A resource allocation method for collecting IoT device data using drones

Abstract

An embodiment of the present invention provides a resource allocation method for collecting IoT device data using drones. The method includes: clustering ground IoT devices using an Affinity Propagation (AP) algorithm, calculating the minimum inter-cluster flight time of drones using a double-layer shortest path algorithm, designing the initial trajectory of drones within each cluster, calculating the minimum acquisition time of drones within each cluster, obtaining the shortest cruising time of drones based on the minimum inter-cluster flight time of drones and the minimum acquisition time of drones within each cluster, and outputting the optimal allocation decision for each resource. Aiming at scenarios where massive IoT device data is collected in areas that are not covered by cellular networks, the present invention uses the AP algorithm to cluster all IoT devices, and divides the overall drone process into two parts based on the clustering results. At the same time, multiple resources such as the drone's trajectory and communication resources are jointly optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things resource allocation, and in particular to a resource allocation method for collecting Internet of Things device data using a drone. Background Art

[0002] With the rapid development of sensor and wireless communication technologies, the concept of the Internet of Things (IoT) has moved from imagination to reality. Both industry and academia have proposed numerous inspiring and practical projects, such as smart grids, smart logistics, and smart cities, to promote the implementation of the IoT concept. Data or information collection has become a crucial foundation for realizing IoT functionality. While many existing communication protocols and routing algorithms can enable data collection for IoT and wireless IoT, cellular network connectivity cannot be effectively guaranteed due to the imprecise deployment and large number of IoT devices. These proposed protocols and routing algorithms may even fail to function properly in emergency situations. Furthermore, cost-effectively sensing or collecting data from IoT devices in uninhabited areas is a key research topic. Traditionally, this data has been transmitted from IoT devices to data centers via wireless links. However, it should be noted that the extremely long-distance communication between IoT devices and data centers inevitably results in significant energy consumption and equipment aging. Furthermore, deploying numerous base stations (BSs) in areas beyond cellular network coverage requires significant human and material resources.

[0003] Due to their high maneuverability and decreasing costs, unmanned aerial vehicles (UAVs) have been widely used in many civilian applications, such as navigation, precision agriculture, photography, and geological exploration. Currently, several companies and universities have successfully designed prototypes of UAVs capable of communication. For example, high-altitude UAV base stations include Facebook's Aquila and Google's Loon, while low-altitude UAV base stations include Nokia's F-cell and Eurecom's perfume. Furthermore, due to the high probability of line-of-sight (LoS) link propagation paths, UAVs can provide a greater sensing range and better wireless communication service quality.

[0004] Compared to traditional base station deployment, drone-based communication offers several advantages: Leveraging the high probability of line-of-sight links, the drone's maneuverability, and the flexibility of three-dimensional positioning, drones can more reliably collect data from IoT devices. The increased transmission distance between the drone and IoT device significantly shortens the drone's flight path.

[0005] In some desolate or harsh environments, precise deployment of IoT devices is difficult, and ensuring network connectivity between IoT devices is also challenging, making data collection from IoT devices a challenge. However, using drones to collect data from IoT devices virtually eliminates the need for network connectivity between them, reducing the complexity of IoT device deployment and management.

[0006] While IoT projects bring convenience and high efficiency, they also introduce a vast number of diverse IoT devices. Furthermore, the dispersed and sparse deployment of IoT devices complicates data collection using drones. Collecting data from a large number of IoT devices individually is illogical and impossible for drones, especially when a sufficient number of drones are not available. Therefore, addressing the data collection challenge for massive IoT devices has become a key research focus in both industry and academia. Furthermore, in addition to the dispersed and sparse deployment of IoT devices, different types of IoT devices often produce varying amounts of data. For example, temperature data generated by wireless temperature sensors is typically text-based data, consisting of only tens or hundreds of KB in size; whereas surveillance videos captured by wireless surveillance cameras typically contain tens or hundreds of MB of video data. Collection tasks requiring varying amounts of data also place varying demands on the communication time or rate between drones and IoT devices. IoT devices with more data to collect often require more communication time or bandwidth. Therefore, data collection tasks for IoT devices with varying amounts of data place higher demands on drone trajectory design. Furthermore, to ensure that drones can complete their IoT device data collection tasks, limited wireless spectrum resources must also be considered.

[0007] As an important supplementary technology for future 6G, drone communication is a hot topic in current research. Currently, existing drone communication methods have the following shortcomings:

[0008] Data perception is often done through a one-to-one sequential sensing between drones and IoT devices, which is not feasible for a large number of IoT devices. Furthermore, current research has not focused much on how to collect data when IoT devices are sparsely deployed.

[0009] When considering the types of IoT devices, it is often easy to assume that all IoT devices are of the same type. However, in reality, there is more than just one type of IoT device. An IoT project often requires the coexistence of multiple IoT devices, which inevitably leads to uneven data distribution within the scenario.

[0010] Only the trajectory optimization of the UAV is focused on, while the present invention simultaneously optimizes multiple resources such as the trajectory of the UAV and communication resources. Summary of the Invention

[0011] An embodiment of the present invention provides a resource allocation method for collecting IoT device data using a drone, thereby enabling the drone to collect IoT device data in areas that are not covered by cellular networks.

[0012] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions.

[0013] A resource allocation method for collecting IoT device data using a drone, comprising:

[0014] Use the proximity propagation AP algorithm to cluster the IoT devices on the ground;

[0015] The minimum flight time of UAVs between clusters is calculated using a two-layer shortest path algorithm;

[0016] Design the initial trajectory of drones within each group and calculate the minimum acquisition time of drones within each group;

[0017] The shortest cruising time of the drone is obtained based on the minimum flying time of the drone between clusters and the minimum collection time of the drone within each cluster, and the optimal allocation decision of each resource is output.

[0018] Preferably, the clustering of ground IoT devices using an AP algorithm includes:

[0019] Obtain the coordinates of each IoT device m, randomly select one or more IoT devices k as the center point, which is the center point of a cluster, set the damping factor and the maximum number of iterations, calculate the similarity information, availability and credibility between each pair of IoT devices, and update it by multiplying it by the damping factor to obtain a similarity matrix S. Based on the similarity information, availability information, credibility and damping factor, calculate the iterative convergence of the AP algorithm. Let the number of iterations increase by one to determine whether iterative convergence is achieved. The criterion for iterative convergence is: the number of iterations reaches the maximum value, or it is available. The reliability and credibility information matrix no longer changes significantly, that is, the difference between two consecutive changes is less than the threshold value; if so, several IoT devices whose sum of the reliability and credibility information matrix is ​​greater than 0 are used as cluster heads, and the IoT devices with the maximum availability calculated by a certain cluster head are divided into the cluster of the cluster head, and the clustering results and the center points of each cluster are output; otherwise, the center point is updated, the similarity information, availability information and credibility between the IoT devices and the center point are recalculated, and the iterative convergence of the AP algorithm is recalculated until convergence is reached or the maximum number of iterations is reached.

[0020] Preferably, the method of calculating the minimum flight time of the UAV between clusters using a double-layer shortest path algorithm includes:

[0021] The simulated annealing algorithm is used to solve the traveling salesman problem TSP. The distance that the drone takes off from the starting point, flies through the centers of each cluster and returns to the starting point is set as energy. Then the l∈L sa The energy expression of the iteration is as follows, where L sa is the total number of iterations:

[0022]

[0023] in, The access order is No. o in k The energy update between two iterations satisfies the Metropolis criterion, that is, the new state is accepted with probability. When the energy E obtained in the l+1th iteration is 1+1 Not greater than the energy E obtained in the first iteration l , then accept the energy E obtained in the l+1th iteration l+1 Otherwise, the probability Choose whether to accept the energy E obtained in the l+1th iteration l+1 , the probability update formula is as follows:

[0024]

[0025] After passing L sa The energy difference of the iteration or multiple times is not greater than the threshold value ∈ sa Then the algorithm is terminated and the access order that minimizes the inter-cluster flight distance is obtained.

[0026] The area of ​​each IoT device cluster is described by an ellipse. For the kth cluster, To represent the center of the ellipse, and The calculation formula is as follows:

[0027]

[0028]

[0029] in and are the maximum and minimum horizontal and vertical coordinate values ​​in the kth cluster respectively, and the horizontal and vertical semi-axis lengths of the ellipse are recorded as L k and W k , the calculation formula is as follows:

[0030]

[0031]

[0032] The points inside the kth ellipse are expressed as follows:

[0033]

[0034] Convert the ellipse to the polar coordinate system. In the polar coordinate system, the angle θ between the line passing through the point on the ellipse boundary and the abscissa of the cluster center is k ∈[0,2π) to represent:

[0035]

[0036] The TSPN optimization problem is to optimize the access location of the UAV to minimize the flight distance between the UAVs under a given access order. The dynamic programming algorithm is used to solve the following problem:

[0037]

[0038] Will It is defined as the drone visiting from the starting point to the oth k The boundary points of the cluster The minimum distance, then the update formula of this minimum distance is as follows:

[0039]

[0040] The algorithm is used to obtain the shortest distance between UAVs flying in clusters. and the optimal boundary point e * , thereby calculating the minimum time for the drone to fly between clusters

[0041] Preferably, the designing of the initial trajectory of the drones in each family and the calculation of the minimum acquisition time of the drones in each family include:

[0042] The TSP-based algorithm is used to design the initial trajectory of the UAV within the cluster, and the flight speed of the UAV is set to the minimum speed V min , and the drone uses 1 / N of the space above each IoT device. k For the kth cluster, the drone starts from the boundary point obtained in the previous step, visits each IoT device in turn, and finally returns to the boundary point. The entire distance is recorded as The flight time of the UAV in the cluster is recorded as:

[0043]

[0044] At the same time, the hovering communication time of each IoT device of the drone is:

[0045]

[0046] Therefore, the flight time of the UAV in the kth cluster is T k The maximum estimated value of is:

[0047]

[0048] Given the collection time within a drone cluster, the feasibility of the relevant resources within the cluster, namely the association variable B between drones and IoT devices, drone trajectory Q, and bandwidth allocation ratio Λ, is analyzed. The alternating optimization method is used to split the feasibility problem into three sub-problems to verify the feasibility separately:

[0049] Problem 1) Fix the drone trajectory Q and bandwidth allocation ratio Λ and verify the feasibility of the association variable B between the drone and the IoT device;

[0050] Problem 2) Fix the association variable B between the drone and the IoT device and the bandwidth allocation ratio Λ to verify the feasibility of the drone trajectory Q;

[0051] Problem 3) Fix the drone trajectory Q and the association variable B between the drone and the IoT device, verify the feasibility of the bandwidth allocation ratio Λ, and propose a double-layer loop minimum drone acquisition time algorithm based on a one-dimensional search method. The acquisition time of all clusters is optimized and calculated in parallel;

[0052] The minimum collection time of each cluster obtained by parallel calculation is added together to obtain the total collection time of the UAV during the intra-cluster collection phase.

[0053] Preferably, the method of obtaining the shortest cruising time of the drone based on the minimum inter-cluster flight time of the drone and the minimum acquisition time of the drone within each cluster and outputting the optimal allocation decision of each resource includes:

[0054] The minimum acquisition time of drones within each cluster and the inter-cluster flight time are added together to obtain the shortest cruising time of the drones as a whole, and the optimal allocation scheme of drone communication resources is obtained. This optimal allocation scheme includes the scheduling resources of drones and IoT devices within the cluster, the allocation of communication bandwidth resources between drones and IoT devices, and the trajectory of drones within the cluster.

[0055] As can be seen from the technical solutions provided by the embodiments of the present invention described above, this invention addresses the need to collect massive amounts of IoT device data in areas beyond cellular network coverage. By introducing drone communications and proposing a comprehensive design solution, it avoids the high costs associated with traditional base station deployment. This approach utilizes an AP algorithm to cluster all IoT devices, and based on the clustering results, the overall drone processing is divided into two parts. Furthermore, multiple resources, such as drone trajectories and communication resources, are jointly optimized.

[0056] Additional aspects and advantages of the present invention will be set forth in part in the following description, will become apparent from the following description, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0058] Figure 1 The present invention proposes a schematic diagram of a resource allocation method for collecting IoT device data using a drone.

[0059] Figure 2 A processing flow chart of a resource allocation method for collecting IoT device data using a drone is proposed for an embodiment of the present invention;

[0060] Figure 3 A flowchart of an AP algorithm based on dynamic programming provided by an embodiment of the present invention;

[0061] Figure 4 A flowchart of a two-layer shortest path algorithm provided by an embodiment of the present invention;

[0062] Figure 5 A drone perception coverage map of an elliptical trajectory provided by an embodiment of the present invention;

[0063] Figure 6 A flow chart of an algorithm for initial trajectory of UAVs within a cluster provided by an embodiment of the present invention;

[0064] Figure 7 A flow chart of a double-layer loop minimum drone acquisition time algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0065] 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 are not to be construed as limiting the present invention.

[0066] Those skilled in the art will appreciate that, unless expressly stated otherwise, the singular forms "a", "an", "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 an embodiment of the present invention refers to an element being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. 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.

[0067] 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 commonly understood by those skilled in the art in the art to which the present invention pertains. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art and, unless defined as such herein, will not be interpreted in an idealized or overly formal sense.

[0068] 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.

[0069] This embodiment of the present invention studies a real-world scenario in which a large number of diverse IoT devices are sparsely distributed in an area beyond cellular network coverage. A drone departs from a data center, senses these IoT devices, and returns to the data center. A joint resource optimization scheme is proposed to minimize the drone's cruising time. This scheme considers clustering strategies, association strategies, drone trajectories, and bandwidth allocation strategies, and simultaneously optimizes multiple resources, including drone trajectories and communication resources.

[0070] To address the order in which drones access these large-scale IoT devices, this paper uses the AP (affinity propagation) algorithm to cluster IoT devices. This clustering algorithm saves time and drone energy by comparing the selection of cluster centers to a leader election. Through information exchange between voters and candidates, a suitable leader, or cluster center, is generated.

[0071] Therefore, based on the above cognition, in order to solve the data collection problem of sparsely distributed IoT devices with different data volumes in areas where cellular networks cannot cover, the embodiment of the present invention proposes a resource allocation method for collecting IoT device data using drones. The implementation principle diagram is as follows: Figure 1 The specific processing flow is as shown in Figure 2 As shown, the processing steps include the following:

[0072] Step S10: clustering the IoT devices on the ground using the AP algorithm.

[0073] Step S20: Calculate the shortest inter-cluster flight path of the UAV using a double-layer shortest path algorithm.

[0074] Step S30: Design the initial trajectory of the drones in each group and calculate the minimum acquisition time of the drones in each group.

[0075] Step S40: output the shortest cruising time of the drone and the optimal allocation decision of each resource based on the shortest inter-cluster flight path of the drone and the minimum acquisition time of the drone in each cluster.

[0076] The minimum acquisition time of drones within each cluster and the inter-cluster flight time are added together to obtain the minimum cruising time of the drones as a whole.

[0077] In step S20, the AP clustering algorithm determines the optimal clustering decision for IoT devices. This clustering decision also reflects the scheduling information between the drone and IoT devices. Only when a drone begins accessing a cluster can the IoT devices in that cluster communicate with the drone.

[0078] In step S30, the optimal flight trajectory of the UAV between clusters and the optimal access order to each cluster can be obtained through the double-layer shortest path algorithm.

[0079] In step S40, the optimal allocation of drone communication resources is obtained by solving the minimum data collection problem for each cluster of drones, including the scheduling resources of drones and IoT devices within the cluster, the communication bandwidth resource allocation between drones and IoT devices, and the trajectory of drones within the cluster.

[0080] The above steps S20 and S30 can be executed in parallel, which can greatly reduce the calculation time.

[0081] Specifically, step S10 includes: This embodiment of the present invention uses an AP algorithm to cluster IoT devices on the ground, thereby determining the most appropriate clustering strategy. Because the location information of IoT devices is known, the AP algorithm does not require the IoT devices to actually communicate; the clustering process can be completed in the data center.

[0082] The flowchart of an AP algorithm based on dynamic programming provided by an embodiment of the present invention is as follows: Figure 3 As shown in Figure 2, the process includes the following: The AP algorithm simultaneously considers all points as potential central points. By recursively exchanging real-valued "information" between these points, all points can elect a set of reasonable central points.

[0083] First, three types of information are defined respectively. In the present invention, the IoT device k is regarded as a potential center point, and m is a general IoT device.

[0084] 1). Similarity information: Similarity information can be obtained by constructing a negative Euclidean distance function.

[0085]

[0086] 2) Responsibility information: This information is sent from IoT devices to the central point, which reflects the accumulated evidence of the central point's suitability as a hub for IoT devices, taking into account other potential central points for IoT devices.

[0087] r(m,k)=s(m,k)-{a(m,k′)+s(m,k)}.

[0088] 3) Availability information: This information is sent from the candidate central point to the IoT device. It indicates the suitability of the IoT device for selecting the central point as its central point. Of course, it also needs to consider the support for the central point from other IoT devices.

[0089]

[0090] Next, the three types of information mentioned above need to be updated. In order to prevent the messages in the AP algorithm from easily falling into repeated oscillations during the update process, a damping multiplier ζ can be added between two consecutive iterative updates. The modified update formula is as follows:

[0091] r′(m,k)=ζr(m,k)+(1-ζ)r′(m,k),

[0092] a′(m,k)=ζa(m,k)+(1-ζ)a′(m,k).

[0093] The above election process can be implemented using the DP (dynamic programming) algorithm. The specific process includes:

[0094] Get the coordinates of each IoT device m, randomly select one or more IoT devices k as the center point, which is the center point of a family, set the damping factor and the maximum number of iterations, calculate the similarity information, availability and credibility between each pair of IoT devices, and update them by multiplying by the damping factor (the function of the damping factor is to prevent the information from oscillating repeatedly and failing to converge to a value during the update), and obtain a similarity matrix S. Based on the above similarity information, availability information, credibility and damping factor, calculate the iterative convergence of the AP algorithm, increase the number of iterations by one, and judge whether it has reached iterative convergence. The convergence criterion is: the number of iterations reaches the maximum value, or the reliability and credibility information matrix no longer changes significantly, that is, the difference between two consecutive changes is less than the threshold value; if so, several IoT devices whose sum of the reliability and credibility information matrix is ​​greater than 0 are used as cluster heads, and the IoT devices with the maximum availability calculated by a certain cluster head are divided into the cluster of the cluster head, and the clustering results and the center points of each cluster are output; otherwise, the center point is updated, the similarity information, availability information and credibility between the IoT devices and the center point are recalculated, and the convergence of the AP algorithm iteration is recalculated until convergence is reached or the maximum number of iterations is reached.

[0095] Specifically, the above-mentioned step S20 includes: after obtaining the optimal clustering strategy of the ground IoT device, the present invention divides the entire drone cruising process into two stages, the first stage is the optimization mechanism of the flight trajectory between drone clusters, and the other stage is the resource optimization mechanism of the drone within the cluster.

[0096] Figure 4 A flowchart of a two-layer shortest path algorithm provided by an embodiment of the present invention. For the inter-cluster flight phase, a traveling salesman problem with neighborhood (TSPN) algorithm is proposed to calculate the shortest inter-cluster flight path for drones, considering minimizing the total distance between clusters.

[0097] This paper uses a two-layer shortest path algorithm based on TSPN to optimize the flight trajectories of inter-cluster drones, thereby minimizing inter-cluster flight time. The first layer of the algorithm uses a simulated annealing (SA) algorithm to solve the query order between clusters, similar to the traveling salesman problem. The second layer uses a dynamic programming algorithm based on the results of the first layer to design the flight trajectories of drones between clusters, thereby optimizing the inter-cluster flight time.

[0098] The first layer uses the simulated annealing algorithm to solve the traveling salesman problem (TSP). This is a probability-based algorithm that starts at a high initial temperature and, as the temperature parameter decreases, randomly searches for the global optimal solution of the objective function in the solution space, combining probabilistic properties. This means that the algorithm can probabilistically escape from the local optimal solution and ultimately converge to the global optimal solution.

[0099] Since the speed of the drone is fixed, minimizing time is equal to minimizing distance. Therefore, in the embodiment of the present invention, the distance that the drone takes off from the starting point, flies through the centers of each cluster, and returns to the starting point is set as energy. Then, sa The energy expression of the iteration is as follows, where L sa is the total number of iterations:

[0100]

[0101] in, The access order is No. o in k The energy update between two iterations satisfies the Metropolis criterion, which is to accept the new state with probability. In other words, when the energy E obtained in the l+1th iteration is 1+1 Not greater than the energy E obtained in the first iteration 1 Then accept the energy E obtained in the l+1th iteration l+1 Otherwise, the probability Choose whether to accept the energy E obtained in the l+1th iteration l+1 The probability update formula is as follows:

[0102]

[0103] After passing L sa The energy difference of the iteration or multiple times is not greater than the threshold value ∈ sa Then the algorithm is terminated and the access order that minimizes the inter-cluster flight distance is obtained.

[0104] The second layer: TSPN solution based on a dynamic programming algorithm. To further reduce the flight time between clusters, this embodiment of the present invention describes the area of ​​each IoT device cluster as a rough ellipse. This allows the aforementioned TSP to be converted into a TSPN. Compared to the TSP, the TSPN results in a shorter path.

[0105] The embodiment of the present invention adopts a simple method to describe the approximate ellipse of each cluster area. For the kth cluster, the embodiment of the present invention uses To represent the center of the ellipse, and The calculation formula is as follows:

[0106]

[0107]

[0108] in and are the maximum and minimum horizontal and vertical coordinate values ​​in the kth cluster respectively. The horizontal and vertical semi-axis lengths of the ellipse are respectively denoted as L k and W k , the calculation formula is as follows:

[0109]

[0110]

[0111] At this time, the point inside the kth ellipse can be expressed as follows

[0112]

[0113] For the sake of simplicity in calculation, the embodiment of the present invention transforms the ellipse into a polar coordinate system. In the polar coordinate system, a point on the boundary of the ellipse can be represented by the angle θ between the straight line between the point and the cluster center and the horizontal coordinate. k ∈[0,2π) to represent

[0114]

[0115] At this time, the TSPN optimization problem is to optimize the access location of the UAV to minimize the flight distance between the UAVs in a given access order, that is, to solve the following problem:

[0116]

[0117] In order to solve this optimization problem, the embodiment of the present invention adopts a dynamic programming algorithm to solve it. It is defined as the drone visiting from the starting point to the oth k The boundary points of the cluster The minimum distance, then the update formula of this minimum distance is as follows:

[0118]

[0119] The algorithm can be used to obtain the shortest distance between drones flying in clusters. and the optimal boundary point e * , so that the minimum time for the drone to fly between clusters can be calculated

[0120] Specifically, step S30 includes the following: For the intra-cluster collection phase, this embodiment of the present invention proposes a collection time minimization problem, which is a mixed-integer non-convex optimization problem. To break the coupling relationship between variables, this embodiment of the present invention uses altering optimization (AO) to decompose the original optimization problem, and then solves the remaining non-convex sub-problems based on the successive convex approximation (SCA) method. Furthermore, because clusters are independent of each other, the collection time of each cluster can be calculated in parallel.

[0121] Since a clustering design is adopted for IoT devices, the entire flight of the drone is divided into an inter-cluster flight phase and an intra-cluster collection phase, and these two parts are independent of each other. Since the duration of intra-cluster perception is a variable that needs to be considered, and this variable exists in both the objective and the constraint of the optimization problem, this makes the problem difficult to solve. Therefore, the embodiment of the present invention uses a one-dimensional search method to search for the perception duration variable and verifies the feasibility of the problem under different duration conditions. At the same time, since the communication range between the drone and the IoT device is limited, and the range of the different IoT device clusters divided by the AP algorithm varies greatly, if the initial trajectory of the drone is not designed properly, it is very likely that some IoT devices will never be able to communicate with the drone.

[0122] Figure 5 The embodiment of the present invention provides a drone perception coverage map of an elliptical trajectory, such as Figure 5 As shown in the figure, when a drone departs from the cluster entry point (asterisk) and flies counterclockwise along an elliptical initial trajectory, due to communication range limitations (grey large circles), some IoT device data can be collected (solid small circles) while some cannot (open small circles). Therefore, the initial trajectory design of the drone also needs to be considered.

[0123] Figure 6 This is a flow chart of an algorithm for initial trajectory of a UAV in a cluster provided by an embodiment of the present invention. The present invention proposes to use a TSP-based algorithm to design the initial trajectory of the UAV in the cluster, and at the same time set the flight speed of the UAV to the minimum speed Vmin , and the drone uses 1 / N of the space above each IoT device. k The purpose of this is to calculate the maximum acquisition time of the UAV in the cluster, which is the maximum acquisition time T of the subsequent UAV in the cluster. k A one-dimensional search provides an initial range.

[0124] For the kth cluster, the drone starts from the boundary point obtained in the previous step, visits each IoT device in turn, and finally returns to the boundary point. The entire journey distance is recorded as The flight time of the UAV in the cluster is recorded as:

[0125]

[0126] At the same time, the hovering communication time of each IoT device of the drone is:

[0127]

[0128] Therefore, the flight time of the UAV in the kth cluster is T k The maximum estimated value of is:

[0129]

[0130] Figure 7 This is a flowchart of a double-layer loop minimum drone acquisition time algorithm provided by an embodiment of the present invention. The resource optimization mechanism for in-cluster drone sensing of IoT device data provided by an embodiment of the present invention includes: given a given drone in-cluster acquisition time, the embodiment of the present invention analyzes the feasibility of relevant resources within the cluster, namely, the associated variable B between drones and IoT devices, the drone trajectory Q, and the bandwidth allocation ratio Λ. Due to the coupling relationship between the variables, the present invention uses an alternating optimization (AO) method to split the original feasibility problem into three sub-problems to verify feasibility separately:

[0131] Problem 1) Fix the drone trajectory Q and bandwidth allocation ratio Λ and verify the feasibility of the association variable B between the drone and the IoT device;

[0132] Problem 2) Fix the association variable B between the drone and the IoT device and the bandwidth allocation ratio Λ to verify the feasibility of the drone trajectory Q;

[0133] Problem 3) With fixed drone trajectories Q and the variables B linking drones and IoT devices, verify the feasibility of the bandwidth allocation ratio Λ. A two-layer loop algorithm for minimizing drone acquisition time is proposed based on a one-dimensional search method. Because each cluster is independent, the acquisition time optimization for all clusters can be calculated in parallel.

[0134] Finally, the minimum collection time of each cluster obtained by parallel calculation is added together to get the total collection time of the UAV in the cluster collection phase. Then, the total cruising time of the UAV is obtained by adding the collection time in the cluster and the flight time between clusters.

[0135] In summary, the embodiments of the present invention are aimed at scenarios where massive IoT device data is collected in areas that are not covered by cellular networks. By introducing drone communications and proposing a complete set of design solutions, the present invention avoids the high costs associated with traditional base station deployment.

[0136] Aiming at the characteristics of dispersed distribution and uneven data volume of IoT devices, the present invention uses the AP algorithm to cluster all IoT devices, and divides the overall process of the drone into two parts based on the clustering results, providing a basis for the subsequent optimization mechanism design.

[0137] In order to optimize the flight distance of drones between IoT device clusters, this paper innovatively designs a two-layer shortest path algorithm based on TSPN to provide the optimal flight path solution for drones flying between clusters.

[0138] To address the varying amounts of data contained by different IoT devices, this paper designs a two-layer, recurring intra-cluster minimum collection time algorithm. This algorithm minimizes intra-cluster collection time by jointly optimizing the connection strategy between drones and IoT devices, drone trajectories, and bandwidth allocation strategies. This enables drones to collect IoT device data in areas beyond cellular network coverage.

[0139] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0140] From the above description of the embodiments, it can be seen that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus the necessary general-purpose hardware platform. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium such as ROM / RAM, a magnetic disk, or an optical disk, 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 various embodiments of the present invention or certain parts of the embodiments.

[0141] Each embodiment in this specification is described in a progressive manner. The same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the relevant parts, refer to the partial description of the method embodiments. 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 can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. A person of ordinary skill in the art can understand and implement it without making any creative efforts.

[0142] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A resource allocation method for collecting IoT device data using drones, characterized in that: include: Use the proximity propagation AP algorithm to cluster the IoT devices on the ground; The minimum flight time of UAVs between clusters is calculated using a two-layer shortest path algorithm; Design the initial trajectory of the UAVs in each cluster and calculate the minimum acquisition time of the UAVs in each cluster; The shortest cruising time of the drone is obtained based on the minimum flying time between clusters and the minimum collection time of the drone within each cluster, and the optimal allocation decision of each resource is output; The method of calculating the minimum flight time of a UAV between clusters using the double-layer shortest path algorithm includes: The simulated annealing algorithm is used to solve the traveling salesman problem TSP. The distance that the drone takes off from the starting point, flies through the centers of each cluster and returns to the starting point is set as energy. Then the l∈L sa The energy expression of the iteration is as follows, where L sa is the total number of iterations: in, The access order is No. o in k The energy update between two iterations satisfies the Metropolis criterion, that is, the new state is accepted with probability. When the energy E obtained in the l+1th iteration is l +1 Not greater than the energy E obtained in the first iteration l , then accept the energy E obtained in the l+1th iteration l+1 Otherwise, the probability Choose whether to accept the energy E obtained in the l+1th iteration l+1 , the probability update formula is as follows: After passing L sa The energy difference of the iteration or multiple times is not greater than the threshold value ∈ sa Then the algorithm is terminated and the access order that minimizes the inter-cluster flight distance is obtained. The area of ​​each IoT device cluster is described by an ellipse. For the kth cluster, To represent the center of the ellipse, and The calculation formula is as follows: in and are the maximum and minimum horizontal and vertical coordinate values ​​in the kth cluster respectively, and the horizontal and vertical semi-axis lengths of the ellipse are recorded as L k and W k , the calculation formula is as follows: The points inside the kth ellipse are expressed as follows: Convert the ellipse to the polar coordinate system. In the polar coordinate system, the angle θ between the line passing through the point on the ellipse boundary and the abscissa of the cluster center is k ∈[0,2π) to represent: The TSPN optimization problem is to optimize the access location of the UAV to minimize the flight distance between the UAVs under a given access order. The dynamic programming algorithm is used to solve the following problem: Will It is defined as the drone visiting from the starting point to the oth k The boundary points of the cluster The minimum distance, then the update formula of this minimum distance is as follows: The algorithm is used to obtain the shortest distance between UAVs flying in clusters. and the optimal boundary point e * , thereby calculating the minimum time for the drone to fly between clusters 2. The method according to claim 1, characterized in that The AP algorithm is used to cluster the IoT devices on the ground, including: Obtain the coordinates of each IoT device m, randomly select one or more IoT devices k as the center point, which is the center point of a cluster, set the damping factor and the maximum number of iterations, calculate the similarity information, availability and credibility between each pair of IoT devices, and update it by multiplying it by the damping factor to obtain a similarity matrix S. Based on the similarity information, availability information, credibility and damping factor, calculate the iterative convergence of the AP algorithm. Let the number of iterations increase by one to determine whether iterative convergence is achieved. The criterion for iterative convergence is: the number of iterations reaches the maximum value, or it is available. The reliability and credibility information matrix no longer changes significantly, that is, the difference between two consecutive changes is less than the threshold value; if so, several IoT devices whose sum of the reliability and credibility information matrix is ​​greater than 0 are used as cluster heads, and the IoT devices with the maximum availability calculated by a certain cluster head are divided into the cluster of the cluster head, and the clustering results and the center points of each cluster are output; otherwise, the center point is updated, the similarity information, availability information and credibility between the IoT devices and the center point are recalculated, and the iterative convergence of the AP algorithm is recalculated until convergence is reached or the maximum number of iterations is reached.

3. The method according to claim 1, characterized in that The design of the initial trajectory of the UAVs in each cluster and the calculation of the minimum acquisition time of the UAVs in each cluster include: The TSP-based algorithm is used to design the initial trajectory of the UAV within the cluster, and the flight speed of the UAV is set to the minimum speed V min , and the drone uses 1 / N of the space above each IoT device. k For the kth cluster, the drone starts from the boundary point obtained in the previous step, visits each IoT device in turn, and finally returns to the boundary point. The entire distance is recorded as The flight time of the UAV in the cluster is recorded as: At the same time, the hovering communication time of each IoT device of the drone is: Therefore, the flight time of the UAV in the kth cluster is T k The maximum estimated value of is: Given the collection time within a drone cluster, the feasibility of the relevant resources within the cluster, namely the association variable B between drones and IoT devices, drone trajectory Q, and bandwidth allocation ratio Λ, is analyzed. The alternating optimization method is used to split the feasibility problem into three sub-problems to verify the feasibility separately: Problem 1) Fix the drone trajectory Q and bandwidth allocation ratio Λ and verify the feasibility of the association variable B between the drone and the IoT device; Problem 2) Fix the association variable B between the drone and the IoT device and the bandwidth allocation ratio Λ to verify the feasibility of the drone trajectory Q; Problem 3) Fix the drone trajectory Q and the association variable B between the drone and the IoT device, verify the feasibility of the bandwidth allocation ratio Λ, and propose a double-layer loop minimum drone acquisition time algorithm based on a one-dimensional search method. The acquisition time of all clusters is optimized and calculated in parallel; The minimum collection time of each cluster obtained by parallel calculation is added together to obtain the total collection time of the UAV during the intra-cluster collection phase.

4. The method according to any one of claims 1 to 3, characterized in that The method of obtaining the shortest cruising time of the drone based on the minimum inter-cluster flight time and the minimum collection time of the drone within each cluster and outputting the optimal allocation decision of each resource includes: The minimum acquisition time of drones within each cluster and the inter-cluster flight time are added together to obtain the shortest cruising time of the drones as a whole, and the optimal allocation scheme for drone communication resources is obtained. This optimal allocation scheme includes the scheduling resources of drones and IoT devices within the cluster, the communication bandwidth resource allocation between drones and IoT devices, and the trajectory of drones within the cluster.

Images