CN113490221A - Sustainable wireless sensor network system construction method based on heuristic algorithm - Google Patents

Abstract

The invention discloses a method for constructing a sustainable wireless sensor network system based on a heuristic algorithm, which comprises the following steps: step 1, inputting a group of candidate position nodes S_locA set of fixed terminal nodes S_terA sink node s, a parameter k, a wireless communication radius R_comAnd a wireless charging radius R_cha(ii) a Step 2, constructing a graph G based on the input; and step 3, for each fixed terminal node,finding the shortest path of the sink node on G from the fixed terminal node, and setting P as the path set of the paths; step 4, finding the minimum set D (charging node set) of the candidate position nodes, enabling each terminal node u to have at least one relay node v to belong to D, and enabling Dist (u, v) to be less than or equal to R_com(ii) a Step 5, if V_relThe number of vertices in (b) is greater than the repeater number limit (parameter k), then V is reduced using a local search method_rel＝V_locThe number of relay nodes of n (D @ (U @) (P)).

Description

Sustainable wireless sensor network system construction method based on heuristic algorithm

Technical Field

The invention relates to the Internet of things, in particular to a sustainable wireless sensor network system construction method based on a heuristic algorithm.

Background

With the improvement of the intelligence level of the manufacturing industry, sensors or intelligent devices with sensing, monitoring, data transmission and simple calculation functions are beginning to be applied to industrial production, so that the development of industrial internet of things is promoted, a large number of problems and challenges are brought, and the attention of a large number of engineers and researchers is brought. Wireless sensor networks hold an important position in various components of the IoT (internet of things). In recent years, as the technology matures, the 5G network gradually replaces the 4G network due to its advantages of high speed, large capacity, low delay, etc., but it has its own disadvantages. Compared with 4G signals, the transmission distance of 5G signals is shorter, and is generally only 300-400 m. Therefore, the coverage of the 5G base station is smaller than that of the 4G base station. A reasonable approach to solve this problem is to deploy a large number of 5G base stations. However, the cost of building a large number of base stations is high. Therefore, how to construct a wireless sensor network system at the lowest cost in a 5G environment becomes an urgent problem to be solved.

Therefore, the invention provides a sustainable wireless sensor network system construction method based on a heuristic algorithm.

Disclosure of Invention

In order to realize the purpose of the invention, the following technical scheme is adopted for realizing the purpose:

a sustainable wireless sensor network system construction method based on heuristic algorithm is disclosed, wherein: the sustainable wireless sensor network system based on the heuristic algorithm comprises a plurality of fixed terminals, a sink node and a plurality of repeaters, wherein the repeaters can transmit data and wirelessly charge the terminals; the method comprises the following steps:

step 1, inputting a group of candidate position nodes S_locA set of fixed terminal nodes S_terA sink node s, a parameter k, a wireless communication radius R_comAnd a wireless charging radius R_cha；

Step 2, constructing a graph G based on the input;

step 3, for each fixed terminal node, searching the shortest path of the sink node on the G from the fixed terminal node, and setting P as a path set of the paths;

step 4, finding the minimum set D (charging node set) of the candidate position nodes, enabling each terminal node u to have at least one relay node v to belong to D, and enabling Dist (u, v) to be less than or equal to R_com；

Step 5, if V_relThe number of vertices in (b) is greater than the repeater number limit (parameter k), then V is reduced using a local search method_rel＝V_locThe number of relay nodes of n (D @ (U @) (P)).

The method, wherein the step 2 comprises:

constructing an undirected graph G (V) with edge weighting and coloring_loc∪V_ter∪{V_sink})，E＝(E_grey∪E_black) ) as follows:

(a) for S_locCreating a candidate position vertex v for each candidate position node a in_aAnd is added to V_locPerforming the following steps;

(b) for S_terCreating a terminal vertex v for each terminal node a in (1)_bAnd adding V_ter；

(c) For a sink node s, a sink vertex v is created_sink；

(d) For each terminal node a and each candidate position node b, if Dist (a, b) ≦ R_chaThen a black border is created (v)_a,v_b) Weight c-Dist (v)_a,,v_b)^qIs added to E_grey；

(e) For each candidate location node a and each candidate location (or terminal) node b, if Dist (v)_a,v_b)≤R_comAnd there is no black border between them, a gray border (v) is created_a,v_b) Weight c-Dist (v)_a,v_b)^qIs added to E_black。

The method, wherein the step 3 comprises: for each vertex V ∈ V_terThe algorithm uses the Dijsktra algorithm to find the shortest path from V to the convergence vertex.

The method, wherein the step 4 comprises: dividing the set of vertices D into two sets D₁And D₂Wherein

Comprises V_terAll the vertexes which are adjacent to a certain vertex through the black edge; let G_black＝((V₁,V₂)，E_black) Is a set of edges E_blackDerived graph, where V₁＝V_ter,V₂＝V(E_black)\V₁，G_blackIs a two-part graph which is composed of a plurality of parts,

let V₂ ^*＝V₂\D₁,V₁ ^*＝V_ter\N_Gblack(D₁) In which N is_Gblack(D₁) Is D₁Is in G_blackA set of vertices in (1); let G^*＝(V^*＝(V₁ ^*,V₂ ^*)，E^*) Is formed by a set of vertices V^*Derived G_blackA subgraph of (1); finding a minimum set using a greedy algorithm

At G^*Middle domination of V₁ ^*All nodes of (2), set of vertices D₂Consisting of all these selected vertices.

The method, wherein the step 5 comprises: if | V_rel＝V_loc∩(D∪V(P))|>k, set forth diagram G_PFor the union of all paths in P, if the vertex V ∈ V_relD is G_PCoating the second-degree vertex of the middle part as black; let v be P (a, v) in path P_sink) A black vertex above, V is a at P (a, V)_sink) Two neighbors of (1), where v⁺，v^-Are each at G_PPaths P (a, v) and P (v, v)_sink) If (v) is⁺,v^-) E.g. E. By inclusion of (v) in P⁺,v,v^-) Is used in all paths of (v)⁺,v^-) Substitution subpath (v)⁺,v,v^-) (ii) a This operation is repeated until | V_relK is less than or equal to | k; if there is a vertex u e (V (P) \ V)_ter) So that (u, v)⁺) E.e, then P (v) is contained in P⁺，v_sink) All paths of (v)⁺U) and P (u, v)_sink) Instead of the subpath P (v)⁺，v⁺) E, wherein P (v)⁺，v_sink)，P(u，v_sink) Are each G_PFrom V⁺To v_sinkAnd from u to v_sinkA path of (a); an evaluation index is defined as

Wherein Δ_cost(v, u) and Δ_relay(v, u) is the inclusion of (P (v) in P⁺，v_sink) Will sub-path P (v) of all paths⁺，v_sink) Replacement by (v)⁺，u)∪P(u，v_sink) Then, choose the minimum value of sigma (v) among all the vertices in B₀) Vertex v of₀And performing a path replacement operation as a local search rule for each step in the algorithm G-SC until a set of relay vertices of at most k is found.

Drawings

FIG. 1 is a schematic diagram of a sustainable wireless sensor network system based on heuristic algorithm;

FIG. 2 is a schematic diagram of an example communication and charging network;

fig. 3 is a simulation experiment result of the CPG algorithm.

Detailed Description

The following detailed description of the embodiments of the present invention is provided in conjunction with the accompanying drawings of fig. 1-3.

As shown in fig. 1, the sustainable wireless sensor network system based on the heuristic algorithm includes a plurality of fixed terminals, a sink node and a plurality of repeaters. The repeater can transmit data and wirelessly charge the terminal. The following two conditions are satisfied, namely 1) communication conditions that the data of each terminal can be transmitted to a sink node through some relays; 2) conditions for sustainable development: each terminal may be charged by at least one repeater. The first condition ensures the basic function of the system, i.e. the communication function. The second condition ensures a sustainable operation of the whole system.

The construction of a sustainable wireless sensor network system needs to complete two tasks, namely (1) the task of deploying the repeater at a limited position; and (2) a routing task (for each terminal, selecting a transmission route from the terminal to the sink node). The optimization goal of the system is to minimize communication costs while meeting communication conditions and sustainability conditions.

For the deployment of relays, the system should focus on the number of relays and the candidate locations to deploy the relays. Because repeaters can be expensive, it is often reasonable to limit the number of repeaters. Since the environment may be complex, e.g. adjacent to lakes, rivers, and the application scenario has some limitations, situations may arise where repeaters cannot be deployed somewhere. This means that the candidate locations to deploy the repeater may be limited.

Due to the power consumption of wireless communication, the repeater and the terminal must remain powered on all the time to receive signals, which consumes power. In addition to this energy consumption, there is also an energy loss rate during radio frequency transmission. The attenuation of energy increases with distanceIs large. In the system of this patent, we only consider the energy loss during the radio frequency transmission, since it is much larger than the energy consumption when the repeater or terminal remains powered on. The relation between the communication energy consumption EC and the distance d can be expressed as a simple energy consumption function, namely EC (d) ═ c × d^qWhere c and q are constants and the value of c is determined by a physical parameter of the material, 2<＝q<4. Since q is not less than 2, d^q>d₁ ^q+d₂ ^qWherein the total distance d ═ d₁+d₂，d₁Is the distance between the signal and the repeater, d₂Is the distance between the repeater and the base station. With such a simple energy consumption function, the repeater can be used not only to extend the wireless transmission range but also to reduce the communication cost.

For wireless charging, there is an effective distance, called the charging radius R, between the repeater and the charging terminal of the fixed terminal_cha. Its value is determined by the type and environment of the repeater. In our proposed system we assume that the charging radius of all repeaters is the same. The charging radius is generally smaller than the wireless communication radius. In addition, in the wireless charging process, besides radio frequency transmission, a circuit for converting radio frequency into energy is also included, and the energy loss rate is usually not lower than 50%, so that the energy loss rate is very high. With the equipment known, the total energy loss rate may be greater than 80% at a charging distance of 5 meters. Therefore, in the sustainable wireless sensor network system of the present patent, it is reasonable to plan one-hop charging.

For the network in our proposed system we assume a fixed aggregation node s which connects the internal communication network with the external internet. Each repeater or terminal may receive and transmit radio frequency signals. In the real world, due to noise and energy loss, the distance between the radio frequency transmitter and receiver cannot be greater than a threshold, called the communication radius. We assume that the communication radius of the terminal, relay and receiver are the same. For wireless communications, the communication radius typically does not exceed 200 meters. For wireless charging, the radius does not exceed 20 meters. Of course, the exact value of the radius is determined by the type, environment and application of the repeater or terminalDetermined by itself. In view of the diversity of terminals, we assume that there is no communication and charging between terminals. However, repeater a can communicate with any repeater or terminal b if Dist (a, b) ≦ R_comWhere Dist (a, b) is the distance between a and b, R_comIs the wireless communication radius.

Fig. 2 is an example of a communication and charging network in the system. In the network, c, e, f and g are deployed relay nodes. We omit all candidate location nodes except the deployed relay nodes. a. b and d are fixed terminal nodes. h is a sink node.

Radius of charge R_cha15m, communication radius R_com25 m. Paths a → c → g → h, b → c → g → h, d → f → g → h are the effective communication paths of the fixed terminal nodes a, b, d to h, respectively. At this time, the terminal c can charge a and b, because Dist (a, c) is Dist (b, c) 10m ≦ R_cha. Relay f cannot charge terminal d because Dist (d, f) is 25m>R_chaRoute d → e → g → h, Dist (e, g) ═ 50m>R_comIt is an invalid communication path. Therefore, in this case, the sustainable wireless sensor network system must use at least 4 repeaters, although only 3 repeaters are used for communication.

Because the positions of the fixed terminal and the sink node are determined, the problem of location selection of the relay is mainly solved when a sustainable wireless sensor network system is built.

Suppose a scenario of a communication system is a three-dimensional space (R)³)。S_terIs a group of fixed terminal nodes, s is a fixed sink node, and aims to deploy a group of relay nodes, so that: (1) for each terminal node v ∈ S_terThere is one relay node u ∈ S_relAnd Dist (u, v) is less than or equal to R_cha，S_relRepresents a set of repeater nodes; (2) for each terminal node v ∈ S_terThere is a communication path P from v to the sink node s_v(one routing path) such that P_vThe distance between any two adjacent nodes is not more than R_com(ii) a (3) Total communication overhead (energy consumption)

Is at a minimum, wherein

In this function, there is an implicit assumption about the total communication cost function that each terminal operates at the same frequency.

The problem of sustainable communication: given distribution in three-dimensional space R³A group of fixed terminal nodes S_terA set of candidate location nodes S_locAnd a sink node S, wherein a communication cost function is EC, and the wireless communication radius and the wireless charging radius are R respectively_comRadius R_chaSelecting a maximum of k candidate node positions (the selected candidate position nodes are used for deploying relay nodes) as much as possible, so that:

(1) for each fixed terminal node v ∈ S_terOne Dist (u, v) is less than or equal to R_chaThe relay node u of (1);

(2) for each terminal node v ∈ S_terThere is a path P from v to the sink node s through the relay node_vIn which P is_vThe distance between any two adjacent nodes is not more than R_com；

(3) The total communication cost is minimal.

The technical scheme for solving the problems is as follows: inputting a set of candidate location nodes S_locA set of fixed terminal nodes S_terA sink node s, a parameter k, a wireless communication radius R_comAnd a wireless charging radius R_chaThe algorithm is as follows (in this context, vertex V and edge E refer to a concept in graph theory):

(A) constructing a graph G based on the input instance of the question;

(B) for each fixed end node, find the shortest path from it to the sink node on G (assuming P is the set of paths for these paths);

(C) finding the minimum set D (charging node set) of candidate position nodes, enabling each terminal node u to have at least one relay node v E D, and enabling Dist (u, v) to be less than or equal to R_com；

(D) If repeater location vertex V_relThe number of vertices in (b) is greater than the repeater number limit (parameter k), then V is reduced using a local search method_rel＝V_locNumber of repeater nodes (V) of n (D @ (U @) (P)) (D @ (U @) (V (P)))_locVertices representing candidate location nodes).

The whole algorithm can be divided into two steps. The first step, consisting of steps (a) (B) (C), is a subroutine for selecting a set of candidate location nodes that satisfy communication conditions and sustainability conditions, but may not satisfy the relay number limit. Algorithm 1 (called CPG) expresses this process.

The second step (D) is to reserve k by two different algorithms (Algorithm 2 and Algorithm 3), called P-SC (Algorithm 2) and G-SC (Algorithm 3), respectively.

Therefore, the first algorithm we solve the problem consists of CPG and P-SC. The second consists of CPG and G-SC.

Now, details of steps (a) to (D) of the algorithm are described in detail.

1) Step (A):

constructing an undirected graph G (V) with edge weighting and coloring_loc∪V_ter∪{V_sink})，E＝(E_grey∪E_black) Is as follows (wherein V)_terIs a vertex, V, representing the location of the terminal device_sinkBeing vertices representing positions of the receiving ends, also called converging vertices, E_greyIs a gray edge, E_blackIs a black border):

(a) for S_locCreating a candidate position vertex v for each candidate position node a in_aAnd is added to V_locPerforming the following steps;

(b) for S_terCreating a terminal vertex v for each terminal node a in (1)_bAnd adding V_ter；

(c) For a sink node s, a sink vertex v is created_sink；

(d) For each terminal node a and each candidate position node b, if Dist (a, b) ≦ R_chaThen a black border is created (v)_a,v_b) Weight c-Dist (v)_a,,v_b)^qIs added to E_grey；

(e) For each candidate location node a and each candidate location (or terminal) node b, if Dist (v)_a,v_b)≤R_comAnd there is no black border between them, a gray border (v) is created_a,v_b) Weight c-Dist (v)_a,v_b)^qIs added to E_black。

2) Step (B) of assigning V e V to each vertex_terThe algorithm uses the Dijsktra algorithm to find the shortest path from V to the convergence vertex.

3) Step (C) of finding a V_terAnd makes it as small as possible, the charging node set D (i.e., the relay node). In our algorithm, the set of vertices D can be divided into two sets D₁And D₂Wherein

Comprises V_terAll vertices where a vertex is bordered by a black edge. In the following, we show just how to find the set of vertices D₂. Let G_black＝((V₁,V₂)，E_black) Is a set of edges E_blackDerived graph, where V₁＝V_ter,V₂＝V(E_black)\V₁. From the construction of G we know that G_blackIs a two-part graph which is composed of a plurality of parts,

let V₂ ^*＝V₂\D₁,V₁ ^*＝V_ter\N_Gblack(D₁) In which N is_Gblack(D₁) Is D₁Is in G_blackSet of vertices in (1). Let G^*＝(V^*＝(V₁ ^*,V₂ ^*)，E^*) Is formed by a set of vertices V^*Derived G_blackIs shown in the figure. Then, we use a simple greedy algorithm to find a minimum set

At G^*Middle domination of V₁ ^*All nodes of (1). Greedy strategy is from V₂ ^*Selecting a vertex with the maximum number of G steps until V₁ ^*All nodes of (2) are controlled, a set of vertices D₂Consisting of all these selected vertices.

4) Step (D) if | V_rel＝V_loc∩(D∪V(P))|>k, then we must reduce V_relUntil it is no greater than k. To achieve this goal, we have two different approaches (two different local search rules). Two different heuristic algorithms are introduced, namely an algorithm P-SC (algorithm 2) and an algorithm G-SC (algorithm 3). Set up the drawing G_PIs the union of all paths in P. If the vertex V ∈ V_relD is G_PThe second vertex in (1), we paint it black. The strategy of our local search rules is to remove black vertices under traffic conditions and sustainability conditions. Let v be P (a, v) in path P_sink) A black vertex above, V is a at P (a, V)_sink) Two neighbors of (1), where v⁺，v^-Are each at G_PPaths P (a, v) and P (v, v)_sink) The above. In the first rule, we consider the case (v)⁺,v^-) E.g. E. If so, we include (v) in P⁺,v,v^-) Is used in all paths of (v)⁺,v^-) Substituted sub-path (v)⁺,v,v^-)。

After such an operation has been carried out, V_relThe number of vertices in (1) is reduced. The algorithm P-SC performs this operation repeatedly until | V_relAnd (5) less than or equal to k. In the second rule, we refine the selection of black vertices as follows. A local search between any two nodes in the entire graph is found. However, the longer the runtime, the lower the communication cost. For the algorithm P-SC, the total run time is (O (| V³)). The running time of the algorithm G-SC is (O (| E | | V |)³) Greater than the run time of the P-SC. If there is a vertex u e (V (P) \ V)_ter) So that (u, v)⁺) E, then IP (v) is contained in the P for people⁺，v_sink) All paths of (v)⁺U) and P (u, v)_sink) Instead of the subpath P (v)⁺，v⁺) E, wherein P (v)⁺，v_sink)，P(u，v_sink) Are each G_PFrom V⁺To v_sinkAnd from u to v_sinkThe path of (2). Such implementation of operations causes an increase in communication costs, but v_relThe number of vertices in (1) is increased by at least 1. For the greedy strategy, we define an evaluation index as

Wherein Δ_cost(v, u) and Δ_relay(v, u) is the inclusion of (P (v) in P⁺，v_sink) Will sub-path P (v) of all paths⁺，v_sink) Replacement by (v)⁺，u)∪P(u，v_sink) Then, in order to make the increment as small as possible, we select the vertex having the minimum value Σ (v) among all the vertices in B₀) Vertex v of₀And performing a path replacement operation as a local search rule for each step in the algorithm G-SC until a set of relay vertices of at most k is found.

A. Performance analysis

As described above, the function of energy consumption (communication cost) is ec (d) kdⁿAnd n is more than or equal to 2 and less than or equal to 4. For simplicity, we set ec (d) ═ d in the experiment². In consideration of practical situations, let us set the wireless communication radius R_com40m, wireless charging radius R _cha15. Each experiment was run 1000 times and the values for each result were averaged. All terminals, candidate locations and receptions are randomly generated in a limited three-dimensional space.

The results of simulation experiments of the CPG algorithm are given in fig. 3. There are four different simulation scenes, the sizes of which are 100m × 10m, 150m × 15m, 200m × 20m and 250m × 25m, and the number of nodes at the candidate positions in the scenes is 100, 225, 400 and 625. In the scenario, the number of end nodes has 5 different values, 10, 20, 30, 40 and 50 respectively. It is not difficult to see that the communication cost and the number of charging nodes increase as the number of terminal nodes increases. The reason is that the total communication cost is related to the number of communication paths, which is determined by the number of terminal nodes.

The invention designs a sustainable wireless sensor network communication system which can maintain the characteristic of sustainable communication by carrying out relay charging on a terminal.

Claims (1)

1. A method for constructing a sustainable wireless sensor network system based on a heuristic algorithm is characterized by comprising the following steps: the sustainable wireless sensor network system based on the heuristic algorithm comprises a plurality of fixed terminals, a sink node and a plurality of repeaters, wherein the repeaters can transmit data and wirelessly charge the terminals; the method comprises the following steps:

step 1, inputting a group of candidate position nodes S_locA set of fixed terminal nodes S_terA sink node s, a parameter k, a wireless communication radius R_comAnd a wireless charging radius R_cha；

Step 2, constructing a graph G based on the input;

step 3, for each fixed terminal node, searching the shortest path of the sink node on the G from the fixed terminal node, and setting P as a path set of the paths;

step 4, finding the minimum set D (charging node set) of the candidate position nodes, enabling each terminal node u to have at least one relay node v to belong to D, and enabling Dist (u, v) to be less than or equal to R_com；

Step 5, if V_relIf the number of vertices in the set is greater than the number limit of repeaters, then V is reduced using a local search method_rel＝V_locThe number of relay nodes of n (D @ (U @) (P)).

Images