Energy-efficient WSN clustering multi-hop transmission method

Abstract

This invention discloses an energy-efficient WSN clustered multi-hop routing method, which includes the following steps: randomly deploying sensor nodes in the area to be tested; initializing node information in the network and determining the minimum node routing cost from the node to the base station based on the node's positional relationship in the network; pre-screening nodes that meet the criteria to become cluster heads through a clustered multi-hop transmission mode, and re-encoding the nodes according to their positional relationships; selecting cluster heads for the re-encoded nodes; performing a global optimal search on the selected cluster heads to determine non-cluster head nodes; and transmitting data with the base station through the non-cluster head node with the minimum routing cost. This invention reduces the distance of each data transmission by using a multi-hop transmission mode, thereby reducing energy consumption during transmission; by introducing node routing cost into the objective function, it balances the load of the communication nodes (CHs) in the network, ensuring that the selected CH is optimal during the data transmission phase, further reducing energy consumption and improving network energy efficiency.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of WSN, and particularly relates to a WSN clustering multi-hop transmission method based on energy efficiency. BACKGROUND

[0002] Wireless sensor network is an important part of Internet of Things, and is a network with less infrastructure and resource constraints. Network needs inter-node cooperative communication to complete network establishment, and the purpose is to collect data and transmit data to the designated receiving point. Wireless sensor network is usually composed of hundreds or thousands of sensor nodes connected, which are randomly distributed in the environment to be detected, used for sensing and natural information collection and processing, and used for evaluating and forwarding and receiving data. With the rapid development of science and technology, the manufacturing cost of sensor nodes is low, which greatly reduces the overall cost of wireless sensors in sensing, processing and transmitting information, so that wireless sensors have higher competitiveness in Internet of Things technology. The architecture of WSN is affected by various parameters such as scalability, fault tolerance, energy efficiency, etc. Energy efficiency is the biggest challenge faced by wireless sensor network, and is the first consideration for network establishment. Each sensor node has limited energy, and current technology cannot widely use energy supplement methods to solve the problem of limited energy. In order to improve the robustness of the network, efficient topology management is the main method to improve energy utilization efficiency and prolong the survival time of the network. Improving the energy efficiency of the network, optimizing the network topology structure, and making the nodes transmit data at the minimum energy consumption are effective ways to prolong the survival cycle of the network.

[0003] The first clustering-based routing protocol is LEACH, which divides the nodes in the network into clusters, randomly selects nodes in the cluster as CH, and sends the collected data to the CH. The cluster structure reduces the amount of data transmission and the delay of data communication, and is convenient for optimizing network management. However, the random selection of CH in the algorithm leads to low network energy efficiency and poor network stability.

[0004] In the cluster structure, the problem of CH selection has always been considered as a non-deterministic problem (NP-Hard) with polynomial complexity. In related research, meta-heuristic optimization algorithm is considered as a high-energy efficient CH selection algorithm. Some nodes are randomly selected as CH by the algorithm, and the objective function of CH is calculated, including the residual energy of CH, the distance from node to CH, the distance from CH to base station, the number of member nodes of CH, and the centrality of CH. The algorithm optimizes the value of the objective function, determines the group of nodes with the minimum objective function value as the CH nodes in the network. Then the optimal path of CH data transmission to BS is found by the meta-heuristic algorithm. However, this method does not consider the random distribution of nodes in the optimization of CH selection, which leads to unsatisfactory optimization effect.

[0005] In a large network, CH far away from the BS usually adopts inter-CH multi-hop mode for data transmission, which is to use other CH as a relay node to finally transmit data to the BS. This method can avoid the rapid energy consumption caused by long-distance data transmission, but the disadvantage is that this transmission method increases the consumption of CH, and the transmission path formed by a small number of CH nodes is not the optimal data transmission path. SUMMARY

[0006] The present application provides a WSN clustering multi-hop transmission method based on energy efficiency to reduce energy consumption from data transmission.

[0007] The present application is achieved in that the WSN clustering multi-hop routing method based on energy efficiency comprises the following steps:

[0008] Step S1: define the size of the space to be measured, randomly deploy sensor nodes in the area to be measured, all sensors have a unique node number, and the node obtains its own position and the distance to the base station;

[0009] Step S2: initialize the node information in the network, and determine the minimum node routing cost of all nodes to the base station according to the position relationship of the nodes in the network;

[0010] Step S3: pre-screen the nodes that meet the CH through the clustering multi-hop transmission mode, and re-encode the nodes according to their position relationship; CH represents cluster head;

[0011] Step S4: design a reasonable objective function; perform global optimal search on the re-encoded nodes; determine the optimal CH combination;

[0012] Step S5: CH searches for the optimal non-cluster head node, finally selects the data transmission path with the minimum routing cost, and determines the optimal multi-hop transmission path of CH to the base station;

[0013] Step S6: the CH in the network receives the data of its member nodes, and after data fusion, transmits data to the base station through multi-hop of non-cluster head nodes.

[0014] In step S6, the CH in the network receives the data of its member nodes, and after data fusion, transmits data to the base station through multi-hop of non-cluster head nodes.

[0015] The optimal distance of the member nodes of the CH node in the network to the CH ensures that the energy consumed by the data communication between the nodes and the CH node is the least. The CH receives the data of the member nodes for data fusion, which reduces the amount of data transmitted to the BS in the network and reduces the energy consumption of the data routing in the network. The multi-hop routing of the non-cluster head node not only balances the energy consumption of the CH, but also avoids the excessive energy consumption of the CH caused by long-distance data transmission.

[0016] Preferably, in step S1, all sensors have a unique node number, which is determined by the distance from the node to the base station (BS) in the network. The nodes are numbered in ascending order based on the distances between the n nodes and the base station, with the node number N being... i Satisfy the following formula:

[0017] D is (N i+1 ,BS)≥D is (N i ,BS),i+1≤n

[0018] BS stands for base station; D stands for D. is (N i ,BS) represents the distance from node Ni to BS.

[0019] Preferably, in step S2, the minimum node routing cost is the minimum energy required for each node to transmit data to the BS; the minimum routing cost of a node is calculated by minimizing the number of transmissions and ensuring the node is in free-space data transmission mode; the calculation formula is as follows:

[0020]

[0021] N m (route) is N m The node routing cost.

[0022] Preferably, step S3, which involves selecting the CH for the re-encoded node, includes the following operations:

[0023] Through two thresholds E m Compare with E0, E m To compute node N m The average remaining energy of all nodes within the range d1, if N m Energy greater than or equal to E m This indicates that the node has an energy advantage, thus becoming a CCH; E m The calculation is as follows:

[0024]

[0025] E(N i ) represents node N i The remaining energy; t is the number of nodes in the d1 range of this node;

[0026] Where t is:

[0027] t=∑D is (N i N m )≤d1

[0028] d1 is a distance parameter, when the nodes in the network are generally energy deficient, E m The value is thus smaller.

[0029] Preferably, the step S3 of re-encoding the nodes according to the positional relationship of the nodes comprises the following operations:

[0030] E0 is introduced for screening, E0 is the energy required for data reception and aggregation; the value of E0 is calculated by the following formula:

[0031]

[0032] W is a parameter; if the energy of the node is higher than two thresholds, it is added to the CCH;

[0033] First, determine the node N i in the CCH closest to the BS, re-number this node as K1, then search for the node in the CCH closest to this node, and encode it as K2; repeat this operation until all nodes are re-encoded; the CCH is:

[0034] CCH = {K1, K2, …, K n}

[0035] n is the number of nodes in the CCH.

[0036] Preferably, the step S4 of selecting the CH for the re-encoded nodes uses an objective function;

[0037] The objective function includes considering the number of CH nodes; the setting of the CH nodes will change with the number of surviving nodes in the network, and will also be related to the location of the base station and the size of the area to be tested; at a position farther from the base station, the range of the cluster needs to be increased to reduce the number of CHs; the number of CHs is:

[0038] f1 = size (unique (C i ))

[0039] f1 is the number of cluster heads; size (unique (C i )) represents the number of nodes in the individual C i that do not repeat.

[0040] Preferably, the objective function includes considering the residual energy of the nodes; the CH needs to collect the data of the member nodes, fuse the data, and send the fused data packet to the next routing node, which consumes a large amount of energy of the CH; in the selection of the CH, a node with higher energy is selected as the CH node; the average residual energy of the CH is represented as follows:

[0041]

[0042] E N(i) represents the residual energy of the ith CH; f1 represents the number of CHs.

[0043] Preferably, the objective function includes a CH routing cost factor, the energy required by the CH node to transmit data to the base station is the main factor affecting the network energy consumption; the routing cost of the CH can reflect the energy efficiency of the node in the data transmission stage, and selecting a CH node with a small routing cost can reduce the energy consumption of the network in the data transmission process; the routing cost is an important indicator for measuring the energy efficiency of the node in transmitting data to the base station, and ensures that the finally selected CH node makes the network energy efficiency optimal. The CH routing cost f3 is represented as follows:

[0044]

[0045] CH i (route) represents the routing cost of the ith node in the CH;

[0046] The objective function includes a CH centrality factor, which describes the total distance from the member node to the CH; the greater the distance from the member node to the CH, the more energy the member node needs to transmit data to the CH, and reducing the distance from the member node to the CH can reduce the energy consumption of the network. Considering the CH centrality in the selection of the CH can ensure that the energy consumption of the network is minimized, and the expression of the CH centrality is as follows:

[0047]

[0048] T CH(j) represents the number of member nodes of the jth CH.

[0049] Preferably, in the step S5, a global optimal search is performed on the selected CH to determine the non-cluster head node, and the meta-heuristic algorithm of LF and the node competition mechanism are adopted;

[0050] After random walk is performed by using the meta-heuristic algorithm of LF, the clustering threshold of the node is calculated by the distance of the node to the base station, and the nodes within the clustering range compete for the CH; in the optimization process, local search and global search are adopted; the new position New(C i ) after search is as follows:

[0051]

[0052] C r represents a random individual in the population C; R() is a random number between 0 and 1; r, a1, and a2 are real parameters; C p is the individual currently obtaining the minimum objective function value;

[0053] If the distance between two nodes in Ci is less than a threshold, then the energy of the two nodes is compared, and the node with lower energy is replaced by the node with higher energy. The distance threshold is:

[0054]

[0055] D ch This is the distance threshold.

[0056] Preferably, in step S6, data transmission is performed between the non-cluster head node and the base station through the non-cluster head node with the lowest routing cost. The data transmission adopts a multi-hop routing method, that is, the non-cluster head is used as the data transmission node of CH.

[0057] During the node selection process, each node selects the shortest transmission path based on the minimum transmission cost. Therefore, after the CH selects the optimal next-hop node, the node receives the routing data from the CH and transmits the data to the base station according to the route with the minimum routing cost.

[0058] The routing cost from CH to all nodes within the range d2 is calculated using the following formula:

[0059] R(N m N k )=α+β1×D is (N m N k ) 2 +N k (route)

[0060] R(N m N k ) represents N m To N k The transmission cost, N m Indicates the current CH, N k Represents node N m The non-cluster head nodes within the d2 range;

[0061] E(N i )>N i (route),R(N m N i )=min(R)

[0062] Ni represents N m The node with the lowest transmission cost within the d2 range; E(N) i ) represents node N i The remaining energy; if N i If it does not exist, the next hop with the lowest routing cost will be used for data transmission.

[0063] The beneficial effects of the present invention are as follows: 1. By using a multi-hop transmission mode, nodes that are farther away from the base station first transmit data to nodes that are closer to the base station, and then the nodes that are closer to the base station transmit the data to the base station, thereby reducing the distance of each data transmission and reducing the energy consumption of the data transmission process.

[0064] 2. By introducing a CH competition mechanism during the optimization process, the optimal number of CHs is determined to minimize network energy consumption. By introducing node routing costs into the objective function, the load of CHs in the network is balanced, ensuring that the selected CHs are optimal during the data transmission phase, further reducing energy consumption and improving network energy efficiency. Attached Figure Description

[0065] Figure 1 This is the data transmission model of the present invention.

[0066] Figure 2 This is a comparison chart of the number of surviving nodes in four scenarios. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of the present invention clearer, further description is provided below in conjunction with the accompanying drawings and specific embodiments. The specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. The invention is further summarized in conjunction with the accompanying drawings.

[0068] like Figure 1 As shown, the energy-efficient WSN clustered multi-hop routing method includes the following steps:

[0069] Step S1: Define the size of the space to be measured, randomly deploy sensor nodes in the area to be measured, and each sensor has a unique node number. The node obtains its own position and the distance to the base station.

[0070] In a WSN, nodes are randomly distributed within the test area. To facilitate the calculation of node routing costs, the nodes in the network are renumbered based on their distance to the base station (BS). The nodes are numbered in ascending order based on the distances of the n nodes to the base station, and the node numbers are N. i The expression that satisfies this condition is:

[0071] D is (N i+1 ,BS)≥D is (N i ,BS),i+1≤n

[0072] BS stands for base station; D stands for D. is (N i ,BS) represents the distance from node Ni to BS.

[0073] Step S2: Initialize the node information in the network, and determine the minimum node routing cost from all nodes to the base station based on the node's positional relationship in the network.

[0074] In a network, the energy consumption factors affecting data transmission from a node to a server (BS) mainly include two aspects: the number of hops during transmission and the distance of each hop. Each additional hop increases the energy consumption for sending and receiving data. The energy required for data transmission increases with distance; when the transmission distance is less than d0, the energy required increases quadratically with the transmission distance; when the transmission distance is greater than d0, the energy required increases quadratically with the transmission distance. The optimal data transmission pattern is the minimum number of hops and a transmission distance less than d0.

[0075] When the distance between a node and a BS is less than d0, direct communication between the node and the BS consumes less energy, and the node routing cost is expressed as:

[0076]

[0077] In the formula, α and β1 are parameters; when the distance from the node to BS is greater than d0, the node uses a multi-hop method for data transmission to avoid consuming more energy for long-distance data transmission. At this time, node N m The routing cost is calculated as follows:

[0078]

[0079] In the formula, m > j, β2 is a parameter; N j Is it to make N m The node with the smallest (route) value.

[0080] Step S3: Pre-select nodes that meet the criteria to become CH through clustered multi-hop transmission mode, and re-encode the nodes according to their positional relationships;

[0081] When selecting CCHs, the main consideration is the remaining energy of the nodes, i.e., through two thresholds E. m Compare it with E0. m By computing node N m The average remaining energy of all nodes within the range d1, if N m Energy greater than or equal to E m This indicates that the node has an energy advantage and can become a CCH. m The calculation is as follows:

[0082]

[0083] E(N i ) represents node N iThe remaining energy; t is the number of nodes in the d1 range of this node; t is:

[0084] t=∑D is (N i N m )≤d1

[0085] d1 is the distance parameter; when the nodes in the network generally lack energy, E m The value is therefore relatively small. To prevent nodes with insufficient remaining energy from becoming CCHs (Consumer Chips), E0 is introduced for filtering. E0 describes the energy required for data reception and aggregation. The value of E0 is calculated using the following formula:

[0086]

[0087] W is a parameter; if a node's energy is higher than two thresholds, it is added to the CCH.

[0088] To improve the search speed of LFMA, the CCH is re-encoded based on its positional relationships. First, the node N in the CCH that is closest to the BS is determined. i The node is renumbered as K1, and then the nearest node in the CCH is searched and encoded as K2; this operation is repeated until all nodes are re-encoded; the CCH is:

[0089] CCH={K1,K2,…,K n}

[0090] n is the number of nodes in CCH.

[0091] Step S4: Design a reasonable objective function; perform a global optimal search on the re-encoded nodes; determine the optimal CH combination.

[0092] All nodes in the network are homogeneous, therefore each node can become a CH (Chain Leader). To ensure optimal network energy efficiency, the following factors need to be considered:

[0093] The number of cluster heads (CHs) is a key factor affecting network lifetime. In clustering protocols, a high CH ratio leads to low data fusion among member nodes, increasing the amount of data that the network needs to transmit and thus increasing network energy consumption. A low CH ratio increases the distance from member nodes to the CHs, resulting in higher energy consumption during transmission from member nodes to the CHs. Furthermore, the increased number of member nodes in a CH requires the CH to consume more energy to receive and fuse data. The optimal number of CHs varies with the number of surviving nodes and is also related to the location of the base station and the size of the area to be tested. At locations farther from the base station (BS), the clustering range needs to be increased to reduce the number of cluster heads, improve data fusion, and reduce network routing burden. The number of cluster heads, f1, is represented as follows:

[0094] f1 = size(unique(C) i ))

[0095] size(unique(C i )) represents individual C i The number of unique nodes in the array.

[0096] The remaining energy of a node is crucial for data collection, fusion, and transmission with routing nodes. If the remaining energy of a node (CH) is lower than required, data loss will occur. Selecting a node with higher remaining energy ensures network performance and balances energy consumption across nodes. A CH needs to collect data from member nodes, fuse the data, and send the fused packets to the next routing node, a process that consumes a significant amount of energy. When selecting a CH, its remaining energy should be sufficient to complete the operation. Choosing a node with higher energy can prevent data loss due to energy depletion, improving network robustness. The average remaining energy of a CH is expressed as follows:

[0097]

[0098] E N(i) f1 represents the remaining energy of the i-th CH. f1 represents the number of CHs.

[0099] CH routing cost: This defines the energy required for a CH to transmit data to the base station. The routing cost of a CH reflects the energy efficiency of a node during data transmission; selecting a CH with a lower routing cost can reduce network energy consumption during data transmission. Routing cost is an important indicator of the energy efficiency of a node transmitting data to the base station. The CH routing cost f3 is expressed as follows:

[0100]

[0101] CH i (route) represents the routing cost of i nodes in CH.

[0102] CH centrality: describes the total distance from member nodes to the CH (Center for Connecting Nodes). A larger distance from a member node to the CH means more energy is required for the member node to transmit data to the CH. Therefore, reducing the distance from member nodes to the CH can reduce the network's energy consumption. The distance from a node to the CH in the network is represented as follows:

[0103]

[0104] T CH(j) This represents the number of member nodes of the j-th CH.

[0105] Based on the four factors influencing CH selection, a final objective function F is formed to evaluate the fitness of the selected CH. The result of F relates to the direction of the CH search and the final CH selection. The objective function F is defined as follows:

[0106] F=κ×(N alive -f1)+f2+f3+β1×f4

[0107] κ and β1 are parameters, N alive This indicates the number of surviving nodes in the network.

[0108] To find the optimal solution, the metaheuristic algorithm of LF and a node competition mechanism were adopted; in the algorithm's iterative update, the step size adopted LF's random walking strategy. LF is inspired by the movement of organisms in nature, and it performs better in searching large unknown spaces; the walking step size S is defined as:

[0109]

[0110] λ∈(0,2) is the distribution index of Levy. U and V are defined as follows:

[0111]

[0112]

[0113] Standard deviation σ u and σ v The calculation formula is:

[0114]

[0115] The gamma function for integers is expressed as:

[0116]

[0117] To avoid getting trapped in local optima during the optimization process, a combination of local and global search is used in the iterative update; the new position New(C) after the search... i )for:

[0118]

[0119] C r Let R represent a random individual in population C; R() is a random number between 0 and 1; r, a1, and a2 are actual parameters; C p It is the individual that currently obtains the minimum objective function value;

[0120] A CH competition mechanism is introduced during the optimization process to determine the optimal number of CHs that minimizes network energy consumption. In C... iIf the distance between two nodes is less than a threshold, then the energies of the two nodes are compared, and the node with lower energy is replaced by the node with higher energy. Threshold D ch The distance threshold D increases with the distance from the node to the BS. Nodes closer to the BS need to receive routing data from more distant CHs. In this case, an excessively large clustering range will cause excessive energy consumption for the nodes. Conversely, expanding the clustering range of nodes farther from the BS can reduce the number of data packets in the network, thus balancing the energy consumption of the nodes. ch for:

[0121]

[0122] D ch Distance threshold;

[0123] The node competition mechanism ensures that the CHs are evenly distributed throughout the network.

[0124] Step S5: CH searches for the optimal non-cluster head node, and finally selects the data transmission path with the minimum routing cost to determine the optimal multi-hop transmission path from CH to the base station;

[0125] Non-cluster head nodes, representing the nodes with the lowest routing cost, transmit data to the base station using a multi-hop routing method. Considering that the CH (Cluster Header) needs to receive data from member nodes and aggregate the received data, a process that consumes a significant amount of energy, non-cluster head nodes are used as data transmission nodes for the CH to balance energy consumption and prevent excessive energy depletion. This routing method reduces the CH's load and energy consumption, and by using non-cluster head nodes as data routing nodes, a shorter transmission path can be found between the CH and the BS (Base Station), further reducing network energy consumption.

[0126] During the point selection process, each node selects the shortest transmission path based on the minimum transmission cost. Therefore, after the CH selects the optimal next-hop node, the node receives the routing data from the CH and transmits the data to the base station according to the route with the minimum routing cost.

[0127] The routing cost from CH to all nodes within the range d2 is calculated using the following formula:

[0128] R(N m N k )=α+β1×D is (N m N k ) 2 +N k (route)

[0129] R(N m N k ) represents N m To N kThe transmission cost, N m Indicates the current CH, N k Represents node N m The non-cluster head nodes within the d2 range;

[0130] E(N i )>N i (route),R(N m N i )=min(R)

[0131] Ni represents N m The node with the lowest transmission cost within the d2 range; E(N) i ) represents node N i The remaining energy; if N i If it does not exist, the next hop with the lowest routing cost will be used for data transmission.

[0132] Step S6: The CH in the network receives data from its member nodes, and after data fusion, transmits the data to the base station through multiple hops of non-cluster head nodes.

[0133] After receiving data from member nodes, the CH (Chief Cluster) aggregates the data. Data aggregation uses a 10% aggregation rate strategy. The size L of the CH aggregation data packet is... agg The calculation formula is as follows:

[0134] L agg =L+(L×R) a ×N member )

[0135] N member R represents the number of members within a cluster. a It is the data aggregation rate. The energy E required to aggregate data (CH) da (L, N) member The calculation formula is as follows:

[0136] E da (L,N member )=EDA×L×N member

[0137] EDA is the energy required to aggregate 1 bit of data.

[0138] The performance of the multi-hop routing protocol algorithm (LFMAR) was simulated using MATLAB 2016a. The experimental scenario made the following assumptions: the base station (BS) possesses unlimited energy, i.e., energy consumption of the BS is not considered; the sensor nodes remain in their positions after deployment; and the node transmission uses a widely accepted first-order radio model as the energy consumption model, with model parameters shown in Table 1.

[0139] Table 1

[0140]

[0141] The formula parameter values ​​in LFMAR are shown in Table 2.

[0142] Table 2

[0143] Parameter Value α 0.4 [Alpha]1 0.0004 <![CDATA[β2]]> 0.000000052 [E0] 0.02 a 900 b 8 W 10 κ 0.42 d1 90 [d2] 60

[0144] To verify the performance of the routing protocol presented in this paper in different monitoring areas, four different scenarios were designed. The parameters for each scenario are shown in Table 3.

[0145] Table 3

[0146] Scenario Number of nodes Network area BS position 1 200 100m*100m (0，100) 2 200 200m*200m (0，200) 3 200 300m*300m (0，300) 4 200 400m*400m (0，400)

[0147] In scenarios 1 and 2, the main factors affecting network energy consumption are intra-cluster data transmission and aggregated data; therefore, the comparison primarily focuses on the optimization of CH selection in node clustering. In scenarios 3 and 4, the test areas are larger, resulting in greater distances between nodes. To improve network efficiency, and considering that the energy required for CH data routing and transmission is a major factor influencing network energy consumption, the primary factor affecting network energy efficiency in scenarios 3 and 4 is the energy consumption generated by CH data routing within the network. The base station is located in one corner of the area.

[0148] like Figure 2 As shown, the LFMAR algorithm used in this invention has a significantly higher number of active nodes during data transmission compared to the three comparison algorithms. In the FND comparison, LFMAR improves efficiency by 60.42%, 68.51%, and 8.83% compared to LEACH-A*, BOA-ACO, and EFC-ISFLA, respectively, indicating that LFMAR outperforms the three comparison algorithms in terms of average energy consumption. It improves efficiency by 37.4%, 22.51%, and 25.8% in HND, and by 59.65%, 55%, and 38.44% in LND, respectively.

[0149] In scenario 2, LFMAR showed improvements of 93.69%, 66.84%, and 12.79% on FND compared to LEACH-A*, BOA-ACO, and EFC-ISFLA, respectively; improvements of 65.29%, 9.54%, and 26.03% on HND; and improvements of 73.97% and 31.12% on LND, respectively. However, EFC-ISFLA showed an improvement of 19.25% on LND compared to LFMAR.

[0150] In scenario 3, LFMAR improved performance by 96.95%, 85.87%, and 2.77% on FND, 78.14%, 54.3%, and 13.68% on HND, and 90.63%, 19.38%, and 7.86% on LND.

[0151] In scenario 4, LFMAR achieved a FND of 270 rounds, representing improvements of 97.77%, 95.55%, and 10% respectively; and improvements of 89.21%, 71.56%, and 9.8% respectively in HND. LFMAR achieved a LND of 341 rounds, outperforming LEACH-A* and EFC-ISFLA by 75.07% and 13.19% respectively; however, BOA-ACO slightly outperformed LFMAR in LND.

[0152] In Scenario 1, a significant number of nodes are within the base station's d0 range, where d0 represents the distance margin. Nodes communicate directly with the BS, reducing the energy consumption required for CH reception and data fusion during clustering. After 1000 rounds, the rate of node death in the network slows down. In Scenario 2, before round 724, the number of surviving nodes per round in LFMAR is greater than or equal to the other three methods. After round 724, EFC-ISFLA has the highest number of active nodes. Between rounds 550 and 724, the number of active nodes in EFC-ISFLA is significantly less than that in the LFMAR method presented in this paper.

[0153] In scenarios 3 and 4, most CH (Choice Channel) data needs to be transmitted to the base station (BS) via multiple hops, making data transmission the main source of network energy consumption. Therefore, the number of multi-hop data transmission links becomes a major factor affecting network energy efficiency, and changes in the clustering range of CHs also affect the amount of CH data. The LFMAR algorithm presented in this paper demonstrates its advantages in multi-hop routing data. Nodes closer to the BS communicate directly with the BS, reducing the energy consumption required for CH reception and data fusion during clustering, reserving more energy for receiving routing data, and avoiding excessive energy consumption from long-distance transmission of CHs farther from the BS. Therefore, in scenarios 3 and 4, LFMAR achieves better results compared to other algorithms in terms of network lifetime.

[0154] Packet delivery rate is a comparison of the amount of data collected by the network and transmitted to the BS. In the four scenarios, the CH in LEACH-A* is randomly selected, resulting in lower network energy efficiency. As the area under test increases, the amount of data reaching the BS varies significantly.

[0155] BOA-ACO is a representative metaheuristic algorithm that uses the butterfly algorithm to elect the Chosen Leaders (CHs) among network nodes. During iterative updates, the random distribution of nodes leads to low inter-node correlation, resulting in slow CH selection optimization. The fixed number of CHs selected (10% of the total number of nodes) cannot guarantee optimal results during iterative updates. Furthermore, the algorithm's fitness function does not consider balancing the energy consumption of network nodes, causing nodes far from the Base Station (BS) to prematurely fail due to excessive energy consumption. In the data routing node selection process, routing paths are chosen between CHs, leading to excessively rapid energy consumption at CH nodes and the selection of suboptimal transmission routes. Therefore, BOA-ACO transmits fewer data packets to the BS than EFC-ISFLA and the algorithm proposed in this paper.

[0156] EFC-ISFLA employs an optimized fuzzy inference system and considers adjusting the clustering range of CHs based on network structure to balance the energy consumption of nodes in the network. However, the EFC-ISFLA protocol uses an improved frog-jumping algorithm to find data routing links between CHs and adopts a data transmission method between CHs, which leads to excessively rapid energy consumption at CHs and a transmission distance that is greater than the data transmission distance of non-cluster head nodes proposed in this paper. Therefore, the number of data packets transmitted to the BS is less than that of the algorithm proposed in this paper.

[0157] The LFMAR protocol proposed in this invention optimizes the metaheuristic algorithm's optimization strategy through node information preprocessing and a node contention mechanism, and introduces node routing costs into the objective function, thus ensuring that the selected CH has the highest energy efficiency. By selecting ordinary nodes as multi-hop nodes during the CH data routing phase, energy consumption of the CH is reduced, the load on the CHs in the network is balanced, and the selected CH also has optimal energy efficiency during the data routing phase.

[0158] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for energy-efficient WSN clustering multi-hop routing, characterized in that: The method comprises the following steps: Step S1: defining the size of the space to be measured, randomly deploying sensor nodes in the to-be-measured region, all sensors having unique node numbers, and the nodes obtaining their own positions and distances from the base station; Step S2: initializing the node information in the network, and determining the minimum node routing cost of all nodes to the base station according to the positional relationship of the nodes in the network; Step S3: pre-screening nodes meeting the condition of becoming CHs through a clustered multi-hop transmission mode, and re-encoding the nodes according to the positional relationship of the nodes; Step S4: designing a reasonable objective function, performing global optimal search on the re-encoded nodes, and determining an optimal CH combination; The objective function includes considering the number of CH nodes, the setting of the CH nodes changing with the number of surviving nodes in the network, and being related to the position of the base station and the size of the to-be-measured region; in a position farther from the base station, the range of the cluster needs to be increased to reduce the number of CHs, and the number of CHs is: f1=size(unique(C i )) Wherein, f1 is the number of cluster heads; size(unique(C i )) represents the number of non-repeated nodes in individual C i ​ The objective function includes considering the residual energy of the nodes; the CH needs to collect the data of the member nodes, fuse the data, and send the fused data packet to the next routing node, which consumes a large amount of energy of the CH; in the selection of the CH, a node with higher energy is selected as the CH node, and the average residual energy of the CH is represented as follows: where E N(i) represents the residual energy of the ith CH; f1is the number of cluster heads; Step S5: the CH searches for an optimal non-cluster head node, selects a data transmission path with the minimum routing cost, and determines an optimal multi-hop transmission path of the CH to the BS; Step S6: the CH in the network receives data of its member nodes, performs data fusion, and performs data transmission with the base station through the non-cluster head node multi-hop.

2. The energy efficiency based WSN clustering multi-hop routing method according to claim 1, characterized in that: All sensors in the step S1 have a unique node number, which is numbered by the distance from the node to the BS in the network, and the nodes are numbered in ascending order according to the distance between the n nodes and the base station, and the node number N i satisfies the following formula: D is (N i+1 ,BS)≥D is (N i ,BS),i+1≤n; where BS is a base station; D is (N i ,BS) denotes the distance of node N i to the BS. 3.The energy efficiency based WSN clustering multi-hop routing method according to claim 1, characterized in that: In step S2, the minimum node routing cost is the minimum energy required for each node to transmit data to the BS; the minimum routing cost of the node is calculated through the minimum transmission times and the free space data transmission mode of the node; the calculation formula is as follows: wherein N m (route) is the node routing cost of N m (route) is the node routing cost of N j (route) is the node routing cost of N m( route ) value minimum node; d0is the distance pre-branch. 4.The energy efficiency based WSN clustering multi-hop routing method according to claim 1, characterized in that: In step S3, the re-encoded nodes are selected as CHs, including the following operations: The judgment is made by two thresholds E m and E0, E m is the average residual energy of all nodes in the d1 range of the computing node N m , if the energy of N m is greater than or equal to E m , it indicates that the node has energy advantage, and then joins the candidate cluster head; the calculation of E m is as follows: where E(N i ) denotes the residual energy of node N i ; t is the number of d1 range nodes of the node; t is: t = ∑D is (N i ,N m )≤d1; where d1 is a distance parameter, E m values are therefore smaller.

5. The energy-efficient based WSN clustering multi-hop routing method according to claim 4, characterized in that: In step S3, the re-encoded nodes are re-encoded according to the positional relationship of the nodes, including the following operations: E0 is introduced for screening, E0 is the energy required for data reception and aggregation; the value of E0 is calculated by the following formula: Wherein, W is a parameter; if the energy of the node is higher than two threshold values, the node is added to CCH; a, β1 are real parameters; d0 represents the distance preset; First, determine the node N closest to the BS in the CCH i , renumber this node as K1, then search for the node closest to this node in the CCH, encode as K2; repeat this operation until all nodes are re-encoded; the CCH is: CCH = {K1, K2,..., K n}; n is the number of nodes in CCH. 6.The energy efficiency based WSN clustering multi-hop routing method according to claim 1, characterized in that: The objective function includes the CH routing cost factor, the energy required for the CH node to transmit data to the base station is the main factor affecting the energy consumption of the network; the routing cost of the CH can reflect the energy efficiency of the node in the data transmission stage, and selecting a CH node with a small routing cost can reduce the energy consumption of the network in the data transmission process; the routing cost is an important indicator for measuring the energy efficiency of the node in transmitting data to the base station, and ensures that the finally selected CH node makes the network energy efficiency optimal; the routing cost f3 of the CH is represented as follows: CH i (route) denotes the routing cost of i nodes in CH; f1 is the number of cluster heads; The target function includes a factor of considering CH centrality, which describes the total distance from the member node to the CH; the greater the distance from the member node to the CH, the more energy the member node needs to transmit data to the CH, reducing the distance from the member node to the CH can reduce the energy consumption of the network, and considering the CH centrality in the selection of the CH can ensure that the energy consumption of the network is minimized, and the expression of the CH centrality is as follows: T CH(j) denotes the number of member nodes of the jth CH. 7.The energy efficiency based WSN clustering multi-hop routing method according to claim 1, characterized in that: The step S4 performs global optimal search, and adopts the meta-heuristic algorithm of LF and the node competition mechanism; After random walk by meta-heuristic algorithm of LF, the clustering threshold of nodes is calculated by the distance from node to base station, and the nodes in clustering range compete for CH; in the optimization process, two parts of local search and global search are adopted; the new position New(C i ) is: C r represents a random individual in population C; R() is a random number between 0 and 1; r, a1, a2 are real parameters; C p is the individual that currently obtains the minimum objective function value; S is a walking step length; In C i two nodes distance below the threshold, then the energy of the two nodes are compared, the energy of the low node is replaced by the high energy node; distance threshold value: D ch is a distance threshold value. 8.The energy efficiency based WSN clustering multi-hop routing method according to claim 1, wherein: In the step S5, the CH searches for an optimal non-cluster head node, selects a data transmission path with the minimum routing cost, and determines an optimal multi-hop transmission path of the CH to the base station; In the node selection process, each node selects the shortest transmission path through the minimum transmission cost, so after the CH selects the optimal next hop node, the node receives the routing data of the CH and transmits the data to the base station according to the routing path with the minimum routing cost; The routing cost of the CH to all nodes within the d2 range is calculated by the following formula: R(N m ,N k ) = a + b1 x D is (N m ,N k ) 2 + b2 x N k (route); R(N m ,N k ) denotes the transmission cost from N m to N k , N m denotes the current CH, N k denotes the non-cluster head nodes within d2 range of node N m ; E(N i ) > N i (route), R(N m ,N i ) = min(R). Ni denotes N m the node with the minimum transmission cost in the range of d2; E(N i ) denotes the residual energy of node N i ; if N i does not exist, the next hop with the lowest routing cost is adopted for data transmission.

Images