Low-power-consumption routing control method based on energy proximity and DBSCAN algorithm

Abstract

The invention discloses a low-power-consumption routing control method based on proximity of energy and a DBSCAN algorithm, which can reduce energy consumption so as to prolong the service life of a network. The method comprises the following specific steps: firstly, clustering sensor nodes through a density-based DBSCAN algorithm; then, the local density of the sum of the shared proximity with respect to the energy with respect to the energy is redefined using the shared proximity with respect to the energy, and a cluster head is selected in combination with the relative position and the local density with respect to the energy. And finally, selecting the most suitable routing node to which the next hop arrives by adopting a formula screening method according to the distance, and transmitting the data to the base station by multiple jumps. According to the method, the network life cycle is prolonged to a certain extent. According to the invention, the sharing proximity about the energy is provided, and the consideration of the energy is increased compared with the original pure sharing proximity. If the sum of the shared proximity of one point and each point in a cluster is large, but the energy of nearby points is small, the communication efficiency is low after the points are selected as cluster heads.

Description

Technical Field

[0001] The invention belongs to the field of wireless sensor network routing and is a low-power consumption routing control method based on energy proximity and DBSCAN algorithm. Background Art

[0002] Nowadays, the Internet of Things is developing rapidly, and wireless sensors are widely used in industrial production, traffic control, medical treatment, experiments, communications, etc. However, the development of wireless sensor networks is restricted due to the limited storage capacity, energy and computing power of nodes.

[0003] In the LEACH protocol, cluster head nodes are randomly selected in a cyclic manner so that the energy load of the entire network is evenly distributed to each node, thereby reducing the energy consumption of the network and increasing the overall survival time of the network.

[0004] However, the LEACH protocol has certain limitations. First, in the assumptions of the LEACH protocol, it is assumed that all nodes have the computing power to support the MAC protocol. This makes the LEACH protocol unsuitable for large-scale wireless sensor networks. Second, the LEACH protocol assumes that "when the cluster head is initially selected, each node has the same energy." In addition, the protocol does not specify how the number of cluster head nodes is distributed, and there is a large degree of randomness in the selection of cluster heads. Summary of the invention

[0005] In order to solve the technical problems mentioned in the above-mentioned back-end technology, the present invention proposes a low-power routing control method based on energy proximity and DBSCAN algorithm, which considers node energy on the basis of considering node proximity. If the sum of the shared proximity of a point and each point in the cluster is large, but the energy of the nearby points is small, the communication efficiency will be low after being selected as the cluster head. Therefore, the shared proximity of energy proposed by the present invention improves the communication efficiency.

[0006] In order to achieve the above technical objectives, the technical solution of the present invention is:

[0007] A low-power routing control method based on energy proximity and DBSCAN algorithm comprises the following steps:

[0008] (1) The base station analyzes the data of each node and divides all nodes into core points and noise points through the DBSCAN algorithm. Clustering is completed according to the number of points in the area. The specific steps are as follows:

[0009] (101) The base station sets parameters and performs initialization. Let eps be the scanning radius and minPts be the minimum number of included points. Initially, all nodes are assumed to be unvisited.

[0010] (102) Randomly select an unvisited point, starting from this point, find all nearby points whose distance to this point is within (including) eps, and mark this point as visited.

[0011] (103) Compare the number of nearby points of the point with minPts. If the number of nearby points is ≥ minPts, mark the point as a core point, initialize a new cluster A, and add the point and the points within eps (including eps) from the point to cluster A; if the number of nearby points is ≤ minPts, but it is in the neighborhood of the core point, mark it as a boundary point and assign it to the cluster where the core point is located; if the number of nearby points is ≤ minPts and it does not belong to the neighborhood of any core point, mark the point as a noise point.

[0012] (104) Expand the cluster. For each point Q in cluster A, obtain the number of points in its eps neighborhood. If the number of points in the eps neighborhood of point P ≥ minPts, add the points in the eps neighborhood of P to cluster A.

[0013] (105) Repeat step (104) until cluster A cannot be expanded.

[0014] (106) Repeat steps (102)-(105) until all points have been visited.

[0015] (107) Noise points are clustered separately.

[0016] (108) Return the cluster list.

[0017] (2) Using the shared proximity of energy, redefine the sum of shared proximity of energy s to optimize the selection of cluster heads. The specific steps are as follows:

[0018] (201) For each node in the cluster, find the K nearest neighbors (KNN), where the K nearest neighbors are the K nodes closest to the current node.

[0019] (202) Calculate the energy-related shared nearest neighbor SNN of each node and other nodes, and the formula is:

[0020] SNN(i,j)=KNN(i)∩KNN(j) (2-1)

[0021] Among them, KNN(i) is the K nearest neighbors of node i, and SNN(i, j) is the shared nearest neighbors of node i and node j.

[0022] (203) Calculate the total residual energy SRE of the shared nearest neighbors. The calculation method of the total residual energy SRE(i, j) of the shared nearest neighbors of nodes i and j is:

[0023] SRE(t,j)=∑RE(m),m∈SNN(i,j) (2-2)

[0024] Among them, RE(m) is the residual energy of node m.

[0025] (204) Define the shared proximity with respect to energy. The shared proximity with respect to energy of any two nodes i and j in the network is calculated as follows:

[0026]

[0027] Among them, d ij is the distance from node i to node j, knnd ij is the sum of the distances from node j and its K nearest neighbors to node i, i.e. knnd ij =∑ p∈KNN(j)∪j d ip .

[0028] (205) Calculate the total shared proximity s and relative distance λ of each node in the cluster with respect to energy, as follows:

[0029] s i =ΣESnd(i, j) (2-4)

[0030]

[0031] Among them, s i is the sum of the shared proximity between node i and all nodes in the cluster regarding energy, d ij is the distance between node i and node j, ESnd(i, j) is the shared proximity of nodes i and j with respect to energy, and λ i is the relative distance of node i, indicating the distance from the point closest to node i among the points near the point whose local energy density is greater than its own to node i.

[0032] (206) Calculate the decision values ​​μ of all nodes in the cluster, take the point with the highest decision value, and use it as the cluster head node. The formula is:

[0033]

[0034] in, is the average value of the local density of energy of the nodes in the cluster, is the average value of the relative distances of the nodes within the cluster, α and β are weight coefficients, satisfying α+β=1, which can be adjusted according to actual needs.

[0035] (3) Select the route. The steps are as follows:

[0036] (301) If there is no candidate routing node near the cluster head node, data is transmitted directly to the base station. If there is a candidate routing node near the cluster head node, the distance from the cluster head node to the base station is compared with the threshold d0 set by the first-order radio model. If the distance from the cluster head node to the base station is less than or equal to the threshold d0, data is transmitted to the base station in a single hop manner; if the distance from the cluster head node to the base station is greater than the threshold d0, the cluster head will establish a selection function based on the total residual energy SRE, angle θ and distance d of the routing nodes near it, and then calculate the selection function value of each nearby routing node, and select the node with the largest selection function value as the next hop routing node. The established selection function is as follows:

[0037]

[0038] Among them, d s is the distance between the candidate routing node and the base station, and cosθ is the cosine value of the angle between the line connecting the current node and the base station and the line connecting the current node and the candidate node.

[0039] The positive effects of adopting the above solution are:

[0040] (1) The method optimizes the clustering and cluster head selection method. The present invention realizes the coordinated operation of the base station, and selects the cluster head by using the nearest neighbor method and comprehensively considering the energy of the node on the basis of clustering using the DBSCAN algorithm.

[0041] (2) Multi-hop routing is adopted between clusters, which comprehensively considers the energy and distance of routing nodes and selects the next hop with the lowest energy consumption. This reduces the energy transmitted between clusters to a certain extent, thereby extending the network life. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is the node clustering graph of the present invention

[0043] Figure 2 This is a schematic diagram of the cluster head of the present invention.

[0044] Figure 3 Schematic diagram of routing selection of the present invention DETAILED DESCRIPTION

[0045] The technical solution of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0046] A low-power routing control method based on energy proximity and DBSCAN algorithm comprises the following steps:

[0047] Step 1: Cluster all nodes in the entire network using the DBSCAN algorithm. The clustering results are as follows: Figure 1 shown.

[0048] Step 2: For each cluster, use the shared neighbors and node energy to redefine the total shared proximity s about energy, and then combine the relative distance to optimize the cluster head selection scheme. The cluster head selection results are shown in Figure 2 shown.

[0049] Step 3: When selecting the route in the data transmission stage, the distance is combined and the formula screening method is used to jump to the next hop routing node and transmit it back to the base station in a multi-hop or single-hop manner, such as Figure 3 shown.

[0050] The specific process of step one is:

[0051] (101) The base station sets parameters and performs initialization. Let eps be the scanning radius and minPts be the minimum number of included points. Initially, all nodes are assumed to be unvisited.

[0052] (102) Randomly select an unvisited point, starting from this point, find all nearby points whose distance to this point is within (including) eps, and mark this point as visited.

[0053] (103) Compare the number of nearby points of the point with minPts. If the number of nearby points is ≥ minPts, mark the point as a core point, initialize a new cluster A, and add the point and the points within eps (including eps) from the point to cluster A; if the number of nearby points is ≤ minPts, but it is in the neighborhood of the core point, mark it as a boundary point and assign it to the cluster where the core point is located; if the number of nearby points is ≤ minPts and it does not belong to the neighborhood of any core point, mark the point as a noise point.

[0054] (104) Expand the cluster. For each point Q in cluster A, obtain the number of points in its eps neighborhood. If the number of points in the eps neighborhood of point P ≥ minPts, add the points in the eps neighborhood of P to cluster A.

[0055] (105) Repeat step (104) until cluster A cannot be expanded.

[0056] (106) Repeat steps (102)-(105) until all points have been visited.

[0057] (107) Noise points are clustered separately.

[0058] (108) Return the cluster list.

[0059] The specific process of step 2 is as follows:

[0060] (201) For each node in the cluster, find the K nearest neighbors (KNN), where the K nearest neighbors are the K nodes closest to the current node.

[0061] (202) Calculate the energy-related shared nearest neighbor SNN of each node and other nodes, and the formula is:

[0062] SNN(t,j)=KNN(i)∩KNN(j) (2-1)

[0063] Among them, KNN(i) is the K nearest neighbors of node i, and SNN(i, j) is the shared nearest neighbors of node i and node j.

[0064] (203) Calculate the total residual energy SRE of the shared nearest neighbors. The calculation method of the total residual energy SRE(i, j) of the shared nearest neighbors of nodes i and j is:

[0065] SRE(tj)=∑RE(m),m∈SNN(t,j) (2-2)

[0066] Among them, RE(m) is the residual energy of node m.

[0067] (204) Define the shared proximity with respect to energy. The shared proximity with respect to energy of any two nodes i and j in the network is calculated as follows:

[0068]

[0069] Among them, d ij is the distance from node i to node j, knnd ij is the sum of the distances from node j and its K nearest neighbors to node i, i.e. knnd ij =∑ p∈KNN(j)∪j d ip .

[0070] (205) Calculate the total shared proximity s and relative distance λ of each node in the cluster with respect to energy, as follows:

[0071] s i =ΣESnd(i, j) (2-4)

[0072]

[0073] Among them, s i is the sum of the shared proximity between node i and all nodes in the cluster regarding energy, d ij is the distance between node i and node j, ESnd(i, j) is the shared proximity of nodes i and j with respect to energy, and λ i is the relative distance of node i, indicating the distance from the point closest to node i among the points near the point whose local energy density is greater than its own to node i.

[0074] (206) Calculate the decision values ​​μ of all nodes in the cluster, take the point with the highest decision value, and use it as the cluster head node. The formula is:

[0075]

[0076] in, is the average value of the local density of energy of the nodes in the cluster, is the average value of the relative distances of the nodes within the cluster, α and β are weight coefficients, satisfying α+β=1, which can be adjusted according to actual needs.

[0077] The specific process of step three is as follows:

[0078] (301) If there is no candidate routing node near the cluster head node, data is transmitted directly to the base station. If there is a candidate routing node near the cluster head node, the distance from the cluster head node to the base station is compared with the threshold d0 set by the first-order radio model. If the distance from the cluster head node to the base station is less than or equal to the threshold d0, data is transmitted to the base station in a single hop manner; if the distance from the cluster head node to the base station is greater than the threshold d0, the cluster head will establish a selection function based on the total residual energy SRE, angle θ and distance d of the routing nodes near it, and then calculate the selection function value of each nearby routing node, and select the node with the largest selection function value as the next hop routing node. The established selection function is as follows:

[0079]

[0080] Among them, d s is the distance between the candidate routing node and the base station, and cosθ is the cosine value of the angle between the line connecting the current node and the base station and the line connecting the current node and the candidate node.

[0081] In summary, when selecting cluster heads, the present invention adds energy consideration to the simple nearest neighbor method, optimizes the cluster head selection method, prolongs the network life cycle, and improves communication efficiency.

Claims

1. A low-power routing control method based on energy proximity and DBSCAN algorithm, characterized in that The routing control is divided into three steps, namely clustering, cluster head selection and next hop routing node selection. The steps are as follows: (1) Use the DBSCAN algorithm to cluster the nodes. The specific steps are as follows: (101) The base station sets parameters and performs initialization. Let eps be the scanning radius and minPts be the minimum number of included points. Initially, all nodes are assumed to be unvisited. (102) Randomly select an unvisited point, starting from this point, find all nearby points whose distance to this point is within (including) eps, and mark this point as visited. (103) Compare the number of nearby points of the point with minPts. If the number of nearby points is ≥ minPts, mark the point as a core point, initialize a new cluster A, and add the point and the points within eps (including eps) from the point to cluster A; if the number of nearby points is ≤ minPts, but it is in the neighborhood of the core point, mark it as a boundary point and assign it to the cluster where the core point is located; if the number of nearby points is ≤ minPts and it does not belong to the neighborhood of any core point, mark the point as a noise point. (104) Expand the cluster. For each point Q in cluster A, obtain the number of points in its eps neighborhood. If the number of points in the eps neighborhood of point P ≥ minPts, add the points in the eps neighborhood of P to cluster A. (105) Repeat step (104) until cluster A cannot be expanded. (106) Repeat steps (102)-(105) until all points have been visited. (107) Noise points are clustered separately. (108) Return the cluster list. (2) Using the shared proximity of energy, redefine the sum of shared proximity of energy s to optimize the selection of cluster heads. The specific steps are as follows: (201) For each node in the cluster, find the K nearest neighbors (KNN), where the K nearest neighbors are the K nodes closest to the current node. (202) Calculate the energy-related shared nearest neighbor SNN of each node and other nodes, and the formula is: SNN(i, j)=KNN(i)∩KNN(j) (2-1) where KNN(i) is the K nearest neighbors of node i, and SNN(i, j) is the shared nearest neighbors of node i and node j. (203) Calculate the total residual energy SRE of the shared nearest neighbors. The calculation method of the total residual energy SRE(i, j) of the shared nearest neighbors of nodes i and j is: SRE(i,j)=ΣRE(m),m∈SNN(i,j) (2-2) Among them, RE(m) is the residual energy of node m. (204) Define the shared proximity with respect to energy. The shared proximity with respect to energy of any two nodes i and j in the network is calculated as follows: Among them, d ij is the distance from node i to node j, knnd ij is the sum of the distances from node j and its K nearest neighbors to node i, i.e. knnd ij =∑ p∈KNN(j) ∪jd ip (205) Calculate the total shared proximity s and relative distance λ of each node in the cluster with respect to energy, the formula is: s i =∑ESnd(i,j) (2-4) Where si is the sum of the shared proximity between node i and all nodes in the cluster regarding energy, d ij is the distance between node i and node j, ESnd(i, j) is the shared proximity of nodes i and j with respect to energy, and λ i is the relative distance of node i, indicating the distance from the point closest to node i among the points near the point whose local energy density is greater than its own to node i. (206) Calculate the decision values ​​μ of all nodes in the cluster, take the point with the highest decision value, and use it as the cluster head node. The formula is: in, is the average value of the local density of energy of the nodes in the cluster, is the average value of the relative distance of nodes within the cluster, α and β are weight coefficients ρ, satisfying α+β=1, which can be adjusted according to actual needs. (3) Select the route. The steps are as follows: (301) If there is no candidate routing node near the cluster head node, data is transmitted directly to the base station. If there is a candidate routing node near the cluster head node, the distance from the cluster head node to the base station is compared with the threshold d0 set by the first-order radio model. If the distance from the cluster head node to the base station is less than or equal to the threshold d0, data is transmitted to the base station in a single hop manner; if the distance from the cluster head node to the base station is greater than the threshold d0, the cluster head will establish a selection function based on the total residual energy SRE, angle θ and distance d of the routing nodes near it, and then calculate the selection function value of each nearby routing node, and select the node with the largest selection function value as the next hop routing node. The established selection function is as follows: Among them, d s is the distance between the candidate routing node and the base station, and cosθ is the cosine value of the angle between the line connecting the current node and the base station and the line connecting the current node and the candidate node.

Images