A drone-assisted post-disaster clustering mine IoT data collection method

By using a clustered method for collecting IoT data in mines with the assistance of drones, the energy consumption of drones and nodes in post-disaster mines has been optimized, solving the problems of limited drone energy and depleted node energy, and improving the reliability and efficiency of IoT data collection in post-disaster mines.

CN114727359BActive Publication Date: 2024-11-12BEIJING JIAOTONG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210355564.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-04-06
Publication Date
2024-11-12
Estimated Expiration
2042-04-06

AI Technical Summary

Technical Problem

Existing technologies for drone-assisted post-disaster mine IoT data collection cannot effectively solve the problems of drone energy limitations and rapid energy depletion of nodes, leading to data transmission interruptions and failing to meet the reliability requirements of data collection in post-disaster mines.

Method used

A clustered data acquisition method for mine IoT assisted by drones is adopted. By using a post-disaster mine IoT node clustering algorithm and a drone data acquisition path planning algorithm, the energy consumption of the cluster head node and the drone flight path are optimized, thereby reducing the total energy consumption of the system.

Benefits of technology

It extended the lifespan of the post-disaster mine IoT network, improved the reliability and efficiency of data collection, and reduced the energy consumption of cluster head nodes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114727359B_ABST
    Figure CN114727359B_ABST
Patent Text Reader

Abstract

The present invention provides a post-disaster clustering mine Internet of Things data collection method assisted by a drone. The method comprises: clustering all Internet of Things nodes in a post-disaster mine through a post-disaster mine Internet of Things node clustering algorithm; finding a group of cluster heads that minimize the energy consumption of the mine Internet of Things data collection system in each cluster through a post-disaster mine Internet of Things drone data collection path planning algorithm, and determining the order of the group of cluster heads as the data collection path of the drone for each cluster head; the drone starts from a starting point, traverses all cluster heads according to the planned flight path to collect data, and finally returns to the starting point. The present invention reduces the energy consumption of cluster head nodes by using drones to assist the mine Internet of Things to complete the post-disaster data collection task, and the proposed drone path planning algorithm comprehensively considers the data collection energy consumption of the Internet of Things nodes and the drone, thereby extending the network survival period of the post-disaster mine Internet of Things.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of media communication technology, and in particular to a UAV (Unmanned Aerial Vehicle)-assisted post-disaster clustering mine Internet of Things data collection method. Background Art

[0002] During mine disaster rescue, quickly and accurately determining the location of trapped people is a key element of accident rescue. The traditional rescue method is for rescuers to enter the disaster area through tunnels to carry out rescue. However, during disasters, the tunnel environment is complex and changeable. If you pass through rashly, it is very dangerous and easy to cause secondary injuries, which brings great difficulties to rescue detection in the disaster area. At present, rescue robots are used to replace rescuers to enter the disaster area, obtain a variety of environmental parameters in time, and transmit them back to the rescue command in real time. However, rescue robots have disadvantages such as large size, poor flexibility and weak obstacle crossing ability, which cannot effectively meet actual needs. UAVs have significant advantages such as small size, low cost, no risk of casualties and strong flexibility. They have been widely used in many fields such as target tracking, emergency communications and environmental monitoring.

[0003] Mine detection drones can enter underground tunnels for real-time detection, use sensors to collect and process information such as gas concentration, oxygen concentration, ambient temperature and location of injured persons, and transmit this information to the rescue command center in real time via data and audio and video, providing decision-making basis for rescue commanders, so as to minimize the possibility of casualties.

[0004] At present, a routing method for multi-hop transmission of IoT nodes in post-disaster mines in the prior art includes: in view of the fact that it is difficult to replenish the energy of IoT sensor nodes in underground mines after disasters in a timely manner, an energy-saving clustering routing protocol based on two-stage cluster head election is proposed. In the initial cluster head election stage, the protocol introduces node energy factors to select nodes with energy levels higher than the average energy of the network as candidate cluster head nodes; according to the connectivity of adjacent nodes, a connectivity factor is introduced, and an election probability function for the post-disaster reconstruction network is designed based on the energy factor and the connectivity factor. Cluster head election is performed by calculating the election probability function to form an initial clustering structure. In the network operation stage, a cluster head election mechanism is constructed, and whether to replace the cluster head is determined by periodically comparing the remaining energy of each cluster head node with the threshold energy.

[0005] The disadvantages of the routing method of multi-hop transmission of mine IoT nodes after a disaster in the above-mentioned prior art are: this method ignores the problem that the number of remaining nodes after the disaster cannot meet the coverage of the entire tunnel area, and a large amount of monitoring data transmission will cause the surviving nodes to die too quickly, thereby interrupting the data communication process between the ground and the accident tunnel.

[0006] Therefore, in order to meet the reliability requirements of post-disaster mine IoT data collection, a UAV (Unmanned Aerial Vehicle)-assisted post-disaster mine IoT data collection system architecture was constructed.

[0007] A method for collecting wireless sensor network data based on drone assistance in the prior art includes: using a robot or drone as a mobile aggregation node to collect data from all IoT nodes through a planned path. The use of robots or drones significantly reduces the energy consumption caused by multi-hop transmission between nodes, increases the life of the network, and reduces transmission delays. In addition, it can improve the routing holes and hot spots that often occur in post-disaster mine IoT. At present, research on the application of drones in mine IoT mainly focuses on positioning and hardware analysis, and there is very little research on post-disaster mine IoT data collection methods assisted by drones.

[0008] The disadvantage of the wireless sensor network data collection method based on drone assistance in the above-mentioned prior art is that the method cannot be directly applied underground. There are the following problems in using drones to assist in data collection in post-disaster mines: first, due to the limited energy carried by the drone itself, the service range of the drone will be limited by the long-term and large-scale flight restrictions in the mine tunnels; second, the battery life of the surviving IoT nodes is limited, and in most cases it is difficult to replace the battery. Frequent data communication with drones will cause the node energy to be quickly exhausted, thereby interrupting data transmission with the drone. Summary of the invention

[0009] The embodiment of the present invention provides a drone-assisted post-disaster clustering mine Internet of Things data collection method, so as to effectively utilize drones to assist the mine Internet of Things to complete the post-disaster data collection task.

[0010] In order to achieve the above object, the present invention adopts the following technical scheme.

[0011] A drone-assisted post-disaster clustering mine IoT data collection method, comprising:

[0012] All IoT nodes in the post-disaster mine are clustered using the post-disaster mine IoT node clustering algorithm;

[0013] Through the post-disaster UAV data collection path planning algorithm, a group of cluster heads that minimize the total energy consumption of the mine IoT data collection system is found in each cluster, and the order of this group of cluster heads is determined as the flight path of the UAV;

[0014] The UAV starts from the starting point, traverses all cluster heads to collect data according to the planned flight path, and finally returns to the starting point.

[0015] Preferably, clustering all IoT nodes in the post-disaster mine using a post-disaster mine IoT node clustering algorithm includes:

[0016] The post-disaster tunnel area is set as a strip-shaped rectangular space with a length of L and a width of W. The N IoT nodes surviving in the post-disaster tunnel area are randomly scattered throughout the network. The IoT nodes are used to collect and transmit environmental parameters at the accident site. It is assumed that the N IoT nodes need to be divided into K clusters, and the cluster set is represented as {Q1,Q2,…,Q k ,…,Q K}, the number of nodes contained in each cluster is represented by |Q k |, k∈[1,K], the cluster head set is represented as {C1,C2,C k ,,C K};

[0017] The optimization objective function of K value is as follows:

[0018]

[0019] Among them, φ is a weighting factor used to weigh the node energy consumption and the UAV trajectory length, E k n It's Q k Midpoint S k n Energy consumption, d UAV→k Indicates that the drone is in cluster Q k distance from the center;

[0020] Assume that the value of K is [K min ,K max ],0<K min ≤K≤K max <N, use the variance of the distance from the node to the cluster center as the value of the cost function, calculated as follows:

[0021]

[0022] Draw the relationship curve between the cost function value SSE and K. When the relationship curve is elbow-shaped, the K value corresponding to the inflection point of the elbow is determined as the clustering number K value;

[0023] Based on the clustering number K value, the N Internet of Things nodes are divided into K clusters through the K-means clustering algorithm.

[0024] Preferably, the method of dividing the N IoT nodes into K clusters by using a K-means clustering algorithm based on the clustering number K value includes:

[0025] 1) Determine the number of IoT nodes N and the number of clusters K after the disaster;

[0026] 2) Randomly select K nodes from N IoT nodes as the initial cluster centers, and the distance between the K initial cluster centers is greater than the pre-set distance threshold d th ;

[0027] 3) For each remaining node, calculate its distance to each cluster center The distance between the node newly added to the cluster and the cluster head must be less than the communication radius r0;

[0028] 4) After completing the division of all nodes, update the cluster center of each cluster The coordinates of the new cluster center are the average of the coordinates of the nodes in the cluster, and one round of iteration is completed;

[0029] 5) Using the new cluster center as the similarity measure, repeat steps 3) and 4) until the nodes in each cluster no longer change, and determine the current IoT node clustering result as the final clustering result.

[0030] Preferably, the post-disaster UAV data collection path planning algorithm is used to find a group of cluster heads that minimize the total energy consumption of the mine Internet of Things data collection system in each cluster, and the order of the group of cluster heads is determined as the flight path of the UAV, including:

[0031] The drone collects all clusters {Q1, Q2…, Q K} data, given the initial position S0 of the drone, the original flight path of the drone is expressed as Define the total energy consumption E of the post-disaster drone-assisted mine IoT data collection system tot is the sum of the total energy consumption of IoT nodes and drones in a round of data collection. The input sequence of the post-disaster mine IoT drone data collection path planning algorithm is {S0,Q1,Q2,…Q K} consists of the starting point of the drone and all clusters, and the output sequence {S0,S1,S2,…S K ,S0} is the access order of a selected set of cluster heads, that is, the data collection path of the UAV. The pointer network of the post-disaster mine IoT UAV data collection path planning algorithm is composed of encoder and decoder networks. The encoder network is used to obtain To represent each element in , a recursive neural network constructed with long short-term memory (LSTM) units is used as the encoder network;

[0032] The post-disaster mine IoT drone data collection path planning algorithm is described as follows:

[0033] Step 1: Change the original path Input to the encoder network, use LSTM unit to convert the input sequence Converted into a sequence of latent memory states ε={e0,…,e K},

[0034] Step 2: Convert the encoder output sequence ε = {e0,,e K} is input to the decoder network. At each time t, the decoder uses the hidden state h of the previous time knowledge t and {e0,…,e t} Output an access decision π t , get the input sequence According to this input sequence Construct a (K+2) search graph of one level for all clusters;

[0035] Step 3: In order to traverse all nodes, during the path search process, the candidate node set to be queried is set to the OPEN list, and the queried nodes are set to the CLOSED list. The neighbor nodes of any node at any layer are defined as all the nodes in its previous and next layers. Each node has a pointer to its parent node, which is implemented by the mapping COME_FROM.

[0036] Initialize OPEN, CLOSED tables and COME_FROM, set f(S0) = 0, and add the initial point S0 to the OPEN table;

[0037] Step 4: Find the node q with the minimum f(q) value in the OPEN table, obtain all neighboring nodes of node q, and calculate the total energy consumption cost for any neighboring node m, that is, Cost = g(q) + E(q→m), where E(q→m) represents the energy consumption generated by the UAV from node q to node m.

[0038] Step 5: If the neighbor node m is in OPEN and cost < g(m), remove the node from OPEN; if the neighbor node m is in CLOSED and cost < g(m), remove the node from CLOSED; otherwise, let g(m) = cost, f(m) = g(m) + h(m), add node m to OPEN, and set its parent node to q;

[0039] Repeat steps 4-5 until the OPEN table is empty. If q = S0, a path from S0 to S0 is constructed through COME_FROM, and the minimum energy consumption E is output. tot and new collection paths

[0040] It can be seen from the technical solution provided by the above-mentioned embodiments of the present invention that the embodiments of the present invention reduce the energy consumption of cluster head nodes by utilizing drones to assist the mine Internet of Things in completing post-disaster data collection tasks. The proposed drone path planning algorithm comprehensively considers the data collection energy consumption of the Internet of Things nodes and drones, thereby extending the network survival period of the mine Internet of Things after the disaster.

[0041] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0043] Figure 1 A schematic diagram of the architecture of a post-disaster UAV-assisted clustered mine Internet of Things system provided by an embodiment of the present invention;

[0044] Figure 2 A flowchart of a post-disaster mine IoT node clustering algorithm proposed in an embodiment of the present invention;

[0045] Figure 3 A schematic diagram of a post-disaster mine IoT drone data collection path planning algorithm proposed in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0047] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items.

[0048] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0049] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0050] Under normal circumstances, the mine Internet of Things uses a large number of IoT nodes randomly deployed in the tunnels to complete the data collection and transmission tasks of the underground environment and personnel information. However, mine accidents will cause most of the IoT nodes to be damaged, and the surviving IoT nodes cannot guarantee reliable data transmission. Therefore, after the accident, drones can be used to enter the mine to complete the data collection and transmission tasks of the accident tunnel. Energy efficiency optimization of drones and IoT nodes is a key issue that needs to be solved in post-disaster mine Internet of Things data collection. In order to improve the energy consumption problem of the post-disaster mine Internet of Things data collection system and extend the network life of the mine Internet of Things, an embodiment of the present invention proposes a drone-assisted post-disaster clustered mine Internet of Things data collection method. It plays an important role in improving the energy efficiency and reliability of the post-disaster mine Internet of Things data collection system.

[0051] A post-disaster UAV-assisted clustered mine IoT system architecture constructed by an embodiment of the present invention is as follows: Figure 1As shown in the figure, the post-disaster tunnel area is set as a strip-shaped rectangular space with a length of L and a width of W. After the disaster, a rescue drone is used to collect data. The surviving N IoT nodes are randomly scattered throughout the network for the collection and transmission of environmental parameters at the accident site. Assume that all AP (Access Point, wireless access point) devices in the tunnel are damaged after the disaster, and only one Sink (aggregation) node is deployed at the entrance of the tunnel. The surviving IoT nodes are divided into K clusters according to a certain clustering algorithm. Only one node in each cluster is selected as the cluster head, and the member nodes transmit the collected data to the cluster head. Then, the drone moves at a constant speed v UAV Starting from the starting point S0, it traverses all cluster heads according to a certain collection path to collect data, and finally returns to the starting point S0.

[0052] Will Figure 1 The set of clusters in is represented as {Q1,Q2,…,Q k ,,Q K}, the number of nodes contained in each cluster is represented by |Q k |, k∈[1,K], the cluster head set is represented by {C1,C2…,C k ,,C K}. The sampling point set of the drone is represented as {S1,,S K}. Using a three-dimensional coordinate system to define the positions of all nodes and drones, cluster Q k The cluster head coordinates are C k (x k ,y k ,0). Assume that the drone is at the cluster head C k Data is collected directly above the sampling point S k The coordinates of (x k ,y k ,H), where H is the fixed altitude of the drone. k The nth member node in The coordinates of n=1,…,|Q k |-1. Assume that the Sink node knows the initial information of all surviving nodes, each IoT node has a unique ID and the same communication radius r0, and the node can exchange information with other nodes or drones within the communication range. In addition, all nodes calculate and store the location information of the nodes according to the existing positioning algorithm, and the drone is equipped with a satellite positioning device for real-time positioning. Since the battery of the node is difficult to replace in the post-disaster mine environment, the IoT node will die if the initial energy is exhausted.

[0053] exist Figure 1On the basis of the above, an embodiment of the present invention proposes a drone-assisted post-disaster clustering mine Internet of Things energy-saving data collection method. The method first proposes a post-disaster mine Internet of Things node clustering algorithm to complete the clustering of all Internet of Things nodes, and then proposes a post-disaster mine Internet of Things drone data collection path planning algorithm to find a group of cluster heads that minimize the energy consumption of the mine Internet of Things data collection system in the determined clusters. The order of the cluster heads is the data collection path of the drone.

[0054] The flowchart of a post-disaster mine IoT node clustering algorithm proposed in an embodiment of the present invention is as follows: Figure 2 As shown, the processing includes the following:

[0055] from Figure 1 It can be seen that, under the condition of a certain tunnel area, as K increases, the distance between nodes in the cluster becomes closer and closer, so the transmission distance between the node and the cluster head gradually decreases, and the energy consumption of the node naturally decreases. However, at the same time, the number of cluster heads that the drone needs to visit increases, the trajectory length of the drone becomes longer, and the energy consumption of the drone also increases. Therefore, the determination of the K value affects the performance of the post-disaster mine IoT clustering, and the optimization of the K value needs to comprehensively consider the energy consumption of the IoT node and the trajectory length of the drone. Its objective function is as follows:

[0056]

[0057] Among them, φ is a weighting factor used to weigh the node energy consumption and the UAV trajectory length, and φ is set to 0.5. It's Q k Midpoint Energy consumption, d UAV→k Indicates that the drone is in cluster Q k Distance from center.

[0058] Assume that the value of K is [K min ,K max ],(0<K min ≤K≤K max <N), and then by plotting the relationship between cost and K in this interval, we can get the value that affects the cost function. Using the distance variance from the node to the cluster center, that is, the sum of squared errors (SSE) as the cost function, the calculation is as follows:

[0059]

[0060] SSE reflects the density of nodes in each cluster after clustering. The smaller the SSE value, the denser the sample points in each cluster. The basis for selecting the optimal K is that when K is less than the optimal number of clusters, the increase in K greatly accelerates the convergence of nodes in each cluster, thereby greatly reducing the SSE of the nodes. When K reaches the optimal number of clusters, the convergence obtained by increasing K rapidly decreases, so the SSE drops sharply, and then gradually flattens as K continues to increase. This means that the relationship curve between SSE and K is elbow-shaped, and the value corresponding to this inflection point is the optimal K.

[0061] Then, the K-means clustering algorithm is used to divide the post-disaster IoT nodes, aiming to divide N nodes into K clusters and ensure that the nodes in the clusters of the post-disaster mine IoT system are as compact as possible and the nodes between different clusters are as separated as possible. The Euclidean distance is used as the similarity measure of clustering, and the distance from the node to the cluster center is used as the optimization objective function. Figure 1 In the example, we assume that the cluster set is Q k The nth node in to the cluster center The distance is calculated as

[0062]

[0063] According to the distance from the node to the cluster center, the optimal cluster with the smallest node distance is obtained through multiple iterations. Then the objective function of clustering is

[0064]

[0065] Therefore, the post-disaster IoT node clustering algorithm based on a given K value is described as follows:

[0066] 1) Determine the number of IoT nodes N and the number of clusters K after the disaster.

[0067] 2) Randomly select K nodes from N IoT nodes as the initial cluster centers. Since the initial cluster center affects the clustering results and the number of iterations, the distance between the K initial cluster centers must be greater than the preset distance threshold d th Assuming that p (0<p<K) initial cluster centers are selected, when selecting the p+1th cluster center, only the one whose distance to the p cluster centers is at least d th The nodes may be selected.

[0068] 3) For each remaining node, calculate its distance to each cluster center The distance between the newly added node and the cluster head must be less than the communication radius r0.

[0069] 4) After completing the division of all nodes, update the cluster center of each cluster The coordinates of the new cluster center are the average of the coordinates of the nodes in the cluster, and one round of iteration is completed.

[0070] 5) Using the new cluster center as the similarity measure, repeat steps 3) and 4) until the nodes in each cluster no longer change, and determine the current IoT node clustering result as the final clustering result.

[0071] A schematic diagram of a post-disaster mine IoT drone data collection path planning algorithm proposed in an embodiment of the present invention is shown in FIG. Figure 3 As shown in FIG. 1 , the specific processing process includes: Figure 1 In the system shown in the figure, after the disaster, all IoT nodes are divided into K clusters, and the drone needs to collect all clusters {Q1, Q2…, Q K Given the initial position S0 of the drone, the original flight path of the drone is expressed as The total energy consumption of the post-disaster drone-assisted mine IoT data collection system is defined as the sum of the total energy consumption of the IoT nodes and the total energy consumption of the drone in one round of data collection, which is calculated as follows:

[0072]

[0073] Among them, ω (0≤ω≤1) is the weight coefficient for adjusting the energy consumption of all IoT nodes and drones after the disaster, and the value here is 0.5.

[0074] For member nodes To cluster head C k Energy consumption for transmitting data, E k is the cluster head C k Energy consumption of receiving data), is the cluster head C k Energy consumption of transmitting data to the drone, E flight is the energy consumption of the UAV flight), For drones in S k The sampling energy consumption at .

[0075] The energy consumption model of IoT nodes is:

[0076] The first-order radio model is used to calculate the energy consumption of IoT nodes. The energy consumption of the transmitting node includes the transmitting circuit and the power amplifier, while the energy consumption of the receiving node is generated by the receiving circuit. If the distance between the transmitting node and the receiving node is less than the threshold, the free space model is used. Otherwise, the multipath fading model is used. Figure 1 In the member node To cluster head C k The energy consumption of transmitting l-bit data is calculated as follows:

[0077]

[0078] Among them, E elec It is the circuit energy consumed to transmit or receive 1 bit of data. is a member node n (n=1,…,|Q k |-1) and cluster head C k The distance between. d0 is the distance threshold, ε fs and ε mp represent the energy parameters of the RF amplifier of the free space model and the multipath fading model respectively. k The energy consumption of receiving l-bit data is calculated as

[0079] E k =lE elec (8)

[0080] Assuming that the amount of data collected by all member nodes in each round is l bit, the cluster head C k The energy consumed in transmitting data to the UAV is expressed as

[0081]

[0082] Among them, T k Is UAV in S k The hovering time.

[0083] In a complete round of data collection tasks, member nodes transmit data to their respective cluster heads, and the cluster heads forward the data to the UAV. The total energy consumption of all IoT nodes is E IoTN Expressed as

[0084]

[0085] The energy consumption model of the drone is: Figure 1 As shown, when the UAV is moving at a fixed speed v UAV Fly to sampling point S k After that, it transmits a beacon frame and sends the corresponding cluster head C k Then, the cluster head C k It starts to collect data from its member nodes through time division multiple access and forwards the collected data to the UAV. The energy consumption of a round of data collection by the UAV is mainly composed of flight energy consumption and sampling energy consumption, among which the sampling energy consumption of the UAV includes hovering energy consumption and communication energy consumption.

[0086] The hovering power of the UAV is calculated as follows

[0087]

[0088] Where g is the earth's gravity, ρ is the air density, and n p is the number of rotors, r p is the rotor radius, m tot is the mass of the UAV. For the convenience of analysis, it is assumed that the hovering time of the UAV is equal to the data transmission time from the cluster head to the UAV. k The sampling energy consumption at is calculated as

[0089]

[0090] Among them, P hover and P com They are the hovering power and communication power of the UAV, is the cluster head C k The total amount of data transmitted to the UAV.

[0091] Assume that the horizontal movement power is the UAV flight speed v UAV The linear function of

[0092]

[0093] Among them, v max is the maximum flight speed of the UAV, P max and P idle are the hardware power of the UAV when it is moving at full speed and in idle state. Since the UAV needs to hover at the sampling point before departure, the flight energy consumption of the UAV consists of hovering energy consumption and moving energy consumption. The flight energy consumption of the UAV is calculated as follows

[0094] E flight =T flight (P hover +P move ) (14)

[0095] Among them, T flight is the total time of UAV flight, expressed as

[0096]

[0097] Where S={S0,S1,…,S K},S k By C k Sure, YesS i With S j The Euclidean distance between is a binary indicator function. If the UAV starts from point S i Move to point S j , otherwise,

[0098] Therefore, the total energy consumption of one round of UAV data collection is expressed as follows

[0099]

[0100] In order to minimize the total energy consumption E of the post-disaster drone-assisted mine IoT data collection system tot , the data collection path planned by the UAV is expressed as where π t Indicates the original path of the drone Any node in Therefore, for a given input path According to the chain rule, the output path The probability of can be factored by the product of conditional probabilities, namely:

[0101]

[0102] where t is the time step, parameterized by θ A random strategy to determine the order of drone collection. Based on the input original drone path and the clustering information collected at time t-1, the conditional probability P(π t |·) can calculate the probability of other clusters being collected at time t.

[0103] The probability of any cluster being visited at time t is modeled. The trained θ can be used to estimate the energy consumption E of the post-disaster drone-assisted mine IoT data collection system. tot The low path is assigned a higher probability, which increases the system energy consumption E tot The high path is assigned a lower probability. Here, the actor-critic method in the reinforcement learning algorithm can be used to train the optimal model strategy θ * .

[0104] Given the number of clusters K, the post-disaster UAV data collection path planning problem can be regarded as a sequential decision problem. Figure 3 The figure shows a schematic diagram of the path planning algorithm for the post-disaster mine IoT drone data collection. Figure 3 In the input sequence {S0,Q1,Q2,…Q K} consists of the starting point of the drone and all clusters, and the output sequence {S0,S1,S2,…S K ,S0} is the access order of a selected set of cluster heads, i.e., the data collection path of the UAV. Figure 3 In , the encoder and decoder networks form the pointer network of the proposed UAV path planning algorithm. The encoder network is used to obtain The representation of each element in . First, a recurrent neural network (RNN) constructed with long short-term memory (LSTM) units that can learn long-term dependencies is used as the encoder. Each item in includes the starting position S0 and the cluster set {Q1,Q2…,Q K}, are converted into a D-dimensional vector space, allowing the strategy to more effectively extract useful features in the converted space. The embedding vector is then input into the LSTM unit. In each encoding step, the LSTM unit reads an embedding item and outputs a latent memory state. Finally, the input sequence Converted into a sequence of latent memory states ε={e0,…,e K}, exist Figure 3 In the encoder, the output {e0,…,e K} is provided to the decoder network. At each decoding time t, the LSTM unit outputs a hidden state that contains the knowledge of the previous time Then, the decoder uses the attention mechanism based on h t and {e0,…,e t} Output an access decision π t At decoding time t, select the one with the maximum conditional probability P(π t |·) as the nodes to be collected, and the calculation is as follows

[0105]

[0106] Among them, W1, is the attention matrix, is the attention vector, W1, W2 and Represented by θ, θ is the learnable parameter in the proposed UAV path planning algorithm network, is the node j(e j ) associated score.

[0107] The original path is calculated by the normalized exponential function The conditional probability of a node that has not been visited is

[0108]

[0109] Among them, the probability P(π t =j|·) represents the degree to which the model points to node j at decoding time t.

[0110] Based on the output sequence of the decoder network, a search graph is constructed for all clusters, where each layer consists of nodes of one cluster. The starting position of the first layer and the ending position of the last layer are both S0. Therefore, the created graph has a total of (K+2) layers. Then, the optimal cluster head is found from each cluster, and the total energy cost E from the starting position to the ending position is constructed. tot The minimum path. In each iteration, the cost of traversing the path and the estimated cost of extending the path to the end need to be calculated to determine whether to extend its partial path into a longer path. Any node m is selected to visit according to the following function

[0111] f(m)=g(m)+h(m) (9)

[0112] Among them, g(m) represents the exact energy consumption of the system when the drone moves from the starting node S0 to the candidate node m, and h(m) is the estimated energy consumption of the drone from the candidate node m to the end point S0. Then, the node with the smallest f(m) value is selected from the candidate nodes as the next node to be visited.

[0113] In order to traverse all nodes, during the path search process, the set of candidate nodes to be queried is set to the OPEN list. The queried nodes are set to the CLOSED list. The neighbor nodes of any node at any layer are defined as all nodes in its previous and next layers. In addition, each node has a pointer to its parent node, which is implemented by the mapping COME_FROM.

[0114] Therefore, the path planning algorithm for post-disaster mine IoT drone data collection is described as follows:

[0115] Step 1: Change the original path Input to the encoder network, use LSTM unit to convert the input sequence Converted into a sequence of latent memory states ε={e0,…,e K},

[0116] Step 2: Transform the encoder output sequence ε={e0,…,e K} is input to the decoder network. At each time t, the decoder uses the hidden state h of the previous time knowledge t and {e0,…,e t} Output an access decision π t . Get the input sequence According to this input sequence Construct a (K+2) search graph of one level for all clusters;

[0117] Step 3: Initialize OPEN, CLOSED tables and COME_FROM, set f(S0) = 0, and add the initial point S0 to the OPEN table;

[0118] Step 4: Find the node q with the minimum f(q) value in the OPEN table, obtain all neighboring nodes of node q, and calculate the total energy consumption cost for any neighboring node m, that is, Cost = g(q) + E(q→m), where E(q→m) represents the energy consumption generated by the UAV from node q to node m.

[0119] Step 5: If the neighbor node m is in OPEN and cost < g(m), remove the node from OPEN; if the neighbor node m is in CLOSED and cost < g(m), remove the node from CLOSED. Otherwise, let g(m) = cost, f(m) = g(m) + h(m), add node m to OPEN, and set its parent node to q.

[0120] Repeat steps 4-5 until the OPEN table is empty. If q = S0, a path from S0 to S0 is constructed through COME_FROM, and the minimum energy consumption E is output. tot and new collection paths

[0121] In summary, the embodiment of the present invention reduces the energy consumption of cluster head nodes by using drones to assist in post-disaster mine IoT data collection. The proposed drone path planning algorithm comprehensively considers the energy consumption of IoT nodes and drone data collection, and prolongs the network survival period of post-disaster mine IoT. By using clustering to reconstruct the networking of IoT nodes that survive after the disaster, the data collection energy consumption of IoT nodes is significantly reduced.

[0122] The embodiment of the present invention solves the clustering problem of the remaining nodes in the post-disaster drone-assisted mine Internet of Things data collection system. A post-disaster Internet of Things node clustering algorithm is proposed, and the optimal clustering number K is determined by comprehensively considering the data collection energy consumption of the post-disaster nodes and the length of the drone collection path. The path planning problem of the drone in the post-disaster drone-assisted mine Internet of Things data collection system is solved, and a post-disaster mine Internet of Things drone data collection path planning algorithm is proposed. The selected group of cluster heads and cluster head sorting comprehensively consider the collection energy consumption of the Internet of Things nodes and the collection energy consumption of the drone.

[0123] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0124] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0125] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0126] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A drone-assisted post-disaster clustering mine IoT data collection method, characterized in that: include: All IoT nodes in the post-disaster mine are clustered using the post-disaster mine IoT node clustering algorithm; Through the post-disaster UAV data collection path planning algorithm, a group of cluster heads that minimize the total energy consumption of the mine IoT data collection system is found in each cluster, and the order of this group of cluster heads is determined as the flight path of the UAV; The UAV starts from the starting point, traverses all cluster heads to collect data according to the planned flight path, and finally returns to the starting point; The method of clustering all IoT nodes in the post-disaster mine using the post-disaster mine IoT node clustering algorithm includes: The post-disaster tunnel area is set as a strip-shaped rectangular space with a length of L and a width of W. The N IoT nodes surviving in the post-disaster tunnel area are randomly scattered throughout the network. The IoT nodes are used to collect and transmit environmental parameters at the accident site. It is assumed that the N IoT nodes need to be divided into K clusters, and the cluster set is represented as {Q1,Q2,...,Q k ,...,Q K }, the number of nodes contained in each cluster is represented by |Q k |, k∈[1,K], the cluster head set is represented by {C1,C2...,C k ,...,C K }; The optimization objective function of K value is as follows: Among them, φ is a weighting factor used to weigh the node energy consumption and the UAV trajectory length. It's Q k Midpoint Energy consumption, d UAV→k Indicates that the drone is in cluster Q k distance from the center; Assume that the value of K is [K min ,K max ],0<K min ≤K≤K max <N, use the variance of the distance from the node to the cluster center as the value of the cost function, calculated as follows: Draw the relationship curve between the cost function value SSE and K. When the relationship curve is elbow-shaped, the K value corresponding to the inflection point of the elbow is determined as the clustering number K value; Dividing the N IoT nodes into K clusters using a K-means clustering algorithm based on the clustering number K value; The method of finding a group of cluster heads that minimizes the total energy consumption of the mine IoT data collection system in each cluster through the post-disaster UAV data collection path planning algorithm, and determining the order of the group of cluster heads as the flight path of the UAV, includes: The drone collects all clusters {Q1,Q2...,Q K } data, given the initial position S0 of the drone, the original flight path of the drone is expressed as Define the total energy consumption E of the post-disaster drone-assisted mine IoT data collection system tot is the sum of the total energy consumption of IoT nodes and the total energy consumption of drones in a round of data collection. The input sequence of the post-disaster mine IoT drone data collection path planning algorithm is {S0,Q1,Q2,...Q K } consists of the starting point of the drone and all clusters, and the output sequence {S0,S1,S2,...S K ,S0} is the access order of a selected set of cluster heads, that is, the data collection path of the UAV. The pointer network of the post-disaster mine IoT UAV data collection path planning algorithm is composed of encoder and decoder networks. The encoder network is used to obtain To represent each element in , a recursive neural network constructed with long short-term memory (LSTM) units is used as the encoder network; The post-disaster mine IoT drone data collection path planning algorithm is described as follows: Step 1: Change the original path Input to the encoder network, use LSTM unit to convert the input sequence Converted into a sequence of latent memory states ε={e0,...,e K }, Step 2: Convert the encoder output sequence ε={e0,...,e K } is input to the decoder network. At each time t, the decoder uses the hidden state h of the previous time knowledge t and {e0,...,e t } Output an access decision π t , get the input sequence According to this input sequence Construct a (K+2) search graph of one level for all clusters; Step 3: In order to traverse all nodes, during the path search process, the candidate node set to be queried is set to the OPEN list, and the queried nodes are set to the CLOSED list. The neighbor nodes of any node at any layer are defined as all the nodes in its previous and next layers. Each node has a pointer to its parent node, which is implemented by the mapping COME_FROM. Initialize OPEN, CLOSED tables and COME_FROM, set f(S0) = 0, and add the initial point S0 to the OPEN table; Step 4: Find the node q with the minimum f(q) value in the OPEN table, obtain all neighboring nodes of node q, and calculate the total energy consumption cost for any neighboring node m, that is, Cost = g(q) + E(q→m), where E(q→m) represents the energy consumption generated by the UAV from node q to node m; Step 5: If the neighbor node m is in OPEN and cost < g(m), remove the node from OPEN; if the neighbor node m is in CLOSED and cost < g(m), remove the node from CLOSED; otherwise, set g(m) = cost. f(m) = g(m) + h(m), where g(m) represents the exact energy consumption of the system when the drone moves from the starting node S0 to the candidate node m, and h(m) is the estimated energy consumption of the drone from the candidate node m to the end point S0. Node m is added to OPEN, and its parent node is set to q; Repeat steps 4-5 until the OPEN table is empty. If q = S0, a path from S0 to S0 is constructed through COME_FROM, and the minimum energy consumption E is output. tot and new collection paths 2. The method according to claim 1, characterized in that The method of dividing N IoT nodes into K clusters by using a K-means clustering algorithm based on the clustering number K value includes: 1) Determine the number of IoT nodes N and the number of clusters K after the disaster; 2) Randomly select K nodes from N IoT nodes as the initial cluster centers, and the distance between the K initial cluster centers is greater than the pre-set distance threshold d th ; 3) For each remaining node, calculate its distance to each cluster center The distance between the node newly added to the cluster and the cluster head must be less than the communication radius r0; 4) After completing the division of all nodes, update the cluster center of each cluster The coordinates of the new cluster center are the average of the coordinates of the nodes in the cluster, and one round of iteration is completed; 5) Using the new cluster center as a similarity measure, repeat steps 3) and 4) until the nodes in each cluster no longer change, and determine the current IoT node clustering result as the final clustering result.