A UAV-assisted data collection system and method

Abstract

The application discloses an unmanned aerial vehicle (UAV) assisted data collection system and method, and belongs to the field of wireless communication. The system comprises: a plurality of sensor nodes, which are uniformly organized into a plurality of clusters; in each cluster, a plurality of sensor nodes are paired into a group and share a frequency resource block; and a UAV, which is used for collecting data of the ground sensor nodes, adopting a serial interference cancellation technology to eliminate multiple access interference in each group to decode signals of the sensor nodes; wherein, by considering optimization of pairing-decoding of the sensor nodes, transmission power and trajectory points of the UAV, minimum throughput of the sensor nodes is maximized. The application maximizes the minimum throughput of the sensor nodes by jointly optimizing the pairing scheme, decoding order, transmission power of the sensor nodes and the trajectory points of the UAV, and provides a flexible, efficient and reliable data collection solution for a wireless sensor network.

Description

Technical Field

[0001] This invention relates to the field of wireless communication, and more particularly to an unmanned aerial vehicle (UAV)-assisted data collection system and method. Background Technology

[0002] Unmanned aerial vehicle (UAV)-assisted data collection in wireless sensor networks (SAVs), i.e., dispatching UAVs to collect data from ground sensor nodes, can overcome the challenges of spatial limitations, energy constraints, and data timeliness in traditional SAV data collection. However, limited spectrum resources restrict the number of sensor nodes that UAVs can serve, thus limiting data collection efficiency and making it impossible to meet the large-scale connectivity requirements of SAVs. As a novel multiple access technology, non-orthogonal multiple access (NOAMI) shows promise as a solution to this problem, improving the efficiency of UAV data collection and serving large-scale SAVs when communication system spectrum resources are limited. Unlike traditional orthogonal multiple access (NOAMI) technologies that allocate each orthogonal time-frequency resource block to a single sensor node, NOAMI allows multiple sensor nodes to share a single time-frequency resource block and distinguishes the signals of different sensor nodes in the power domain or code domain. Research shows that communication systems based on NOAMI can achieve more efficient utilization of time-frequency resources through heterogeneous channel conditions, improving spectrum efficiency, reducing latency, and supporting large-scale connectivity.

[0003] Existing technical solutions have been studied for both orthogonal and non-orthogonal multiple access (MOA) scenarios, optimizing UAV trajectories, ground sensor node transmission power, and scheduling schemes to maximize minimum throughput. Numerical results show that the minimum throughput achieved in non-orthogonal MOA scenarios is no less than that in orthogonal MOA scenarios. However, in non-orthogonal MOA scenarios, if a large number of sensor nodes access the same time-frequency resource block, the receiver will suffer severe interference when decoding various signals, resulting in higher decoding latency. Furthermore, when the strength differences between received signals are small, signal decoding is prone to errors, leading to continuous error propagation. In other words, multiple sensor nodes sharing the same time-frequency resource block increases the design complexity of the signal receiver and imposes stricter accuracy requirements. Summary of the Invention

[0004] In order to at least partially solve one of the technical problems existing in the prior art, the purpose of this invention is to provide a UAV-assisted data collection system and method based on wireless sensor node pairing access.

[0005] The technical solution adopted in this invention is:

[0006] A drone-assisted data collection system includes:

[0007] Multiple sensor nodes are evenly organized into M clusters; within each cluster, several sensor nodes are paired into a group and share a frequency resource block, with different groups using different frequency resource blocks;

[0008] Unmanned aerial vehicles (UAVs) are used to collect data from ground sensor nodes and, after receiving the data, employ serial interference cancellation technology to eliminate multiple access interference in each group in order to decode the signals of each sensor node.

[0009] This involves optimizing sensor node pairing-decoding, transmission power, and UAV trajectory points to maximize the minimum throughput of sensor nodes and improve fairness among them.

[0010] Another technical solution adopted in this invention is:

[0011] A design method for the above-described UAV-assisted data collection system includes the following steps:

[0012] Construct a system model, and determine the mathematical programming problem based on the constructed system model;

[0013] The mathematical programming problem is broken down into two subproblems: the matching and decoding order subproblem, and the transmission power and UAV trajectory point problem;

[0014] Design a joint optimization algorithm based on block coordinate descent to solve the pairing and decoding order subproblem and the transmit power and UAV trajectory point problem;

[0015] Design heuristic algorithms to optimize the solutions to the pairing and decoding order subproblems and the problems of transmission power and UAV trajectory points;

[0016] The initial feasible point of the UAV is designed based on the joint optimization algorithm and the heuristic algorithm.

[0017] Furthermore, two sensor nodes are paired to form a group, which is called a NOMA group;

[0018] make Let represent the set of indices for all clusters in a wireless sensor network. Let represent the set of indices of all sensor nodes in the m-th cluster, and let This represents the k-th sensor node in the m-th cluster. Represents sensor nodes The location, among which,

[0019] set up Represents sensor nodes Whether or not Pairing and the decoding order during pairing, i.e., when sensor nodes and Paired into a NOMA group and signal strength When the signal is decoded later, otherwise, in,

[0020] set up For sensor nodes The transmission power, Let m be the trajectory point of the UAV in the m-th cluster;

[0021] Among them, variables describing the pairing and decoding order of sensor nodes will be included. Simply put, it's called pairing-decoding. In each NOMA group, the sensor node whose signal is decoded first is referred to as the first-decoding node, and the other sensor node is referred to as the second-decoding node.

[0022] Furthermore, the construction of the system model, and the determination of the mathematical programming problem based on the constructed system model, includes:

[0023] The communication link between the UAV and the ground sensor node is a line-of-sight propagation link, and the sensor node... Channel power gain between drones Represented as:

[0024]

[0025] In the formula, ρ0 is the channel power gain, and H is the flight altitude of the UAV;

[0026] sensor nodes The corresponding signal-to-interference-plus-noise ratio is expressed as:

[0027]

[0028] in, This is the ratio of channel power gain to noise power, i.e.:

[0029]

[0030] N0 is the power spectral density of additive white Gaussian noise, and B is the channel bandwidth.

[0031] Consider optimizing sensor node pairing and decoding. Transmit power and the drone's trajectory points To maximize the minimum throughput of the sensor nodes, the problem is modeled as follows, denoted as P1:

[0032]

[0033] Where η represents the minimum throughput in bits per second (bps), and β≥1 is a predetermined power difference ratio (used to ensure that the receiver can distinguish between two received signals and apply SIC technology for decoding). For sensor nodes Maximum transmit power, This is an index set of drone trajectory points. As the starting point for drones, Let L be the endpoint of the UAV and L be the maximum allowable trajectory length; constraint (5) ensures that η is the minimum throughput; constraint (6) specifies the pairing-decoding. The range of values; constraint (7) ensures that each sensor node participates in pairing and is paired with only one node. At the same time, constraint (7) avoids "when hour, and The signals are all considered as interference; constraint (8) ensures that the receiver on the UAV prioritizes decoding the high-power signals (in order to maximize the minimum throughput); constraint (9) ensures that the transmit power of each sensor node does not exceed its maximum transmit power; constraint (10) limits the length of the UAV's flight path to ensure that the UAV can collect data before its onboard energy is exhausted.

[0034] Furthermore, the design is based on a joint optimization algorithm using block coordinate descent to solve the pairing and decoding order subproblem and the transmit power and UAV trajectory point problem, including:

[0035] In each iteration, firstly, with the transmit power P and trajectory point Q fixed, the pairing and decoding order subproblem is solved to optimize the pairing-decoding A; then, with the pairing-decoding A fixed, the transmit power and UAV trajectory point problem is solved to optimize the transmit power P and trajectory point Q.

[0036] Furthermore, the pairing and decoding order subproblem includes:

[0037] Given the transmit power P and the UAV trajectory point Q, problem P1 simplifies to the following problem, denoted as P2:

[0038]

[0039] st(5)-(8)

[0040] To further transform problem P2 into a more manageable form, we first introduce auxiliary variables. And it stipulates:

[0041]

[0042] at this time,

[0043]

[0044] Therefore, problem P2 is transformed into the following, denoted as P3:

[0045]

[0046] The optimal solution to problem P3 can only be obtained when constraint (12) holds true on the equality side. Therefore, problem P3 is equivalent to problem P2.

[0047] Furthermore, since both constraints (6) and (15) are non-convex constraints, problem P3 is a non-convex programming problem. Therefore, a penalty-based successive convex approximation algorithm is used to solve problem P3 in order to obtain an approximate optimal solution to problem P3.

[0048] Furthermore, the issue of the transmission power and the drone trajectory points includes:

[0049] Given the pairing-decoding problem A, problem P1 becomes the following programming problem, denoted as P6:

[0050]

[0051] st(5),(8)–(10)

[0052] The SCA algorithm is used to solve the problem of transmit power and UAV trajectory point; in particular, after obtaining the optimal trajectory point Q, the optimal transmit power P is then obtained.

[0053] Furthermore, the design heuristic algorithm includes:

[0054] In each iteration of the solution, the sensor node with the minimum throughput is first found. Paired sensor nodes Cluster

[0055] Target nodes are determined according to preset criteria. And attempt to make the first The trajectory points corresponding to the clusters are directed towards the target node. Move in the direction it is in;

[0056] Update sensor node pairing, decoding, and transmit power to improve minimum throughput.

[0057] Furthermore, the design of the initial feasible points of the UAV based on the joint optimization algorithm and heuristic algorithm includes:

[0058] Use the centroid of the sensor node in each cluster as the initial trajectory point; if constraint (10) is satisfied, directly use it as the feasible trajectory point and skip the subsequent steps; otherwise, proceed to the next step.

[0059] Calculate the centroid of all trajectory points, and move each trajectory point toward the centroid position to shorten the trajectory length, while ensuring that the minimum throughput does not decrease and the trajectory length does not increase. To avoid getting stuck in a loop, when a trajectory point can no longer be moved, the trajectory point will be marked. When all trajectory points are marked, this step ends.

[0060] If the trajectory points obtained in the previous step satisfy constraint (10), then exit this step; otherwise, move the trajectory point farthest from the drone's starting point q0 towards the starting point q0 by a fixed step length s. c This causes the center of gravity of the trajectory point to move towards the starting point q0, and then returns to the previous step.

[0061] The beneficial effects of this invention are as follows: This invention proposes a UAV-assisted data collection scheme based on paired access, allowing sensor nodes to transmit data in pairs using Non-Orthogonal Multiple Access (NOMA) technology. By jointly optimizing the pairing scheme, decoding order, transmission power, and UAV trajectory points of the sensor nodes, the minimum throughput of the sensor nodes is maximized, providing a flexible, efficient, and reliable data collection solution for wireless sensor networks. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following description is provided with accompanying drawings of the relevant technical solutions in the embodiments of the present invention or the prior art. It should be understood that the accompanying drawings described below are only for the purpose of clearly illustrating some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 This is a schematic diagram of an unmanned aerial vehicle (UAV) assisted data collection system according to an embodiment of the present invention;

[0064] Figure 2 This is a design flowchart of the UAV-assisted data collection system in an embodiment of the present invention;

[0065] Figure 3 This is a flowchart of the joint optimization algorithm in an embodiment of the present invention. Detailed Implementation

[0066] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0067] In the description of this invention, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., are based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0068] In the description of this invention, "several" means one or more, "multiple" means two or more, "greater than," "less than," and "exceeding" are understood to exclude the stated number, while "above," "below," and "within" are understood to include the stated number. If "first" or "second" is used, it is only for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the order of the indicated technical features. Furthermore, "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0069] In the description of this invention, unless otherwise explicitly defined, terms such as "set up," "install," and "connect" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this invention in conjunction with the specific content of the technical solution.

[0070] Terminology Explanation:

[0071] Non-Orthogonal Multiple Access (NOMA): NOMA is a novel multi-user communication technology that allows multiple users to transmit signals on the same time and frequency while differentiating users in the power domain. Based on appropriate power control and resource allocation strategies, NOMA can improve the frequency efficiency and system capacity of wireless communication systems.

[0072] Successive Interference Cancellation (SIC): SIC is a technique used to eliminate interference in multi-user communication systems. It eliminates interference step by step according to the power of different user signals in order to achieve the separation and decoding of user signals.

[0073] To address the receiver design complexity caused by multiple sensor nodes accessing the same time-frequency resource block, this invention proposes a pairing-based non-orthogonal multiple access scheme. Sensor nodes are paired in pairs, thus reducing system complexity while maintaining system performance. To improve data collection efficiency and ensure fairness among sensor nodes, this invention maximizes the minimum system throughput by jointly optimizing the pairing scheme, decoding order, transmission power, and UAV trajectory points. This provides an efficient and reliable data collection solution for wireless sensor networks, further promoting the development of wireless sensor network technology.

[0074] like Figure 1 This embodiment provides a drone-assisted data collection system, including:

[0075] Multiple sensor nodes are evenly organized into M clusters; within each cluster, several sensor nodes are paired into a group and share a frequency resource block, with different groups using different frequency resource blocks;

[0076] Unmanned aerial vehicles (UAVs) are used to collect data from ground sensor nodes and, after receiving the data, employ serial interference cancellation technology to eliminate multiple access interference in each group in order to decode the signals of each sensor node.

[0077] This involves optimizing sensor node pairing-decoding, transmission power, and UAV trajectory points to maximize the minimum throughput of sensor nodes and improve fairness among them.

[0078] Regarding the system described above, this embodiment also provides a design method, including the following steps:

[0079] S1. Construct a system model and determine the mathematical programming problem based on the constructed system model;

[0080] S2. Break down the mathematical programming problem into two subproblems: the matching and decoding order subproblem, and the transmission power and UAV trajectory point problem;

[0081] S3. Design a joint optimization algorithm based on block coordinate descent to solve the pairing and decoding order subproblem and the transmit power and UAV trajectory point problem;

[0082] S4. Design heuristic algorithms to optimize the solutions to the pairing and decoding order subproblems and the transmission power and UAV trajectory point problems;

[0083] S5. Design the initial feasible points of the UAV based on the joint optimization algorithm and the heuristic algorithm.

[0084] The following is in conjunction with the appendix Figure 2 The above will be explained in detail with specific embodiments.

[0085] I. Constructing a system model and formulating a mathematical programming problem:

[0086] See Figure 1 The embodiments of the present invention consider the following wireless sensor network data collection scenario: a rotary-wing UAV collects data from a wireless sensor network containing N ground sensor nodes. Specifically, these sensor nodes are uniformly organized into M clusters, that is, each cluster contains There are 10 sensor nodes, among which... This indicates rounding down. Assume the UAV collects data from ground sensor nodes according to a flight-hover-communication protocol. That is, the UAV hovers at each trajectory point for a period of time, collecting data from all sensor nodes in the corresponding cluster, and then flies to the next trajectory point. To achieve high spectral efficiency and reduce system complexity, this invention specifies that sensor nodes in each cluster use a non-orthogonal multiple access (NOMA) scheme to transmit data. Specifically, in each cluster of the wireless sensor network, every two sensor nodes are paired into a group and share a frequency resource block (this embodiment assumes that each cluster contains an even number of sensor nodes, i.e., K is even; in fact, a resource block can be allocated to a single sensor node, thus extending this embodiment to scenarios where K is odd). Different groups use different frequency resource blocks. For ease of explanation, this embodiment refers to the group formed by pairing two sensor nodes as a NOMA group. After receiving data from the sensor nodes, the UAV uses Successive Interference Cancellation (SIC) technology to eliminate multiple access interference in each NOMA group and decodes the signals of each sensor node.

[0087] Without loss of generality, let Let represent the set of indices for all clusters in a wireless sensor network. Let represent the set of indices of all sensor nodes in the m-th cluster, and let This represents the k-th sensor node in the m-th cluster. Represents sensor nodes The location, among which,

[0088] set up Represents sensor nodes Whether or not Pairing and the decoding order during pairing, i.e., when sensor nodes and Paired into a NOMA group and signal strength When the signal is decoded later, otherwise, in, set up For sensor nodes The transmission power, Let be the trajectory point of the UAV in the m-th cluster. For ease of explanation, this embodiment will describe the variables of sensor node pairing and decoding order. Simply put, it's called pairing-decoding. In each NOMA group, the sensor node whose signal is decoded first is referred to as the "first-decoding node," and the other sensor node is referred to as the "later-decoding node." Furthermore, in the following text, unless otherwise specified, all subscripts k, l, and m are arbitrary, i.e.,

[0089] According to 3GPP specifications, when a drone flies at an altitude of 40m or higher in rural areas, or 100m or higher in urban areas, the communication link between the drone and the ground sensor node can almost entirely be considered a line-of-sight (LAS) link. Therefore, the sensor node... Channel power gain between drones It can be represented as:

[0090]

[0091] Where ρ0>0 represents the channel power gain at a reference distance of 1m, and H represents the UAV's flight altitude. When the receiver uses SIC technology to decode the signal of a sensor node, it will treat undecoded signals within the same NOMA group as interference. Therefore, sensor nodes can be... The corresponding signal-to-interference-plus-noise ratio (SINR) is expressed as follows:

[0092]

[0093] in, This is the ratio of channel power gain to noise power, i.e.:

[0094]

[0095] N0 is the power spectral density of additive white Gaussian noise, and B is the channel bandwidth. Specifically, this invention assumes a system bandwidth B s The channel bandwidth is divided equally among all NOMA groups within each cluster; that is, for each NOMA group, the channel bandwidth is...

[0096] To maximize the throughput of sensor nodes and ensure fairness among them, this embodiment considers optimizing the pairing and decoding of sensor nodes. Transmit power and the drone's trajectory points This maximizes the minimum throughput of the sensor nodes. Therefore, the problem can be modeled (denoted as P1) as follows:

[0097]

[0098] Where η represents the minimum throughput in bits per second (bps), and β≥1 is a predetermined power difference ratio (used to ensure that the receiver can distinguish between two received signals and apply SIC technology for decoding). For sensor nodes Maximum transmit power, This is an index set of drone trajectory points. As the starting point for drones, Let η be the endpoint of the UAV, and L be the maximum allowable trajectory length. Constraint (5) ensures that η is the minimum throughput; constraint (6) specifies the pairing-decoding. The range of values; constraint (7) ensures that each sensor node participates in pairing and is paired with only one node. At the same time, constraint (7) can also avoid "when hour, and The signals are all considered as interference; constraint (8) ensures that the receiver on the UAV prioritizes decoding the high-power signals (in order to maximize the minimum throughput); constraint (9) ensures that the transmit power of each sensor node does not exceed its maximum transmit power; constraint (10) limits the length of the UAV's flight path to ensure that the UAV can collect data before its onboard energy is exhausted.

[0099] II. Propose a joint optimization algorithm:

[0100] Given the transmit power P and the UAV trajectory point Q, problem P1 can be simplified into the following problem, denoted as P2:

[0101]

[0102] st(5)-(8).

[0103] To further transform problem P2 into a more manageable form, this invention first introduces auxiliary variables. And it stipulates:

[0104]

[0105] at this time,

[0106]

[0107] Therefore, problem P2 is transformed into the following, denoted as P3:

[0108]

[0109] The optimal solution to problem P3 can only be obtained when constraint (12) holds true on the equality side. Therefore, problem P3 is equivalent to problem P2.

[0110] Since constraints (6) and (15) are non-convex, problem P3 is a non-convex programming problem. Therefore, this embodiment proposes a penalty-based successive convex approximation (PSCA) algorithm to obtain an approximate optimal solution to problem P3.

[0111] First, constraint (6) can be rewritten in the following equivalent form:

[0112]

[0113] Furthermore, the constraint (17) can be used as a penalty term in the objective function using the penalty function method. That is, problem P3 can be restated as follows, denoted as P4:

[0114]

[0115] Here, ξ>0 is the penalty factor, and π≥0 is the auxiliary variable, which characterizes the violation value of constraint (17). According to the characteristics of the penalty function method, when ξ is large enough (i.e., ξ→∞), problem P4 is equivalent to problem P3.

[0116] Therefore, the non-convex constraints (15) and (19) can be transformed into convex constraints using the Successive Convex Approximation (SCA) technique. Specifically, the convex function on the left side of constraint (15) is always greater than its first-order Taylor expansion at any feasible point, i.e.:

[0117]

[0118] in, At the feasible point The first-order Taylor expansion at that point is:

[0119]

[0120] Therefore, constraint (15) can be rewritten as:

[0121]

[0122] Similarly, for a given feasible point Constraint (19) can be rewritten as:

[0123]

[0124] Therefore, problem P4 can be restated as follows, denoted as P5:

[0125]

[0126] st(7)–(8),(12),(16),(22),(23).

[0127] It is easy to see that problem P5 is a convex programming problem, and therefore can be solved using standard convex optimization tools (e.g., CVX). Furthermore, by iteratively solving problem P5 and updating feasible points, the solution to problem P4 can be obtained.

[0128] Based on the above analysis, the PSCA algorithm can be summarized as follows: For each iteration, first at a given point... and Based on this, solve problem P5, and then use the solution to problem P5 to update the given point A. (i) and X (i) Specifically, the PSCA algorithm sets a very small initial value for the penalty factor ξ to provide sufficient degrees of freedom for A, and multiplies it by a constant in each iteration. Increase ξ until it reaches a predefined upper limit ξ. max It is worth noting that when ξ = ξ max At this point, the PSCA algorithm becomes the standard SCA algorithm. The specific steps of the PSCA algorithm are shown in Algorithm 1, where ∈1 and ∈2 are convergence criteria. It can be proven that Algorithm 1 can guarantee convergence to a solution that satisfies the Karush-Kuhn-Tucker (KKT) conditions of problem P4, that is, the KKT point of problem P4, which is also a stationary point of problem P2.

[0129]

[0130]

[0131] II-B. Solving the problem of transmit power versus UAV trajectory points:

[0132] Given the pairing-decoding A, problem P1 becomes the following planning problem, denoted as P6:

[0133]

[0134] st(5),(8)–(10).

[0135] For ease of description, the present invention defines As a pairing decoding set, it contains the pairing scheme and decoding order information for each NOMA group. Unless otherwise specified, the notations (m,k,l) ​​appearing in the remaining content of this embodiment are sets. Any element in, that is,

[0136] Since the logarithmic throughput is a monotonically increasing function of the signal-to-interference-plus-noise ratio (SINR), maximizing the minimum throughput can be equivalently transformed into maximizing the minimum SINR. That is, the optimal solution to problem P6 can be obtained by solving the following problem, denoted as P7:

[0137]

[0138] st(8)–(10).

[0139] in, and Sensor nodes and The signal-to-interference-to-noise ratio.

[0140] Proposition 1: The following optimization problem is denoted as...

[0141]

[0142] st 0≤x≤s, (28)

[0143] 0≤y≤t, (29)

[0144] cx≤y, (30)

[0145] Essentially equivalent to, denoted as

[0146]

[0147] Furthermore, the optimal solution for x and y is:

[0148]

[0149] y = t, (33)

[0150] Where c>0 is a constant.

[0151] Proof: First, observe the problem. It can be observed that y should be maximized to provide a larger feasible region for x, while simultaneously maximizing the objective function value. That is, y = t. Furthermore, the problem... It can be written as:

[0152]

[0153] 0≤x≤s, (35)

[0154]

[0155] According to the function The monotonicity can be proven as follows:

[0156]

[0157] That is, as x increases, the objective function First increase, then decrease. Therefore, to maximize the objective function value, the following conclusion holds: when and When the optimal x is Otherwise, the optimal x is That is, the optimal solution for x is:

[0158]

[0159] In particular,

[0160]

[0161] According to formulas (37) and (39), maximizing This is equivalent to maximizing x. Therefore, the problem... Equivalent to:

[0162]

[0163] The proof is complete.

[0164] Will Substituting c = β into proposition 1, we get: Problem P7 is equivalent to, denoted as P8:

[0165]

[0166] st(10),

[0167] Furthermore, the optimal transmit power P is:

[0168]

[0169] in, In fact, the objective function of problem P8 is to minimize the signal-to-interference-plus-noise ratio.

[0170] Furthermore, problem P8 can be rewritten in a more manageable form, denoted as P9:

[0171]

[0172] st(10),

[0173]

[0174] Where μ is an auxiliary variable, and, That is the minimum signal-to-interference-plus-noise ratio.

[0175] Note that constraints (45) and (46) can be equivalently transformed into:

[0176]

[0177] Furthermore, constraint (47) can be rewritten as:

[0178]

[0179] Right now,

[0180]

[0181] For the nonconvexity of constraint (51), the SCA technique can be applied to approximate it. Specifically, for a given μ (j) convex function Using its first-order Taylor expansion as the lower bound, that is:

[0182]

[0183] Therefore, constraint (51) can be approximated as:

[0184]

[0185] Therefore, problem P9 can be approximated as P10:

[0186]

[0187] st(10),(48),(49),(53).

[0188] Problem P10 is a convex programming problem, which can be solved directly using the convex optimization toolkit.

[0189] By applying the SCA technique, problem P10 can be solved iteratively, and point μ can be updated after each iteration.(j) Thus, an approximate optimal solution to problem P9 is obtained. It can be proven that the objective function value of problem P10 is monotonically non-increasing during the iteration process, and correspondingly, the objective function value of problem P9 is monotonically non-decreasing during the iteration process. Furthermore, the solution obtained by iteratively solving problem P10 can be guaranteed to converge to a KKT point of problem P9. After obtaining the solution Q of problem P9, the optimal transmission power P can be obtained through formulas (42) and (43). The complete algorithm for solving problem P6 can be found in Algorithm 2, where ∈3 is the convergence accuracy.

[0190]

[0191]

[0192] II-C. Design Iterative Algorithms:

[0193] See Figure 3 Based on the above analysis, this embodiment of the invention proposes a joint optimization algorithm based on block coordinate descent to obtain an approximate optimal solution to problem P1. Specifically, in each iteration, Algorithm 1 is first used to solve the pairing-decoding subproblem P2, and then Algorithm 2 is used to solve the transmit power and UAV trajectory point subproblem P6. The specific steps are shown in Algorithm 3, where ∈4 represents the convergence accuracy of the joint optimization algorithm.

[0194]

[0195] The convergence analysis of Algorithm 3 is as follows. Let η[A] (r) ,P (r) Q (r) Let ] be the objective function value of Algorithm 3 when the r-th iteration is completed. Then:

[0196] η[A (r) ,P (r) Q (r) ]≤η[A (r+61) ,P (r) Q (r) (55a)

[0197] ≤η[A (r61) ,P (r+1) Q (r+1) (55b)

[0198] Inequality (55a) holds because problem P2 is solved approximately optimally by algorithm 1, and inequality (55b) holds because problem P6 is solved approximately optimally by algorithm 2. Inequality (55) shows that the objective function value is monotonically non-decreasing during the iteration of algorithm 3. Since the objective function value of problem P1 always has an upper bound, algorithm 3 can guarantee convergence.

[0199] The worst-case time complexity of Algorithm 3 is... Where I1, I2, and I3 are the number of iterations of Algorithm 1, Algorithm 2, and Algorithm 3, respectively; n1 = NK + N + 2 is the number of variables in problem P5; n2 = 2M + 1 is the number of variables in problem P10; and ∈ > 0 is the accuracy of the interior point method in the optimization toolkit.

[0200] III. Propose a heuristic algorithm:

[0201] Because Algorithm 3 has relatively high computational complexity, it often cannot obtain a solution to problem (P1) in a short time. To address this, this invention proposes a heuristic algorithm with low complexity to efficiently obtain an approximate solution to problem (P1).

[0202] III-A. Design a drone trajectory point update scheme:

[0203] Given an initial feasible solution Based on this, the throughput of each sensor node can be calculated, and the minimum throughput η among all sensor nodes can be obtained. (n) For ease of explanation, let Let represent the sensor node with the lowest throughput (if there is more than one sensor node corresponding to the lowest throughput, any one of them can be selected), and let . Indicates and Paired sensor nodes, express and The cluster in which it is located. According to formula (41), if the transmission power P (n) The minimum throughput η is determined by formulas (42) and (43). (n) In fact, it was or The constraints are as follows. Therefore, while ensuring that the drone trajectory length constraint is met, the trajectory points can be... Towards or The location was moved to improve... or This increases the minimum throughput η (n) .

[0204] For the sake of simplicity, this invention will use trajectory points The sensor node that should be nearby is called the target node. Therefore, the target node can be determined according to the following criteria:

[0205] 1) If and but For the post-decoding node, and:

[0206]

[0207] Right now:

[0208]

[0209] Combining formula (41), we can see that This limits the minimum throughput. Therefore, Should be close To increase This further improves the minimum throughput, i.e., the target node is

[0210] 2) Otherwise,

[0211]

[0212] Right now, This limits the minimum throughput. Therefore, the target node is... Should The location is nearby.

[0213] III-B. Design of Pairing-Decoding and Transmit Power Update Scheme:

[0214] Since the pairing-decoding and transmission power of sensor nodes are closely related to the UAV's trajectory points, the pairing-decoding and transmission power of sensor nodes should also be updated after modifying the UAV's trajectory points. This invention aims to provide a heuristic criterion for determining the update scheme for the pairing-decoding and transmission power of sensor nodes.

[0215] According to serial interference cancellation techniques, for a given signal, the signal-to-interference-plus-noise ratio (SNR) obtained by decoding the signal first will be smaller than that obtained by decoding it later (because decoding earlier is subject to interference from other signals). In other words, the throughput corresponding to decoding earlier is less than that of decoding later. Therefore, sensor nodes with low transmit power and relatively poor channel conditions should be used as later decoding nodes to maximize their throughput.

[0216] For ease of description, in this embodiment of the invention, the product of the maximum transmit power of each sensor node and the channel power gain is referred to as the Maximum Received Power of Signal (MRPS). This MRPS is used to characterize the maximum signal-to-interference-plus-noise ratio (SNR) achievable by the sensor node signal and is used as an evaluation metric to determine the pairing and decoding order of sensor nodes. Specifically, the pairing-decoding scheme for sensor nodes in each cluster is as follows:

[0217] a) First, sort all sensor nodes in the cluster according to their MRPS values ​​from largest to smallest. Then, form a strong set with half of the sensor nodes having the largest MRPS values, and a weak set with the remaining half.

[0218] b) Further, pair the first node in the strong set with the first node in the weak set. Continue in this manner until all nodes have been paired.

[0219] c) To maximize minimum throughput, nodes in the strong set should be decoded first, and paired nodes in the weak set should be decoded last.

[0220] Therefore, the pairing-decoding schemes for UAV trajectory points and sensor nodes have been determined, and the transmission power of sensor nodes can be updated directly according to formulas (42) and (43).

[0221] III-C. Algorithm Description:

[0222] Based on the above discussion, the following heuristic algorithm can be proposed: In each iteration, first find the sensor node with the minimum throughput. Paired sensor nodes Cluster Furthermore, the target node is determined according to the criteria in section III-A. And attempt to make the first The trajectory points corresponding to the clusters are directed towards the target node. Move in the direction of movement; finally, update the pairing-decoding and transmit power of the sensor nodes according to section III-B to improve minimum throughput.

[0223] It is worth noting that each adjustment to the drone's trajectory points and pairing-decoding scheme may reduce the first... The throughput of other sensor nodes in the cluster may also cause constraint (10) to fail. Therefore, it is necessary to check the feasibility of each adjustment. In particular, if the adjusted minimum throughput is not less than the original minimum throughput and constraint (10) is met, the adjustment can be accepted (even if the minimum throughput does not change after adjusting the UAV trajectory points, the adjustment can still be accepted. This is because there may be multiple sensor nodes with the minimum throughput, in which case multiple adjustments are needed to improve the system's minimum throughput). Otherwise, the adjustment should be revoked. In addition, each adjustment may change the sensor node corresponding to the minimum throughput, so it is necessary to update after accepting the adjustment. as well as For simplicity, the algorithm proposed in this invention will search at the beginning of each iteration. and

[0224] The specific steps of the heuristic algorithm are shown in Algorithm 4, where ζ0 is the initial movement step size of the UAV trajectory point, ∈5 is the convergence accuracy, and m L This is the number of the trajectory point from the last update.

[0225]

[0226]

[0227] The computational complexity of Algorithm 4 is analyzed as follows: In the worst case, each sensor node is selected as the target node once, and the movement of the UAV trajectory point towards each target node requires log(ζ0 / ∈5) iterations. Therefore, the worst-case time complexity of Algorithm 4 is... It can be observed that the complexity of Algorithm 4 is linear with respect to N, which is much lower than the polynomial complexity of Algorithm 3.

[0228] IV. Initial Feasibility Points for Joint Optimization Algorithm and Heuristic Algorithm Design:

[0229] Since both the proposed joint optimization algorithm and the heuristic algorithm depend on the initial feasible solution, i.e., A (0) P (0) Q (0) η (0) This invention also needs to design an initialization scheme to generate the initial points of Algorithm 3 and Algorithm 4.

[0230] Specifically, the centroid of the sensor nodes in each cluster can be considered as the UAV trajectory point corresponding to that cluster, so as to provide a relatively fair channel power gain for all sensor nodes in that cluster. However, the UAV trajectory points obtained in this way may not necessarily meet the UAV trajectory length constraint (10). In this case, the trajectory points need to be modified to shorten the trajectory length. Therefore, the main process of obtaining the initial feasible trajectory points is as follows:

[0231] 1) First, use the centroid of the sensor node in each cluster as the initial trajectory point. If constraint (10) is satisfied, it can be directly used as a feasible trajectory point and subsequent steps can be skipped. Otherwise, proceed to the next step.

[0232] 2) Calculate the centroid of all trajectory points, and move each trajectory point towards the centroid to shorten the trajectory length, ensuring that the minimum throughput does not decrease and the trajectory length does not increase. Note that for each trajectory point, the minimum throughput can be calculated by determining the pairing-decoding scheme and transmit power of the sensor nodes in Section III-B. Additionally, to avoid looping, a trajectory point is marked when it can no longer be moved. This step ends when all trajectory points are marked.

[0233] 3) If the trajectory points obtained in the previous step satisfy constraint (10), then exit this step. Otherwise, move the trajectory point farthest from the UAV starting point q0 towards q0 by a fixed step length s. c (where s c >0), so that the center of gravity of the trajectory point moves toward q0, and then return to the previous step.

[0234] The feasible UAV trajectory point Q is obtained through the above steps. (0) Then, the initial pairing-decoding A can be obtained according to section III-B. (0) and transmit power P (0) Thus, the initial points for Algorithm 3 and Algorithm 4 have been obtained.

[0235] In summary, UAVs, with their excellent maneuverability and ability to establish line-of-sight propagation links, have demonstrated great potential in data collection for wireless sensor networks. However, with the continuous increase in sensor nodes and demands, such as low latency and massive access, improving the spectral efficiency of UAV data collection and serving large-scale wireless sensor networks under the constraint of limited spectrum resources in communication systems has become a more challenging requirement. To address the problem of low data collection efficiency caused by limited spectrum resources, this invention proposes a UAV-assisted data collection scheme based on paired access, allowing sensor nodes to transmit data in pairs using Non-Orthogonal Multiple Access (NOMA) technology. By jointly optimizing the pairing scheme, decoding order, transmission power, and UAV trajectory points of sensor nodes, the minimum throughput of sensor nodes is maximized, providing a flexible, efficient, and reliable data collection solution for wireless sensor networks.

[0236] Compared with the prior art, the present invention has at least the following advantages and beneficial effects:

[0237] (1) By grouping and pairing sensor nodes and allowing sensor nodes in each group to share the same frequency resource block, this invention reduces the implementation complexity of non-orthogonal multiple access and achieves a balance between performance and complexity of UAV-assisted data collection system.

[0238] (2) By maximizing the minimum throughput of sensor nodes, the present invention not only improves the throughput of the UAV-assisted data collection system, but also ensures fairness among sensor nodes.

[0239] (3) The present invention fully considers the heterogeneity of sensor nodes, allowing each sensor node to have a different maximum transmit power, and supports each sensor node to control its maximum transmit power according to its own state (e.g., remaining energy), which helps to improve the robustness of wireless sensor networks.

[0240] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0241] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0242] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for designing a drone-assisted data collection system, characterized in that, The method comprises the following steps: constructing a system model, determining a mathematical programming problem according to the constructed system model; splitting the mathematical programming problem into two sub-problems: a pairing and decoding order sub-problem and a transmission power and unmanned aerial vehicle trajectory point sub-problem; designing a joint optimization algorithm based on a block coordinate descent method to solve the pairing and decoding order sub-problem and the transmission power and unmanned aerial vehicle trajectory point sub-problem; designing a heuristic algorithm to optimize the solving of the pairing and decoding order sub-problem and the transmission power and unmanned aerial vehicle trajectory point sub-problem; designing an initial feasible point of the unmanned aerial vehicle according to the joint optimization algorithm and the heuristic algorithm; pairing two sensor nodes to form a group, and the group is referred to as a NOMA group; make Let represent the set of indices for all clusters in a wireless sensor network. Indicates the first The set of indices of all sensor nodes in the cluster, let Indicates the first No. 1 in the cluster One sensor node, Represents sensor nodes Location; Set The sensor node is paired with a sensor node and the decoding order when paired, i.e. when the sensor node is paired with a sensor node into a NOMA group and the signal of the sensor node is decoded later than the signal of the sensor node , otherwise ; Let be the transmit power of the sensor node , be the trajectory point of the UAV in the cluster; wherein a variable describing the pairing of the sensor nodes with the decoding order abbreviated as pair-decode abbreviated as pair-decode abbreviated as pair-decode The constructing a system model, determining a mathematical programming problem according to the constructed system model comprises: The communication link between the UAV and the ground sensor nodes is a line-of-sight propagation link, the sensor nodes Channel power gain between the sensor nodes and the UAV is represented as: In the formula, is a channel power gain, is a flight height of the UAV; Sensor node The corresponding signal-to-interference-plus-noise ratio is expressed as: wherein is the ratio of the channel power gain to the noise power, i.e.: the power spectral density of the additive white Gaussian noise, the channel bandwidth, ; Considering optimizing pairing-decoding of sensor nodes , transmit power and trajectory points of a drone , such that the minimum throughput of the sensor nodes is maximized, the problem is modeled as follows, denoted as P1: in, This represents the minimum throughput. It is a given power difference ratio. For sensor nodes Maximum transmit power, This is an index set of drone trajectory points. As the starting point for drones, The destination of the drone, The maximum allowable trajectory length; constraint (5) ensures To minimize throughput; constraint (6) specifies pair-decoding The range of values; constraint (7) ensures that each sensor node participates in pairing and is paired with only one node; constraint (8) ensures that the receiver on the UAV prioritizes decoding signals with high power; constraint (9) ensures that the transmission power of each sensor node does not exceed its maximum transmission power; constraint (10) limits the flight path length of the UAV to ensure that the UAV can collect data before its onboard energy is exhausted; The unmanned aerial vehicle assisted data collection system comprises: a plurality of sensor nodes, the plurality of sensor nodes are uniformly organized into a plurality of clusters; in each cluster, a plurality of sensor nodes are paired into a group and share a frequency resource block, and different groups use different frequency resource blocks; an unmanned aerial vehicle, configured to collect data of the ground sensor nodes, and after receiving the data of the sensor nodes, adopt a successive interference cancellation technology to cancel multiple access interference in each group to decode signals of the sensor nodes. The joint optimization algorithm based on the block coordinate descent method is designed to solve the pairing and decoding order sub-problem and the transmission power and unmanned aerial vehicle trajectory point sub-problem.

2. A design method according to claim 1, characterized in that The pairing and decoding order sub-problem comprises: In each optimization iteration, the transmit power is first fixed with the trajectory points , the pairing-decoding order subproblem is solved to optimize the pairing-decoding ; then with the pairing-decoding optimized, the transmit power and UAV trajectory points subproblem is solved to optimize the transmit power and the trajectory points .

3. A design method according to claim 1 or 2, characterized in that, At this time, Given transmit power With the UAV trajectory point At time t, the problem P1 is simplified to the following problem, denoted as P2: To further transform the problem P2 into a more tractable form, first introduce an auxiliary variable and define: Therefore, the problem P3 is equivalent to the problem P2. Since the constraint (6) and the constraint (15) are both non-convex constraints, the problem P3 is a non-convex programming, and a successive convex approximation algorithm based on punishment is adopted to solve the problem P3 to obtain an approximate optimal solution of the problem P3. The transmission power and unmanned aerial vehicle trajectory point sub-problem comprises:

4. A design method according to claim 3, wherein, The designing a heuristic algorithm comprises:

5. A design method according to claim 1 or 2, characterized in that, updating the pairing-decoding and transmission power of the sensor nodes to improve the minimum throughput. When pair-decoding Given time, problem P1 becomes the following program problem, denoted as P6: The SCA algorithm is used to solve the sub-problems of the launch power and the UAV trajectory point. After obtaining the optimal trajectory point , the optimal launch power is obtained .

6. The method of claim 1, wherein, The designing an initial feasible point of the unmanned aerial vehicle according to the joint optimization algorithm and the heuristic algorithm comprises: In each iteration of the solving, first find the sensor node with the smallest throughput , its paired sensor node , the cluster it belongs to ; determining the target node according to a preset criterion and the first trajectory point corresponding to the cluster moves to the direction where the target node is located; taking the center of gravity of the sensor nodes in each cluster as an initial trajectory point; if the constraint (10) is satisfied, the initial trajectory point is directly taken as a feasible trajectory point and the subsequent steps are skipped; otherwise, the next step is performed; 7. The method of claim 1, wherein calculating the center of gravity of all trajectory points, and moving each trajectory point to the center of gravity position to shorten the trajectory length under the premise of ensuring that the minimum throughput is not reduced and the trajectory length is not increased; and ​ ​ If the trajectory point obtained in the last step satisfies the constraint (10), exit this step; otherwise, move the trajectory point obtained in the last step by a fixed step length in the direction from the starting point of the UAV to the farthest trajectory point obtained in the last step, so that the barycenter of the trajectory points moves towards the starting point of the UAV, and then return to the last step. the farthest trajectory point obtained in the last step by a fixed step length so that the barycenter of the trajectory points moves towards the starting point of the UAV and then return to the last step.

Images