blind area judgment module for obtaining the corresponding positions of trees in the forest in the two-dimensional planar graph and marking them as interference points; simultaneously obtaining the diameter information of each tree and filling the interference circle marks corresponding to the diameter of the tree at the corresponding interference point position; then according to the actual growth distribution law of the trees in the forest, sequentially and continuously associating adjacent interference points in the two-dimensional planar graph according to the spatial position correlation to form a plurality of polygonal regions; for each edge line of each polygonal region, calculating the total length of the projection of all interference circle marks on the edge line in the edge length direction, comparing and operating with the total length of the edge line to obtain the interference circle mark proportion of each edge line, and judging whether the polygonal region is a detection blind area according to a preset judgment strategy; an optimization module for adding network nodes in the detection blind area based on the distribution of the detection blind area in the basic node distribution map to cover the detection blind area and output the optimized node distribution result.

[0008] Further, the preset judgment strategy is as follows: When the interference circle mark proportion of all edge lines in the polygonal region is greater than or equal to 70%, it is determined that the polygonal region is a detection blind area; otherwise, it is determined to be a non-detection blind area.

[0009] Further, the optimization module is further used to obtain the polygon boundary coordinates and the corresponding area of the detection blind area after marking the detection blind area, and calculate the maximum coverage range of a single network node in the detection blind area. If the maximum coverage range of a single network node can completely cover the detection blind area, it is determined that no new node is needed; otherwise, through the coverage subset optimization algorithm of combinatorial mathematics, the number of nodes required to cover the detection blind area and the optimal point coordinates of each node are calculated with the "minimum number of nodes" as the target until the detection blind area is completely covered.

[0010] Further, the number of edge lines of the polygonal region is 3-10.

[0011] Further, the blind area judgment module is further used for merging processing of a plurality of interference circle marks with close distances in the forest two-dimensional plane map; when the plurality of interference circle marks are merged, the two circle marks farthest apart are first screened, a straight line distance between the center points of the two circle marks is calculated, the straight line distance is added to the radius values of the two interference circle marks respectively, and a diameter of a new interference circle mark after merging is obtained; the center point coordinates of the new interference circle mark are taken as the midpoint of the line connecting the center points of the two circle marks farthest apart; after the merging is completed, the new interference circle mark is used to replace all the interference circle marks in the original merging group, and interference point data is updated, and the new interference point is the center point of the new interference circle mark.

[0012] Further, in the blind area judgment module, the merging trigger condition is: When a straight line distance between edges of any two interference circle marks ≤ a preset merging interval threshold, the two interference circle marks are included in the same merging group.

[0013] Further, the combination module specifically includes a parameter configuration subunit, a coverage model construction subunit, an optimization solving subunit and a deployment adjustment subunit. The parameter configuration subunit is used for acquiring geographical boundary data of the forest two-dimensional plane map and basic performance parameters of the network nodes; wherein the basic performance parameters include a standard detection radius and a standard communication radius of the nodes. The coverage model construction subunit is used for discretizing the forest two-dimensional plane area to be monitored into a plurality of square grid units based on a set covering algorithm of combinatorics, forming a set of elements to be covered; and taking a set of grid units covered by a circle with a grid node deployment position as the center and the detection radius as the radius as a covering subset, a covering subset family is constructed. The optimization solving subunit is used for solving the covering subset family by adopting a greedy algorithm with “the least number of nodes covering all grid units” as the target.

[0014] The deployment adjustment subunit is used for adjusting the network node positions according to the node positions output by the optimization solving subunit, and generating a basic node distribution map containing the network node positions and corresponding detection ranges.

[0015] Further, the parameter configuration subunit sets an area with a distance <0.3 m from any interference point center point as a node forbidden area, and synchronizes to the coverage model construction subunit.

[0016] Further, when the blind area judgment module forms a polygonal area according to the spatial position correlation of adjacent interference points, a K-nearest neighbor clustering and boundary fitting combined mode is adopted to determine the spatial position correlation.

[0017] Further, for each interference point, the K closest other interference points are screened to form an initial association group; for each association group, a triangular mesh is generated by a Delaunay triangulation algorithm, and the common edges of adjacent triangular meshes are merged to form a polygonal region containing all the interference points of the association group; wherein the value of K is dynamically adjusted according to the average distance between forest trees, K=3 when the distance is less than 5m, and K=5 when the distance is greater than or equal to 5m.

[0018] The above scheme has the following beneficial effects: the scheme can solve the core pain points of traditional wireless sensor network node deployment in forest scenarios, and realize multiple optimization of monitoring efficiency, resource utilization and scene adaptation. The system combines the set covering algorithm and the greedy solving logic to discretize the forest area into grid units, generates a basic node distribution map with the goal of "covering the whole area with the least nodes", reduces invalid deployment from the source, controls hardware and maintenance costs, and reduces the energy imbalance problem caused by excessive nodes; the 0.3m node forbidden area set by the parameter configuration subunit further reduces the node shielding or damage caused by being close to the tree, ensuring that each type of node can play an effective monitoring role. In the blind area processing, the blind area judgment module quantizes the shielding range corresponding to the tree diameter through the interference circle, combines K-neighbor clustering (the value of K is dynamically adjusted according to the distance between trees) and Delaunay triangulation to construct a polygonal region, and then identifies the blind area with the clear standard that the edge line interference circle mark ratio is greater than or equal to 70%, restores the monitoring dead angle formed by the three-dimensional shielding of the tree; the optimization module first judges the single node coverage capability for the blind area, and then calculates the minimum number of supplementary points and the optimal coordinates through the combination mathematical algorithm, which not only eliminates the blind area and reduces the early warning lag caused by the shielding of infrared radiation of fire source and the delay of smoke diffusion, but also meets the needs of early discovery and early disposal of forest fires, and does not blindly increase points to cause waste. The overall scheme is supported by a quantitative model, the process is standardized, and the key parameters can be adapted to different forest types, taking into account monitoring effectiveness, cost controllability and network longevity, providing a reliable solution for wireless sensor network deployment in complex forest environments.

[0019] Additional aspects and advantages of the application will be set forth in part in the description which follows, and in part will become apparent to those skilled in the art upon examination of the following and / or can be learned by practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The flowchart of the embodiment of the wireless sensor network node deployment optimization system based on combination mathematics coverage of the application is shown. DETAILED DESCRIPTION

[0021] The technical solutions of the present application will be described clearly and completely below in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0022] The specific embodiments are further described in detail below: Embodiment 1

[0023] As shown in the wireless sensor network node deployment optimization system based on combinatorial mathematics coverage includes: combination module, blind area judgment module and optimization module; Figure 1 Specifically, the combination module specifically includes a parameter configuration subunit, a coverage model construction subunit, an optimization solving subunit and a deployment adjustment subunit. The parameter configuration subunit is used to obtain the geographical boundary data of the forest two-dimensional plane and the basic performance parameters of the network node; wherein the basic performance parameters include the standard detection radius and the standard communication radius of the node. The coverage model construction subunit is used to discretize the forest two-dimensional plane area to be monitored into a plurality of square grid units (for example, 0.3*0.3m) based on the set covering algorithm of combinatorial mathematics, forming a set of elements to be covered; and then taking each grid node deployment position as the center and the detection radius as the radius, the grid unit set covered by the circle is taken as the covering subset, and the covering subset family is constructed. The optimization solving subunit is used to adopt a greedy algorithm to solve the covering subset family with the target of “the minimum number of nodes covering all grid units”. The deployment adjustment subunit is used to adjust the network node position according to the node position output by the optimization solving subunit, and generate a basic node distribution map containing the network node position and the corresponding detection range.

[0024] ​Specifically, the blind area judgment module is configured to obtain corresponding positions of trees in a forest in a two-dimensional plan view, and mark the positions as interference points; obtain diameter information of each tree, and fill interference circle marks corresponding to the diameter of the tree at the corresponding interference point positions; according to actual growth distribution rules of the trees in the forest, sequentially and continuously arrange adjacent interference points in the two-dimensional plan view according to spatial position correlation to form a plurality of polygonal regions (the number of side lines of the polygonal regions is 3-10); for each side line of each polygonal region, calculate a total length of projections of all interference circle marks on the side line in a length direction of the side line, compare the total length of the projections with a total length of the side line, and obtain an interference circle mark proportion of each side line; and according to a preset judgment strategy, determine whether the polygonal region is a detection blind area. The preset judgment strategy is specifically as follows: when the interference circle mark proportions of all side lines in the polygonal region are all greater than or equal to 70%, the polygonal region is determined to be a detection blind area; otherwise, the polygonal region is determined to be a non-detection blind area.

[0025] Secondly, in the blind area judgment module, a plurality of interference circle marks with close distances in the two-dimensional plan view of the forest are processed by merging; when the plurality of interference circle marks are merged, two interference circle marks with the farthest distance are first selected, a straight line distance between the centers of the two interference circle marks is calculated, the straight line distance is added to the radii of the two interference circle marks, and a diameter of a new interference circle mark after merging is obtained; a center point coordinate of the new interference circle mark is the midpoint of a line connecting the centers of the two interference circle marks with the farthest distance; after the merging is completed, the new interference circle mark is used to replace all the interference circle marks in an original merging group, and interference point data is updated, and a new interference point is the center point of the new interference circle mark. A merging trigger condition is that when a straight line distance between edges of any two interference circle marks is less than or equal to a preset merging interval threshold (1 / 10 of a node sensing radius), the two interference circle marks are included in a same merging group.

[0026] When the blind area judgment module forms the polygonal region according to the spatial position correlation of the adjacent interference points, a K-nearest neighbor clustering and boundary fitting combined mode is used to determine the spatial position correlation, and the mode is specifically as follows: For each interference point, the K nearest other interference points are selected to form an initial correlation group; for each correlation group, a triangular mesh is generated by using a Delaunay triangulation algorithm, and common edges of adjacent triangular meshes are merged to form a polygonal region containing all interference points of the correlation group; wherein the value of K is dynamically adjusted according to an average distance of the trees in the forest, K=3 when the distance is less than 5 m, and K=5 when the distance is greater than or equal to 5 m.

[0027] Specifically, the optimization module is configured to, after detecting the blind area, acquire polygon boundary coordinates and a corresponding area of the detection blind area, calculate a maximum coverage range of a single network node in the detection blind area, and determine that no new node is needed if the maximum coverage range of the single network node can completely cover the detection blind area. Otherwise, a covering subset optimization algorithm of combinatorics is used to calculate a number of nodes and optimal distribution coordinates of each node required to cover the detection blind area until the detection blind area is completely covered, and an optimized node distribution result is output.

[0028] Based on the above scheme, the following experimental comparison is made: Experimental area: Two typical sample plots in a temperate forest are selected. Sample plot A is a dense tree area (average tree spacing of 4 m, meeting the adjustment condition of K=3), and sample plot B is a relatively sparse tree area (average tree spacing of 6 m, meeting the adjustment condition of K=5). Both sample plots have an area of 100 m x 100 m and contain broadleaf trees and coniferous trees of different diameters (5-30 cm), simulating the distribution characteristics of real forest trees.

[0029] Comparison scheme: The experiment is divided into two groups. The control group uses a traditional static combinatorial mathematical deployment scheme (only calculates the minimum number of nodes based on the sensing range of the node, without processing the blind area caused by tree obstruction), and the experimental group uses the present scheme.

[0030] Experimental data: Table 1-Comparison of related parameters Experimental scheme Plot type Total number of nodes Number of detection blind spots Simulation fire source identification rate (%) Fire warning lag time (s) Network expected lifetime (months) Traditional scheme (control) Dense tree area (A) 28 12 60.7 2400 18 Inventive scheme (experiment) Dense tree area (A) 32 0 96.6 350 22 Traditional scheme (control) Sparse tree area (B) 22 8 86.8 1800 20 Inventive scheme (experiment) Sparse tree area (B) 25 0 100 280 25 Combining table 1 can be obtained, the number of detection blind areas of the experimental group (the scheme of the application) in the dense and sparse areas of trees is 0, while the control group has 12 and 8 blind areas respectively, indicating that the application can solve the monitoring loopholes caused by the three-dimensional occlusion of trees in the traditional scheme through the logic of "interference circle quantization of occlusion range + polygon area judgment of blind area + supplementary node coverage", realize full-area dead-angle-free coverage, and the recognition rate of simulated fire source points reaches 96.6%-100%, which is much higher than that of the control group of 60.7%-86.6%. Compared with the control group, the total number of nodes of the experimental group only increases by 14%-18% (from 28 to 32 in the dense area and from 22 to 25 in the sparse area), but the blind area is reduced and the early warning efficiency is improved; at the same time, the expected life of the network of the experimental group is 22%-25% longer than that of the control group, because the application only supplements the nodes in the blind area as needed, reduces the energy waste caused by the "blind increase of points" in the traditional scheme, and reduces the deployment of invalid nodes by setting the node forbidden area, thereby reducing the data transmission conflict. In the two sample plots with different tree densities, the application can accurately construct the polygon occlusion area by dynamically adjusting the K value (K=3 in the dense area and K=5 in the sparse area), and the fire warning lag time is shortened to 280-350s, which is only 15%-19% of that of the control group (1800-2400s), fully meeting the core needs of "early discovery and early disposal" of forest fires, and verifying the adaptation ability of the system to different forest scenarios with different tree densities.

[0031] Obviously, the above embodiments are only examples for the purpose of clarity, and are not limitations of the embodiments. Based on the above description, those skilled in the art can make other different forms of changes or modifications. Here, all the embodiments need not and cannot be exhausted. The obvious changes or modifications derived therefrom are still within the protection scope of the present application.

Claims

1. A wireless sensor network node deployment optimization system based on combinatorial mathematical coverage, characterized in that, include: The ensemble module is used to deploy network nodes in a two-dimensional planar graph of the forest based on the detection range of a single network node and a combinatorial mathematics-based coverage subset optimization algorithm, generating a basic node distribution map; the basic node distribution map shows the location of each network node and its corresponding detection range; The blind zone detection module is used to obtain the corresponding positions of trees in the forest on a two-dimensional planar map and mark them as interference points; at the same time, it obtains the diameter information of each tree and fills the corresponding interference point position with interference circles corresponding to the diameter of the tree; then, according to the actual growth distribution pattern of trees in the forest, it arranges adjacent interference points in an orderly and continuous manner on the two-dimensional planar map according to their spatial position correlation to form several polygonal regions; for each edge of each polygonal region, it calculates the total projection length of all interference circles on the edge in the side length direction, compares it with the total length of the edge, obtains the proportion of interference circles on each edge, and determines whether the polygonal region is a detection blind zone according to a preset judgment strategy; The optimization module is used to add network nodes within the detection blind zone based on the distribution of the detection blind zone in the basic node distribution map, thereby covering the detection blind zone and outputting the optimized node distribution result.

2. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 1, characterized in that: The preset judgment strategy is as follows: When the proportion of interference circles on all edges within a polygonal region is ≥70%, the polygonal region is determined to be a detection blind zone; otherwise, it is determined to be a non-detection blind zone.

3. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 2, characterized in that: The optimization module is also used to obtain the polygon boundary coordinates and corresponding area of ​​the detection blind zone after marking the detection blind zone, and to calculate the maximum coverage of a single network node in the detection blind zone; If the maximum coverage of a single network node can completely cover the detection blind zone, then it is determined that no new node is needed; otherwise, the coverage subset optimization algorithm of combinatorial mathematics is used to calculate the number of nodes required to cover the detection blind zone and the optimal placement coordinates of each node with the goal of "minimum number of nodes" until the detection blind zone is completely covered.

4. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 3, characterized in that: The polygonal region has 3-10 edges.

5. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 4, characterized in that: The blind zone detection module is also used to merge several interferometers that are close to each other in the two-dimensional planar map of the forest. When merging interferometers, the two interferometers that are farthest apart are first selected, the straight-line distance between the center points of the two interferometers is calculated, and then the straight-line distance is added to the radius value of each of the two interferometers to obtain the diameter of the new interferometer after merging. The center point coordinate of the new interferometer is taken as the midpoint of the line connecting the center points of the two interferometers that are farthest apart. After merging, all the interferometers in the original merged group are replaced with the new interferometer, and the interferometer data is updated. The new interferometer is the center point of the new interferometer.

6. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 5, characterized in that: In the blind spot detection module, the merged trigger condition is: When the straight-line distance between the edges of any two interference circles is less than or equal to the preset merging spacing threshold, the two interference circles are included in the same merging group.

7. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 6, characterized in that: The combined module specifically includes a parameter configuration subunit, a coverage model construction subunit, an optimization solution subunit, and a deployment and adjustment subunit; The parameter configuration subunit is used to obtain the geographic boundary data of the forest two-dimensional planar map and the basic performance parameters of the network nodes; among which, the basic performance parameters include the standard detection radius and standard communication radius of the nodes; The coverage model constructs sub-units for a set coverage algorithm based on combinatorial mathematics. It discretizes the two-dimensional planar region of the forest to be monitored into several square grid units to form a set of elements to be covered. Then, the set of grid units covered by the circle with the deployment position of each grid node as the center and the detection radius as the radius is used as the coverage subset to construct a family of coverage subsets. The optimized solution for the subcell is achieved using a greedy algorithm, with the objective of "minimum number of nodes covering all grid cells" to solve for the covered subset family. The deployment adjustment subunit is used to adjust the network node positions based on the node positions output by the optimization solution subunit, and generate a basic node distribution map containing the network node positions and their corresponding detection ranges.

8. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 7, characterized in that: The parameter configuration subunit sets the area with a distance of less than 0.3m from the center point of any interference point as the node no-distribution zone and synchronizes it to the coverage model construction subunit.

9. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 8, characterized in that: When the blind zone detection module forms polygonal regions based on the spatial location correlation of adjacent interference points, it uses a combination of K-nearest neighbor clustering and boundary fitting to determine the spatial location correlation.

10. The wireless sensor network node deployment optimization system based on combinatorial mathematical coverage according to claim 9, characterized in that: The spatial correlation is determined as follows: For each interference point, the K nearest other interference points are selected to form an initial correlation group; for each correlation group, a triangular mesh is generated using the Delaunay triangulation algorithm, and the common edges of adjacent triangular meshes are merged to form a polygonal region containing all interference points of the correlation group; among them, the value of K is dynamically adjusted according to the average spacing of forest trees, K=3 when the spacing is <5m, and K=5 when the spacing is ≥5m.

Images