A method for optimizing the location of a drone relay deployment based on user location uncertainty

Abstract

The application discloses a kind of unmanned aerial vehicle relay deployment position optimization method based on user position uncertainty, comprising: obtaining the position uncertainty information of target user;Obtain the long-term link parameter of unmanned aerial vehicle relay communication system, establish relay constraint, and determine candidate unmanned aerial vehicle position;For each candidate unmanned aerial vehicle position, according to the position uncertainty information of target user, and the long-term link parameter of unmanned aerial vehicle relay communication system generates statistical robust communication representation;According to the statistical injection capability of the first hop between base station and unmanned aerial vehicle, the noise amplification influence of unmanned aerial vehicle amplify and forward relay, and statistical robust communication representation, generate the communication availability index of candidate unmanned aerial vehicle position;From multiple candidate unmanned aerial vehicle positions, screening and determining target communication position, the present application can be used for emergency communication, shielding area communication and temporary coverage scene under unmanned aerial vehicle relay deployment position optimization.

Description

Technical Field

[0001] This invention relates to the field of UAV relay communication technology, and specifically to a method for optimizing the deployment location of UAV relays based on user location uncertainty. Background Technology

[0002] Unmanned aerial vehicles (UAVs) offer advantages such as flexible deployment, good line-of-sight propagation, and rapid access to temporary coverage areas, making them suitable for emergency communications, disaster relief, temporary hotspot coverage, and communication in obstructed areas. When a direct link between a base station and a ground user is absent, blocked, or of poor quality, UAVs can be used as aerial amplify-and-forward (AF) relays to establish a two-hop communication link between the base station, the UAV relay, and the user. However, while relaying useful signals, the AF relay amplifies noise at the UAV relay receiver. Therefore, the UAV's location cannot be determined solely by distance; the quality of the first and second hop links, the impact of noise amplification, and the spatial separability of multiple users must also be considered.

[0003] Existing UAV relay deployment, trajectory planning, or communication resource optimization schemes typically assume that Channel State Information (CSI) at the user's or candidate location is already obtained before optimization. In emergency communication, communication in obstructed areas, or temporary deployment scenarios, the user's location is usually obtained from pilot measurements, network positioning, multi-source sensing, historical observations, or map priors. This information can provide user location estimates and may also provide location uncertainty information. For two-hop amplified relay communication, the UAV's location not only affects the distance between it and the user but also the base station-to-UAV link, the UAV-to-user link, the directional relationship of the multi-user array, and the propagation effect of amplified relay noise. Therefore, candidate UAV location evaluation requires a comprehensive judgment combining location uncertainty information and relay link characteristics.

[0004] Furthermore, the future instantaneous channel state information at candidate drone locations is typically unavailable before the drone reaches those locations. Performing complete instantaneous channel estimation and joint optimization of communication parameters for each candidate drone location would increase online computational complexity and make deployment location optimization dependent on future observations. Therefore, without obtaining the instantaneous channel state information of candidate locations, it is necessary to utilize available user location information, location uncertainty information, and long-term link parameters to generate a communication availability evaluation result for the candidate drone locations, thereby determining the target communication location. The target communication location refers to the location used for downlink relay communication after the drone stops moving. For scenarios with protected objects such as primary users, sensitive receivers, potential eavesdropping nodes, or no-interference zones, signal leakage risk also needs to be considered during the candidate drone location evaluation stage to avoid excessive sacrifice of user communication performance in subsequent leakage suppression. Summary of the Invention

[0005] The purpose of this invention is to provide a method for optimizing the deployment location of UAV relays based on user location uncertainty. This method can be applied to emergency communication, communication in obstructed areas, and temporary coverage scenarios. It can optimize the deployment location of UAV relays when the user location is uncertain and the instantaneous channel state information of the candidate location is not obtained in advance.

[0006] To achieve the above functions, this invention designs a method for optimizing the deployment location of UAV relays based on user location uncertainty. For a UAV relay communication system including a base station, UAVs, and multiple target users, the following steps S1-S5 are executed to complete the optimization of the UAV relay deployment location:

[0007] Step S1: Obtain location uncertainty information for multiple target users, including location estimates and location uncertainty parameters for multiple target users;

[0008] Step S2: Using the communication link between the base station and the UAV as the first hop and the communication link between the UAV and the target user as the second hop, obtain the long-term link parameters of the UAV relay communication system, establish relay constraints, and determine the positions of multiple candidate UAVs based on the relay constraints.

[0009] Step S3: For each candidate UAV location, generate a statistically robust communication representation based on the target user's location estimate and location uncertainty parameters, as well as the long-term link parameters of the UAV relay communication system. This representation is used to characterize the relationship between the second-hop communication link and the directions of multiple target users.

[0010] Step S4: Based on the statistical injection capability of the first hop between the base station and the UAV, the noise amplification effect of the UAV amplified forwarding relay, and the statistical robust communication characterization, generate communication availability indicators for multiple candidate UAV locations.

[0011] Step S5: Based on the communication availability indicators of multiple candidate UAV locations, filter and determine the target communication location from the multiple candidate UAV locations, and use the target communication location as the location for performing downlink relay communication after the UAV stops moving. Calculate the physical indicators of the communication phase after the UAV stops moving, and complete the optimization of UAV relay deployment location.

[0012] As a preferred technical solution of the present invention: In step S1:

[0013] The location estimates for multiple target users include at least one of the posterior mean or the maximum a posteriori estimate;

[0014] The location uncertainty parameters of multiple target users include at least one of the following: posterior covariance, modified covariance, posterior confidence region, posterior sample set, particle set and its weights, location probability density function, or compressed representation of location probability density function.

[0015] As a preferred technical solution of the present invention, the specific steps of step S2 are as follows:

[0016] Step S2.1: Obtain the long-term link parameters of the UAV relay communication system, including the first hop path loss index, the second hop path loss index, the first hop reference distance channel coefficient, the second hop reference distance channel coefficient, the first hop Rice factor, the second hop Rice factor, the base station array response parameters, the UAV array response parameters, the first hop statistical channel covariance, the base station transmit power budget, the UAV relay transmit power budget, and the UAV received noise power or the target user received noise power;

[0017] Based on the long-term link parameters of the first hop, the distance from the base station to the candidate UAV location and the large-scale link coefficient of the first hop are calculated as follows:

[0018] ;

[0019] ;

[0020] in, This indicates the distance from the base station to the candidate drone's location. This represents the first-hop large-scale link coefficient. Indicates the location of the candidate drone. Indicates the location of the base station. This represents the first-hop reference distance channel coefficient. This represents the first-hop path loss exponent;

[0021] Step S2.2: Establish relay constraints for the UAV relay communication system, including:

[0022] Unmanned aerial vehicle (UAV) flight area constraints: ;

[0023] Flight altitude constraints: ;

[0024] Single-step movement distance constraint: ;

[0025] Maximum hovering time slot constraint: ;

[0026] Base station transmit power constraints: ;

[0027] UAV relay transmission power constraints: ;

[0028] Communication availability threshold constraint: ;

[0029] in, Indicates the permitted flight area. This represents the altitude coordinates of the candidate drone's location. Indicates the drone's flight altitude. Indicates the maximum single-step movement distance. Indicates the maximum number of hovering time slots allowed. and These represent the maximum transmit power of the base station and the maximum transmit power of the UAV relay, respectively. Indicates a communication availability threshold;

[0030] Step S2.3: Record the first... The set of candidate UAV locations that satisfy the relay constraint at each optimization time is: .

[0031] As a preferred technical solution of the present invention, the specific steps of step S3 are as follows:

[0032] Step S3.1: For each candidate drone location, generate the statistically robust equivalent distance of the target user relative to the candidate drone location according to the following formula:

[0033] ;

[0034] in, Indicates the first At each optimization moment, the target user's position relative to the candidate drones The statistically robust equivalent distance, Indicates the first The optimization time is the first... Location estimate of a target user For the target user set, Represents the matrix trace operation. Indicates the first The optimization time is the first... Posterior robust covariance of the location of each target user; subscript The superscript 'rob' indicates that the posterior robust covariance of the position has been robustly or contracted. Indicates by the first The observation information obtained at each optimization time point is updated accordingly;

[0035] The location posterior robust covariance is generated from the original posterior covariance:

[0036] ;

[0037] in, Indicates the first The optimization time is directly obtained from the location posterior distribution. The original posterior covariance of each target user; its subscript and Meaning and The same applies; the superscript "raw" indicates the raw posterior statistic without robustness processing. This represents the covariance fusion weight or shrinkage coefficient, and ; Indicates the dimension of the position coordinates; express 3D identity matrix;

[0038] Step S3.2: Generate a statistically robust second-hop line-of-sight representation based on the statistically robust equivalent distance, the long-term link parameters of the second hop, and the UAV array response parameters.

[0039] ;

[0040] in, Indicates the first Candidate drone positions at each optimization moment The corresponding number Statistically robust second-hop sight distance representation for each target user Indicates the first The line-of-sight weight or equivalent channel weight of the second-hop link for each target user. This represents the second-hop reference distance channel coefficient. This represents the loss index of the second hop path. Indicates the drone array relative to the first Array response in the direction of the target user's estimated location;

[0041] Step S3.3: Combine the statistical robust second-hop sight distance representations corresponding to multiple target users to obtain the statistical robust second-hop matrix representation:

[0042] ;

[0043] in, Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. Indicates the number of drone array units. This indicates the conjugate transpose. Indicates the number of target users;

[0044] Step S3.4: Construct the second-hop Gram matrix based on the statistically robust second-hop matrix representation:

[0045] ;

[0046] in, Indicates the first The second hop Gram matrix of the UAV array at each optimized time point;

[0047] Step S3.5: Generate the user-level spatial separability cost based on the regularized inverse matrix of the second-hop Gram matrix, and further generate the user-level second-hop equivalent gain:

[0048] ;

[0049] ;

[0050] in, Represents the regularization coefficient. express 3D identity matrix Represents the first of the matrix diagonal elements, Indicates the first The optimization time is the first... The cost of user-level space separability for each target user. Indicates the first The optimization time is the first... The equivalent gain of the second hop at the user level for each target user.

[0051] As a preferred technical solution of the present invention, the specific steps of step S4 are as follows:

[0052] Step S4.1: Calculate the statistical injection capability of the first hop between the base station and the drone as follows:

[0053] ;

[0054] ;

[0055] in, Indicates the location of candidate drones The corresponding base station to the first hop channel of the drone, This represents the first-hop Gram matrix. Indicates the number of target users. This represents the principal eigenvalues ​​of the first-hop Gram matrix. This represents the equivalent single-stream base station transmit power. This indicates the statistical injection capability of the first hop;

[0056] Step S4.2: Calculate the amplification and forwarding normalization coefficient as follows:

[0057] ;

[0058] in, This represents the budget for the equivalent drone relay transmission power. Indicates the received noise power of the drone. This represents the amplification and forwarding normalization coefficient;

[0059] Step S4.3: Based on the statistical injection capability of the first hop and the amplified forwarding normalization coefficient, calculate the single-user communication contribution as follows:

[0060] ;

[0061] in, Indicates the first The received noise power of the target user Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user;

[0062] Step S4.4: Generate communication availability metrics for candidate UAV locations based on single-user communication contribution.

[0063] As a preferred embodiment of the present invention, the method for generating the communication availability index of candidate UAV locations is one of the following methods:

[0064] The communication availability metric is generated by taking the minimum value of the communication contributions from multiple individual users:

[0065] ;

[0066] in, Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics; Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user. For the target user set;

[0067] The communication availability metric is generated by weighted aggregation of the communication contributions of multiple individual users:

[0068] ;

[0069] in, Indicates the first Aggregate weight of each target user Indicates the first Candidate drone positions generated by weighted aggregation at each optimization time The corresponding communication availability metrics.

[0070] As a preferred embodiment of the present invention: In step S4, if there are protected objects in the UAV relay communication system, the set of protected objects is denoted as... ,in Indicates the first One protected object; construct the target user direction subspace projection matrix based on the statistically robust second-hop matrix representation:

[0071] ;

[0072] in, Indicates the first The target user direction subspace projection matrix corresponding to the UAV array at each optimized time point. Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. This indicates the conjugate transpose. Represents the regularization coefficient. express 3D identity matrix;

[0073] For the For each protected object, calculate the predicted cost of leakage:

[0074] ;

[0075] in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object This represents the equivalent leakage characteristic of the protected object.

[0076] As a preferred embodiment of the present invention: the method for filtering and determining the target communication location from multiple candidate UAV locations in step S5 is one of the following methods:

[0077] Select the candidate drone location with the highest communication availability index as the target communication location;

[0078] ;

[0079] in, Indicates the target communication location. For the set of candidate drone locations; Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics;

[0080] From the set of candidate drone locations, select the candidate drone location with the lowest deployment cost as the target communication location;

[0081] ;

[0082] in, Indicates the communication availability threshold. Indicates the current location of the drone Move to candidate drone location The deployment cost includes at least one of flight distance cost, flight energy consumption cost, or flight time cost;

[0083] The flight distance cost is expressed as:

[0084] ;

[0085] in, Indicates the current location of the drone Move to candidate drone location The cost of flight distance;

[0086] The cost of flight time is expressed as:

[0087] ;

[0088] in, Showing the current location of the drone Move to candidate drone location The cost of flight time; This indicates the maximum flight speed of the drone. Indicates the length of a single flight time slot. Indicates rounding up;

[0089] When candidate drone locations are used to determine the next hovering location, search for candidate reference locations in the preset candidate search area that meet preset conditions for communication availability metrics:

[0090] ;

[0091] in, Indicates the first Candidate reference positions for each optimization time point;

[0092] Based on the one-step reachability constraint of the UAV, the candidate reference position is projected onto the one-step reachable area of ​​the UAV to obtain the next hovering position:

[0093] ;

[0094] ;

[0095] in, Indicates the first An optimized set of locations that a drone can reach in one step. Indicates the next hover position;

[0096] Set the stop conditions as follows:

[0097] ;

[0098] in, Indicates the distance threshold. Indicates the threshold value of the indicator gain. Indicates a positive number to prevent the denominator from being too small;

[0099] When the candidate moving distance is less than the distance threshold and the communication availability index gain is less than the index gain threshold, the current drone position or the next hovering position is determined as the target communication position.

[0100] As a preferred embodiment of the present invention: if there is a protected object in the UAV relay communication system, the specific method of step S5 is as follows:

[0101] Based on the location, location uncertainty, orientation information, or sampling points in the restricted area of ​​the protected object, generate equivalent leakage representations of the protected object corresponding to the locations of the multiple candidate UAVs;

[0102] Construct the target user direction subspace projection matrix based on the statistically robust second hop matrix representation;

[0103] Based on the projection intensity of the equivalent leakage characterization of the protected object onto the target user direction subspace, the predicted leakage cost corresponding to multiple candidate UAV locations is generated.

[0104] The method for filtering and determining the target communication location from multiple candidate UAV locations is one of the following:

[0105] Based on communication availability metrics and predicted leakage costs, the target communication location is selected from multiple candidate drone locations:

[0106] ;

[0107] in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object A set of protected objects; Indicates the first Leakage risk threshold for each protected object;

[0108] Determine the target communication location using a penalty-based approach:

[0109] ;

[0110] in, This indicates the penalty weight.

[0111] As a preferred embodiment of the present invention: In step S5, the method for calculating the physical indicators of the communication phase after the UAV stops moving is as follows:

[0112] Calculate the base station's transmitted signal as follows:

[0113] ;

[0114] in, Represents the base station precoding matrix. Represents the target user data symbol vector;

[0115] Calculate the amplified relay received signal of the UAV. as follows:

[0116] ;

[0117] in, Indicates the target communication location The corresponding base station to the first hop channel of the drone, This indicates the noise received by the drone;

[0118] Calculate the amplified and relayed signal of the UAV as follows:

[0119] ;

[0120] in, This represents the drone amplification relay matrix;

[0121] No. End-to-end signal-to-interference-plus-noise ratio for each target user as follows:

[0122] ;

[0123] in, Indicates drones up to the The second-hop instantaneous channel for the target user Indicates the first The actual location of the target user Indicates the corresponding number in the base station precoding matrix Precoded vectors of data streams for each target user Indicates the received noise power of the drone. Indicates the first The target user receives noise power. For the target user set.

[0124] Beneficial effects: Compared with the prior art, the advantages of the present invention include:

[0125] (1) By introducing position uncertainty parameters to generate statistically robust communication representations, the second-hop link representation of candidate UAV positions includes the impact of target user position estimation error; by inversely generating user-level spatial separability cost through Gram matrix regularization, and further obtaining user-level second-hop equivalent gain, the communication availability index includes multi-target user spatial separability.

[0126] (2) By combining the statistical injection capability of the first-hop base station to UAV link and the noise amplification effect of the UAV amplified relay, the evaluation of the candidate UAV location is matched with the two-hop amplified relay link;

[0127] (3) By predicting the leakage cost, the candidate UAV location screening or penalty is included in the target communication location determination stage to include the leakage risk of the protected object direction. Attached Figure Description

[0128] Figure 1 This is a flowchart of a method for optimizing the deployment location of unmanned aerial vehicle (UAV) relays based on user location uncertainty, according to an embodiment of the present invention.

[0129] Figure 2 This is a schematic diagram of a UAV relay communication system model provided according to an embodiment of the present invention;

[0130] Figure 3This is a comparison chart of the worst user signal-to-interference-plus-noise ratio (SINNR) under the current stopping condition for different schemes provided in the embodiments of the present invention;

[0131] Figure 4 This is a graph showing the cumulative distribution of the worst user signal-to-interference-plus-noise ratio at the target communication location using different methods provided in embodiments of the present invention.

[0132] Figure 5 This is a comparison chart of the leakage suppression effect of the protected object provided by the embodiments of the present invention. Detailed Implementation

[0133] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0134] This invention provides a method for optimizing the deployment location of UAV relay systems based on user location uncertainty. This method is applicable to UAV relay communication systems comprising a base station, UAVs, and multiple target users. Figure 1 Perform the following steps S1-S5 to complete the optimization of the drone relay deployment location:

[0135] Step S1: Obtain location uncertainty information for multiple target users, including location estimates and location uncertainty parameters for multiple target users;

[0136] Let the target user set be... ,in Indicates the number of target users. Indicates the first One target user; Indicates the first Optimize deployment location timing, movement steps, or hovering time slots; Indicates the location of the candidate drone. Indicates the first The current location of the drone at any given moment.

[0137] In the At each deployment location optimization moment, the system acquires location uncertainty information for multiple target users. This location uncertainty information can be obtained by updating the information after receiving uplink pilot signals from target users via a UAV amplified relay, or it can be obtained by fusing at least one of the following: base station-side positioning results, network-side fused positioning results, external sensing results, historical observation results, or prior map information. The system extracts location estimates and location uncertainty parameters from this location uncertainty information.

[0138] The location estimates for multiple target users include at least one of the posterior mean or the maximum a posteriori estimate;

[0139] The location uncertainty parameters for multiple target users include at least one of the following: posterior covariance, corrected covariance, posterior confidence region, posterior sample set, particle set and its weights, location probability density function, or compressed representation of location probability density function. The compressed representation of the location probability density function includes Gaussian mixture model parameters, kernel density estimation parameters, or orthogonal basis expansion coefficients.

[0140] Step S2: Using the communication link between the base station and the UAV as the first hop and the communication link between the UAV and the target user as the second hop, obtain the long-term link parameters of the UAV relay communication system, establish relay constraints, and determine the positions of multiple candidate UAVs based on the relay constraints.

[0141] The specific steps of step S2 are as follows:

[0142] Step S2.1: Obtain the long-term link parameters of the UAV relay communication system. Long-term link parameters refer to the link statistical parameters or model parameters obtainable during the candidate location evaluation phase, excluding future instantaneous channel state information at the candidate location. Specifically, these include the first-hop path loss index, the second-hop path loss index, the first-hop reference distance channel coefficient, the second-hop reference distance channel coefficient, the first-hop Rice factor, the second-hop Rice factor, the base station array response parameters, the UAV array response parameters, the first-hop statistical channel covariance, the base station transmit power budget, the UAV relay transmit power budget, and the UAV received noise power or the target user received noise power.

[0143] Based on the long-term link parameters of the first hop, the distance from the base station to the candidate UAV location and the large-scale link coefficient of the first hop are calculated as follows:

[0144] ;

[0145] ;

[0146] in, This indicates the distance from the base station to the candidate drone's location. This represents the first-hop large-scale link coefficient. Indicates the location of the candidate drone. Indicates the location of the base station. This represents the first-hop reference distance channel coefficient. This represents the first-hop path loss exponent;

[0147] Step S2.2: Establish relay constraints for the UAV relay communication system, including:

[0148] Unmanned aerial vehicle (UAV) flight area constraints: ;

[0149] Flight altitude constraints: ;

[0150] Single-step movement distance constraint: ;

[0151] Maximum hovering time slot constraint: ;

[0152] Base station transmit power constraints: ;

[0153] UAV relay transmission power constraints: ;

[0154] Communication availability threshold constraint: ;

[0155] in, Indicates the permitted flight area. This represents the altitude coordinates of the candidate drone's location. Indicates the drone's flight altitude. Indicates the maximum single-step movement distance. Indicates the maximum number of hovering time slots allowed. and These represent the maximum transmit power of the base station and the maximum transmit power of the UAV relay, respectively. Indicates a communication availability threshold;

[0156] Step S2.3: Record the first... The set of candidate UAV locations that satisfy the relay constraint at each optimization time is: The elements in the candidate UAV location set satisfy flight area constraints, flight altitude constraints, and power or energy consumption constraints set according to the specific scenario, and can be further filtered according to the communication availability threshold after the communication availability index is generated.

[0157] In this embodiment, the system performs spatial meshing within the UAV's flight area, generating a spatial mesh of candidate UAV positions at preset intervals within a plane that satisfies flight altitude constraints. As an alternative implementation, the system determines a candidate search region based on the centroid direction of the target user's estimated position, the direction of the line connecting the base station to the target user, and flight area constraints, and generates a set of candidate UAV positions by uniformly sampling or adaptively refining the sampling within the candidate search region.

[0158] Candidate UAV locations include at least one of the following: candidate hovering locations, next hovering locations, or target communication locations. The candidate hovering locations are candidate points used for communication availability evaluation; the next hovering locations are the next locations the UAV will actually fly to after the current optimization time; and the target communication locations are the locations where the UAV will perform downlink relay communication after it stops moving. The determination logic for these three is as follows: first, in the set of candidate UAV locations... The system evaluates candidate reference positions and then determines the next hovering position based on the one-step reachability constraint of the UAV. When the stopping condition is met or the mission time slot ends, the current UAV position, the next hovering position, or a candidate position that meets the screening conditions is determined as the target communication position.

[0159] Step S3: For each candidate UAV location, generate a statistically robust communication representation based on the target user's location estimate and location uncertainty parameters, as well as the long-term link parameters of the UAV relay communication system. This representation is used to characterize the relationship between the second-hop communication link and the directions of multiple target users.

[0160] The specific steps of step S3 are as follows:

[0161] Step S3.1: For each candidate drone location, generate the statistically robust equivalent distance of the target user relative to the candidate drone location according to the following formula:

[0162] ;

[0163] in, Indicates the first At each optimization moment, the target user's position relative to the candidate drones The statistically robust equivalent distance, Indicates the first The optimization time is the first... Location estimate of a target user For the target user set, Represents the matrix trace operation. Indicates the first The optimization time is the first... Posterior robust covariance of the location of each target user; subscript The superscript 'rob' indicates that the posterior robust covariance of the position has been robustly or contracted. Indicates by the first The observation information obtained at each optimization time point is updated; the above statistical robust equivalent distance introduces the variance term corresponding to the position uncertainty parameter into the distance square term, so that the second-hop link evaluation of the candidate UAV position can consider the impact of user position estimation bias on the equivalent distance and equivalent link gain.

[0164] The location posterior robust covariance is generated from the original posterior covariance:

[0165] ;

[0166] in, Indicates the first The optimization time is directly obtained from the location posterior distribution. The original posterior covariance of each target user; its subscript and Meaning and The same applies; the superscript "raw" indicates the raw posterior statistic without robustness processing. This represents the covariance fusion weight or shrinkage coefficient, and ; Indicates the dimension of the position coordinates; express The above correction method reduces the impact of single-direction covariance estimation bias on the candidate UAV position evaluation by weighting and fusing the original posterior covariance with the isotropic components.

[0167] Step S3.2: Generate a statistically robust second-hop line-of-sight representation based on the statistically robust equivalent distance, the long-term link parameters of the second hop, and the UAV array response parameters.

[0168] ;

[0169] in, Indicates the first Candidate drone positions at each optimization moment The corresponding number Statistically robust second-hop sight distance representation for each target user Indicates the first The line-of-sight weight or equivalent channel weight of the second-hop link for each target user. This represents the second-hop reference distance channel coefficient. This represents the loss index of the second hop path. Indicates the drone array relative to the first The array response in the direction of the target user's position estimate; the statistically robust second-hop line-of-sight characterization is used for candidate UAV position evaluation, and does not represent the true second-hop channel obtained by instantaneous channel estimation at the candidate UAV's position.

[0170] Step S3.3: Combine the statistical robust second-hop sight distance representations corresponding to multiple target users to obtain the statistical robust second-hop matrix representation:

[0171] ;

[0172] in, Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. Indicates the number of drone array units. This indicates the conjugate transpose. Indicates the number of target users;

[0173] Step S3.4: Construct the second-hop Gram matrix based on the statistically robust second-hop matrix representation:

[0174] ;

[0175] in, Indicates the first The second hop Gram matrix of the UAV array at each optimized time point;

[0176] Step S3.5: Generate the user-level spatial separability cost based on the regularized inverse matrix of the second-hop Gram matrix, and further generate the user-level second-hop equivalent gain:

[0177] ;

[0178] ;

[0179] in, Represents the regularization coefficient. express 3D identity matrix Represents the first of the matrix diagonal elements, Indicates the first The optimization time is the first... The cost of user-level space separability for each target user. Indicates the first The optimization time is the first... The user-level second-hop equivalent gain for each target user. When the equivalent directions of multiple target users are nearly collinear or spatial separability decreases. Increase and Decrease.

[0180] Step S4: Based on the statistical injection capability of the first hop between the base station and the UAV, the noise amplification effect of the UAV amplified forwarding relay, and the statistical robust communication characterization, generate communication availability indicators for multiple candidate UAV locations.

[0181] The specific steps of step S4 are as follows:

[0182] Step S4.1: Calculate the statistical injection capability of the first hop between the base station and the drone as follows:

[0183] ;

[0184] ;

[0185] in, Indicates the location of candidate drones The corresponding base station to the first hop channel of the drone, This represents the first-hop Gram matrix. Indicates the number of target users. This represents the principal eigenvalues ​​of the first-hop Gram matrix. This represents the equivalent single-stream base station transmit power. This indicates the statistical injection capability of the first hop; in the embodiment, the statistical injection capability of the first hop can also be generated based on at least one of the following: distance from base station to UAV, first hop path loss, first hop Rice factor, base station array response, UAV array response, first hop statistical channel covariance, or base station transmit power budget.

[0186] Step S4.2: Calculate the amplification and forwarding normalization coefficient as follows:

[0187] ;

[0188] in, This represents the budget for the equivalent drone relay transmission power. Indicates the received noise power of the drone. It represents the amplification and forwarding normalization coefficient; it is used to characterize the amplification and forwarding normalization effect and its noise amplification effect under relay transmission power constraints.

[0189] Step S4.3: Based on the statistical injection capability of the first hop and the amplified forwarding normalization coefficient, calculate the single-user communication contribution as follows:

[0190] ;

[0191] in, Indicates the first The received noise power of the target user Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user;

[0192] Step S4.4: Generate communication availability metrics for candidate UAV locations based on single-user communication contribution.

[0193] The communication availability index for candidate UAV locations is generated using one of the following methods:

[0194] The communication availability metric is generated by taking the minimum value of the communication contributions from multiple individual users:

[0195] ;

[0196] in, Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics; Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user. The target user set; when generated using the minimum value, it is used to reflect the worst-case communication guarantee capability.

[0197] The communication availability metric is generated by weighted aggregation of the communication contributions of multiple individual users:

[0198] ;

[0199] in, Indicates the first Aggregate weight of each target user Indicates the first Candidate drone positions generated by weighted aggregation at each optimization time The corresponding communication availability index. In this embodiment, the communication availability index for the candidate UAV location can be generated using the geometric mean or the harmonic mean.

[0200] If there are protected objects in the UAV relay communication system, let the set of protected objects be denoted as . ,in Indicates the first The system generates multiple candidate UAV locations corresponding to the protected object's equivalent direction or equivalent leakage characterization based on at least one of the following: the protected object's location, location uncertainty, direction information, or sampling points in a no-interference zone. The equivalent leakage characterization represents the directional leakage relationship between the candidate UAV's location and the protected object's direction or area. The protected object includes at least one of the following: primary user, neighboring system users, non-serving users, sensitive receivers, potential eavesdropping nodes, or sampling points in a no-interference zone.

[0201] The system generates a target user direction subspace based on statistical robust communication representation. The target user direction subspace is generated based on at least one of the following: a projection matrix of the statistical robust second-hop matrix representation, a regularized projection matrix, a column space basis, a row space basis, or a singular vector subspace. The target user direction subspace projection matrix is ​​constructed based on the statistical robust second-hop matrix representation:

[0202] ;

[0203] in, Indicates the first The target user direction subspace projection matrix corresponding to the UAV array at each optimized time point. Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. This indicates the conjugate transpose. Represents the regularization coefficient. express 3D identity matrix;

[0204] For the For each protected object, calculate the predicted cost of leakage:

[0205] ;

[0206] in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object This represents the equivalent leakage characteristic of the protected object.

[0207] Step S5: Based on the communication availability indicators of multiple candidate UAV locations, filter and determine the target communication location from the multiple candidate UAV locations, and use the target communication location as the location for performing downlink relay communication after the UAV stops moving. Calculate the physical indicators of the communication phase after the UAV stops moving, and complete the optimization of UAV relay deployment location.

[0208] In step S5, the method for filtering and determining the target communication location from multiple candidate UAV locations is one of the following:

[0209] Select the candidate drone location with the highest communication availability index as the target communication location;

[0210] ;

[0211] in, Indicates the target communication location. For the set of candidate drone locations; Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics;

[0212] From the set of candidate drone locations, select the candidate drone location with the lowest deployment cost as the target communication location;

[0213] ;

[0214] in, Indicates the communication availability threshold. Indicates the current location of the drone Move to candidate drone location The deployment cost includes at least one of flight distance cost, flight energy consumption cost, or flight time cost;

[0215] The flight distance cost is expressed as:

[0216] ;

[0217] in, Indicates the current location of the drone Move to candidate drone location The cost of flight distance;

[0218] The cost of flight time is expressed as:

[0219] ;

[0220] in, Showing the current location of the drone Move to candidate drone location The cost of flight time; This indicates the maximum flight speed of the drone. Indicates the length of a single flight time slot. Indicates rounding up;

[0221] The energy cost of flight is calculated based on the flight distance and the UAV's flight power model.

[0222] When candidate drone locations are used to determine the next hovering location, search for candidate reference locations in the preset candidate search area that meet preset conditions for communication availability metrics:

[0223] ;

[0224] in, Indicates the first Candidate reference positions for each optimization time point;

[0225] Based on the one-step reachability constraint of the UAV, the candidate reference position is projected onto the one-step reachable area of ​​the UAV to obtain the next hovering position:

[0226] ;

[0227] ;

[0228] in, Indicates the first An optimized set of locations that a drone can reach in one step. Indicates the next hover position;

[0229] The preset candidate search area is determined according to one of the following methods: generated within a preset search radius centered on the current position of the UAV; generated within a preset range centered on the estimated centroid of the target user's position; or adaptively determined based on the gradient direction of the communication availability index at the previous optimization time. When the candidate movement distance is less than a preset distance threshold and the communication availability index gain is less than a preset index gain threshold, the system determines the current UAV position or the next hovering position as the target communication position.

[0230] Set the stop conditions as follows:

[0231] ;

[0232] in, Indicates the distance threshold. Indicates the threshold value of the indicator gain. Indicates a positive number to prevent the denominator from being too small;

[0233] When the candidate moving distance is less than the distance threshold and the communication availability index gain is less than the index gain threshold, the current drone position or the next hovering position is determined as the target communication position.

[0234] If there is a protected object in the UAV relay communication system, the specific method for step S5 is as follows:

[0235] Based on the location, location uncertainty, orientation information, or sampling points in the restricted area of ​​the protected object, generate equivalent leakage representations of the protected object corresponding to the locations of the multiple candidate UAVs;

[0236] Construct the target user direction subspace projection matrix based on the statistically robust second hop matrix representation;

[0237] Based on the projection intensity of the equivalent leakage characterization of the protected object onto the target user direction subspace, the predicted leakage cost corresponding to multiple candidate UAV locations is generated.

[0238] The method for filtering and determining the target communication location from multiple candidate UAV locations is one of the following:

[0239] Based on communication availability metrics and predicted leakage costs, the target communication location is selected from multiple candidate drone locations:

[0240] ;

[0241] in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object A set of protected objects; Indicates the first Leakage risk threshold for each protected object;

[0242] Determine the target communication location using a penalty-based approach:

[0243] ;

[0244] in, This indicates the penalty weight.

[0245] After determining the target communication location, the system controls the UAV amplification and relay relay to arrive at or remain at the target communication location. After the UAV stops moving, the system acquires the instantaneous channel state information at the target communication location and generates communication parameters based on the instantaneous channel state information. The communication parameters include at least one of base station precoding parameters, UAV amplification and relay matrix, or leakage suppression parameters; wherein, the base station precoding parameters are used to implement beamforming on the base station side, and the UAV amplification and relay matrix is ​​used to implement amplification and forwarding processing on the relay side. The communication parameters can be generated using joint optimization methods known in the art, such as fractional programming, successive convex approximation, semidefinite relaxation, or gradient iteration.

[0246] Figure 2 This is a schematic diagram of a UAV relay communication system. The base station transmits downlink communication signals; the UAV amplification and relay relay receives signals from the base station and amplifies and forwards them; multiple target users receive communication signals relayed by the UAV amplification and relay relay. The direct link between the base station and the target users may be nonexistent, blocked, or have a link quality below a preset threshold.

[0247] The user location uncertainty region represents the range of uncertainty in the target user's location estimation. The estimated location and the actual location represent the relationship between the location uncertainty information and the actual location. The pilot link is used to update the target user's location uncertainty information. The communication link is used for relay communication between the base station, the UAV amplification relay, and the target user. The UAV trajectory represents the UAV's movement path during the deployment location optimization process.

[0248] Figure 2 The protected object and leakage suppression direction are used to generate the equivalent direction or equivalent leakage characterization of the protected object, and participate in the calculation of predicted leakage cost during the candidate UAV position determination stage. The equivalent leakage characterization of the protected object represents the directional leakage relationship from the candidate UAV position to the direction or region of the protected object. During the communication stage after the UAV stops moving, the leakage suppression direction can serve as a directional reference for leakage constraints or leakage suppression processing.

[0249] The calculation method for physical indicators during the communication phase after the drone has stopped moving is as follows:

[0250] Calculate the base station's transmitted signal as follows:

[0251] ;

[0252] in, Represents the base station precoding matrix. Represents the target user data symbol vector;

[0253] Calculate the amplified relay received signal of the UAV. as follows:

[0254] ;

[0255] in, Indicates the target communication location The corresponding base station to the first hop channel of the drone, This indicates the noise received by the drone;

[0256] Calculate the amplified and relayed signal of the UAV as follows:

[0257] ;

[0258] in, This represents the drone amplification relay matrix;

[0259] No. End-to-end signal-to-interference-plus-noise ratio for each target user as follows:

[0260] ;

[0261] in, Indicates drones up to the The second-hop instantaneous channel for the target user Indicates the first The actual location of the target user Indicates the corresponding number in the base station precoding matrix Precoded vectors of data streams for each target user Indicates the received noise power of the drone. Indicates the first The target user receives noise power. The target user set is defined. The signal-to-interference-plus-noise ratio (SINR) serves as a physical indicator for the communication phase after the UAV has stopped moving. During the candidate UAV location determination phase, this invention does not require prior acquisition of the instantaneous channel state information at the location of each candidate UAV.

[0262] The following are three application examples of the present invention:

[0263] Example 1:

[0264] In this embodiment, the system includes a base station, a UAV amplification relay, and four target users. The base station is equipped with sixty-four transmitting antennas, and the UAV is equipped with sixteen array elements.

[0265] Base station coordinates are meters, the initial coordinates of the drone are meters, the drone's flight altitude is Meters. Target users are distributed in the following areas: With meters as the center and radius Within a meter-wide area.

[0266] The maximum transmission power of the base station is The maximum transmission power of the drone relay is The received noise power of the drone and the received noise power of the target user are both The Rice factor for the first jump and the Rice factor for the second jump are both... The maximum single-step movement distance of the drone is Meters, the maximum permissible number of hovering time slots is .

[0267] In this embodiment, the UAV receives the uplink pilot signal from the target user and updates its position uncertainty information during each hovering time slot. Based on the position uncertainty information and long-term link parameters, the system calculates the communication availability index of each candidate UAV's location and drives the UAV to move towards an area with a higher communication availability index.

[0268] The comparative schemes include a perception-first-fly-then-direct-fly scheme, a positioning-only planning scheme, a statistical planning scheme with known true location, and a non-causal upper bound scheme. The perception-first-fly-then-direct-fly scheme involves the UAV hovering at its initial position for several time slots to update user location uncertainty information before flying directly to the estimated centroid of the target user's location. The positioning-only planning scheme drives the UAV's movement solely based on improved positioning accuracy, without considering communication availability metrics. The statistical planning scheme with known true location assumes the target user's true location is known and performs statistical planning based on long-term link parameters. The non-causal upper bound scheme is an ideal performance upper bound that assumes all future instantaneous channel state information can be obtained in advance.

[0269] Figure 3 The results show a comparison of the worst user signal-to-interference-plus-noise ratio (SINNR) for different schemes under the current stopping condition. Figure 3 It can be seen that in the first During a hovering time slot, the worst-case SINR of the current stopped user in this invention is approximately [value missing]. The location planning scheme is approximately The "sensing first, then direct flight" approach is approximately In the first During a hovering time slot, the solution of this invention is approximately... The location planning scheme is approximately The "sensing first, then direct flight" approach is approximately The above results demonstrate that the solution of the present invention can reach relay deployment areas that support higher current-stop communication quality earlier.

[0270] After the UAV reaches the target communication location and stops moving, the system acquires the instantaneous channel state information at the target communication location and generates communication parameters at the target communication location. Figure 4 The cumulative distribution of the worst user signal-to-interference-plus-noise ratio (SINR) for different methods at the target communication location is shown.

[0271] Depend on Figure 4 It can be seen that the worst user SINR at the target communication location in the scheme of the present invention is mainly distributed in to The distribution of the positioning planning schemes shifted relatively to the left, with about half of the samples below the interval; The distribution center of the "sensing first, then direct flight" scheme is approximately located in... Non-causal upper bound schemes are mainly distributed in to The above results indicate that, by combining location uncertainty, first-hop statistical injection capability, amplified forwarding noise, and second-hop multi-user spatial separability in deployment location optimization, the communication quality at the target communication location can be improved.

[0272] Example 2:

[0273] This embodiment provides a specific implementation of protected object leakage suppression. Based on the basic scenario of Embodiment 1, a protected object is added, located near the target user distribution area, for example, at coordinates [coordinates missing]. Meters. Predicting the leakage risk threshold. The penalty weight is set according to the protection level of the protected object. Typical values ​​were compared in the simulation. The system generates predicted leakage costs during the candidate UAV location determination stage and uses these costs to filter, penalize, or rank the candidate UAV locations. This embodiment compares different leakage handling methods to demonstrate the impact of predicted leakage costs on leakage suppression conditions of the protected object after participating in candidate UAV location screening. Specifically, communication-stage nullification refers to applying leakage suppression only during the communication parameter generation stage after the UAV stops moving; joint nullification refers to introducing predicted leakage costs during the candidate UAV location evaluation stage and applying leakage suppression during the communication stage after the UAV stops moving.

[0274] Figure 5 The comparison of leakage suppression effects on the protected objects is shown. Figure 5 As shown in subgraph (a), without considering leakage suppression, the relative reference leakage power near the direction of the protected object remains at approximately to Within the protected area, no significant leakage depression was formed near the protected direction. After employing communication phase nulling or combined nulling, a leakage depression formed near the protected direction, where the relative reference leakage power at the protected direction could be reduced to approximately [value missing]. .

[0275] Depend on Figure 5 As shown in subgraph (b), without considering leakage suppression, the worst-case user SINR is approximately The protected direction relative to the reference leakage power is close. When using zero-depression in the communication phase, the leakage power can be reduced to approximately The worst user SINR is approximately When using a combined zero-depression method, the leakage power can be controlled to approximately [value missing]. Nearby, the worst user's SINR is approximately The above results indicate that, by incorporating the predicted leakage cost into the selection or penalty of candidate UAV locations, more favorable UAV relay deployment location conditions can be provided for subsequent leakage mitigation.

[0276] Example 3:

[0277] This embodiment provides a specific implementation of a UAV relay deployment location optimization system. The system includes a location uncertainty information acquisition module, a system parameter acquisition module, a candidate UAV location determination module, a statistical robust characterization generation module, a communication availability index generation module, and a target communication location determination module.

[0278] The location uncertainty information acquisition module is used to acquire location uncertainty information of multiple target users; the system parameter acquisition module is used to acquire long-term link parameters and relay constraints; the candidate UAV location determination module is used to determine the locations of multiple candidate UAVs; the statistical robust characterization generation module is used to generate statistical robust communication characterizations corresponding to the candidate UAV locations; the communication availability index generation module is used to generate communication availability indices for candidate UAV locations based on the statistical robust communication characterizations, the statistical injection capability of the link from the first-hop base station to the UAV, and the noise amplification effect of the UAV amplified forwarding relay; the target communication location determination module is used to filter and determine the target communication location.

[0279] In an extended embodiment, the system further includes a predicted leakage processing module. This module generates a predicted leakage cost based on the projection intensity of the equivalent direction of the protected object or the equivalent leakage characterization of the protected object onto the target user's directional subspace, and provides the predicted leakage cost to the target communication location determination module. These modules can be implemented using dedicated hardware circuits, processor-executed software programs, programmable logic devices, or combinations thereof. The steps in these modules are the same as those in the method embodiments and will not be repeated here.

[0280] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A method for optimizing the deployment location of UAV relays based on user location uncertainty, characterized in that, For a drone relay communication system that includes a base station, drones, and multiple target users, perform the following steps S1-S5 to optimize the drone relay deployment location: Step S1: Obtain location uncertainty information for multiple target users, including location estimates and location uncertainty parameters for multiple target users; Step S2: Using the communication link between the base station and the UAV as the first hop and the communication link between the UAV and the target user as the second hop, obtain the long-term link parameters of the UAV relay communication system, establish relay constraints, and determine the positions of multiple candidate UAVs based on the relay constraints. Step S3: For each candidate UAV location, generate a statistically robust communication representation based on the target user's location estimate and location uncertainty parameters, as well as the long-term link parameters of the UAV relay communication system. This representation is used to characterize the relationship between the second-hop communication link and the directions of multiple target users. Step S4: Based on the statistical injection capability of the first hop between the base station and the UAV, the noise amplification effect of the UAV amplified forwarding relay, and the statistical robust communication characterization, generate communication availability indicators for multiple candidate UAV locations. Step S5: Based on the communication availability indicators of multiple candidate UAV locations, filter and determine the target communication location from the multiple candidate UAV locations, and use the target communication location as the location for performing downlink relay communication after the UAV stops moving. Calculate the physical indicators of the communication phase after the UAV stops moving, and complete the optimization of UAV relay deployment location.

2. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, In step S1: The location estimates for multiple target users include at least one of the posterior mean or the maximum a posteriori estimate; The location uncertainty parameters of multiple target users include at least one of the following: posterior covariance, modified covariance, posterior confidence region, posterior sample set, particle set and its weights, location probability density function, or compressed representation of location probability density function.

3. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, The specific steps of step S2 are as follows: Step S2.1: Obtain the long-term link parameters of the UAV relay communication system, including the first hop path loss index, the second hop path loss index, the first hop reference distance channel coefficient, the second hop reference distance channel coefficient, the first hop Rice factor, the second hop Rice factor, the base station array response parameters, the UAV array response parameters, the first hop statistical channel covariance, the base station transmit power budget, the UAV relay transmit power budget, and the UAV received noise power or the target user received noise power; Based on the long-term link parameters of the first hop, the distance from the base station to the candidate UAV location and the large-scale link coefficient of the first hop are calculated as follows: ； ； in, This indicates the distance from the base station to the candidate drone's location. This represents the first-hop large-scale link coefficient. Indicates the location of the candidate drone. Indicates the location of the base station. This represents the first-hop reference distance channel coefficient. This represents the first-hop path loss exponent; Step S2.2: Establish relay constraints for the UAV relay communication system, including: Unmanned aerial vehicle (UAV) flight area constraints: ; Flight altitude constraints: ; Single-step movement distance constraint: ; Maximum hovering time slot constraint: ; Base station transmit power constraints: ; UAV relay transmission power constraints: ; Communication availability threshold constraint: ; in, Indicates the permitted flight area. This represents the altitude coordinates of the candidate drone's location. Indicates the drone's flight altitude. Indicates the maximum single-step movement distance. Indicates the maximum number of hovering time slots allowed. and These represent the maximum transmit power of the base station and the maximum transmit power of the UAV relay, respectively. Indicates a communication availability threshold; Step S2.3: Record the first... The set of candidate UAV locations that satisfy the relay constraint at each optimization time is: .

4. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, The specific steps of step S3 are as follows: Step S3.1: For each candidate drone location, generate the statistically robust equivalent distance of the target user relative to the candidate drone location according to the following formula: ； in, Indicates the first At each optimization moment, the target user's position relative to the candidate drones The statistically robust equivalent distance, Indicates the first The optimization time is the first... Location estimate of a target user For the target user set, Represents the matrix trace operation. Indicates the first The optimization time is the first... Posterior robust covariance of the location of each target user; subscript The superscript 'rob' indicates that the posterior robust covariance of the position has been robustly or contracted. Indicates by the first The observation information obtained at each optimization time point is updated accordingly; The location posterior robust covariance is generated from the original posterior covariance: ； in, Indicates the first The optimization time is directly obtained from the location posterior distribution. The original posterior covariance of each target user; its subscript and Meaning and The same applies; the superscript "raw" indicates the raw posterior statistic without robustness processing. This represents the covariance fusion weight or shrinkage coefficient, and ; Indicates the dimension of the position coordinates; express 3D identity matrix; Step S3.2: Generate a statistically robust second-hop line-of-sight representation based on the statistically robust equivalent distance, the long-term link parameters of the second hop, and the UAV array response parameters. ； in, Indicates the first Candidate drone positions at each optimization moment The corresponding number Statistically robust second-hop sight distance representation for each target user Indicates the first The line-of-sight weight or equivalent channel weight of the second-hop link for each target user. This represents the second-hop reference distance channel coefficient. This represents the loss index of the second hop path. Indicates the drone array relative to the first Array response in the direction of the target user's estimated location; Step S3.3: Combine the statistical robust second-hop sight distance representations corresponding to multiple target users to obtain the statistical robust second-hop matrix representation: ； in, Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. Indicates the number of drone array units. This indicates the conjugate transpose. Indicates the number of target users; Step S3.4: Construct the second-hop Gram matrix based on the statistically robust second-hop matrix representation: ； in, Indicates the first The second hop Gram matrix of the UAV array at each optimized time point; Step S3.5: Generate the user-level spatial separability cost based on the regularized inverse matrix of the second-hop Gram matrix, and further generate the user-level second-hop equivalent gain: ； ； in, Represents the regularization coefficient. express 3D identity matrix Represents the first of the matrix diagonal elements, Indicates the first The optimization time is the first... The cost of user-level space separability for each target user. Indicates the first The optimization time is the first... The equivalent gain of the second hop at the user level for each target user.

5. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, The specific steps of step S4 are as follows: Step S4.1: Calculate the statistical injection capability of the first hop between the base station and the drone as follows: ； ； in, Indicates the location of candidate drones The corresponding base station to the first hop channel of the drone, This represents the first-hop Gram matrix. Indicates the number of target users. This represents the principal eigenvalues ​​of the first-hop Gram matrix. This represents the equivalent single-stream base station transmit power. This indicates the statistical injection capability of the first hop; Step S4.2: Calculate the amplification and forwarding normalization coefficient as follows: ； in, This represents the budget for the equivalent drone relay transmission power. Indicates the received noise power of the drone. This represents the amplification and forwarding normalization coefficient; Step S4.3: Based on the statistical injection capability of the first hop and the amplified forwarding normalization coefficient, calculate the single-user communication contribution as follows: ； in, Indicates the first The received noise power of the target user Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user; Step S4.4: Generate communication availability metrics for candidate UAV locations based on single-user communication contribution.

6. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 5, characterized in that, The communication availability index for candidate UAV locations is generated using one of the following methods: The communication availability metric is generated by taking the minimum value of the communication contributions from multiple individual users: ； in, Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics; Indicates the first Candidate drone positions at each optimization moment For the first Single-user communication contribution of each target user. For the target user set; The communication availability metric is generated by weighted aggregation of the communication contributions of multiple individual users: ； in, Indicates the first Aggregate weight of each target user Indicates the first Candidate drone positions generated by weighted aggregation at each optimization time The corresponding communication availability metrics.

7. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, In step S4, if there are protected objects in the UAV relay communication system, the set of protected objects is denoted as... ,in Indicates the first One protected object; construct the target user direction subspace projection matrix based on the statistically robust second-hop matrix representation: ； in, Indicates the first The target user direction subspace projection matrix corresponding to the UAV array at each optimized time point. Indicates the first The statistically robust second-hop matrix representation of the UAV array at each optimization time point. This indicates the conjugate transpose. Represents the regularization coefficient. express 3D identity matrix; For the For each protected object, calculate the predicted cost of leakage: ； in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object This represents the equivalent leakage characteristic of the protected object.

8. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, In step S5, the method for filtering and determining the target communication location from multiple candidate UAV locations is one of the following: Select the candidate drone location with the highest communication availability index as the target communication location; ； in, Indicates the target communication location. For the set of candidate drone locations; Indicates the first Candidate drone positions at each optimization moment The corresponding communication availability metrics; From the set of candidate drone locations, select the candidate drone location with the lowest deployment cost as the target communication location; ； in, Indicates the communication availability threshold. Indicates the current location of the drone Move to candidate drone location The deployment cost includes at least one of flight distance cost, flight energy consumption cost, or flight time cost; The flight distance cost is expressed as: ； in, Indicates the current location of the drone Move to candidate drone location The cost of flight distance; The cost of flight time is expressed as: ； in, Showing the current location of the drone Move to candidate drone location The cost of flight time; This indicates the maximum flight speed of the drone. Indicates the length of a single flight time slot. Indicates rounding up; When candidate drone locations are used to determine the next hovering location, search for candidate reference locations in the preset candidate search area that meet preset conditions for communication availability metrics: ； in, Indicates the first Candidate reference positions for each optimization time point; Based on the one-step reachability constraint of the UAV, the candidate reference position is projected onto the one-step reachable area of ​​the UAV to obtain the next hovering position: ； ； in, Indicates the first An optimized set of locations that a drone can reach in one step. Indicates the next hover position; Set the stop conditions as follows: ； in, Indicates the distance threshold. Indicates the threshold value of the indicator gain. Indicates a positive number to prevent the denominator from being too small; When the candidate moving distance is less than the distance threshold and the communication availability index gain is less than the index gain threshold, the current drone position or the next hovering position is determined as the target communication position.

9. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, If there is a protected object in the UAV relay communication system, the specific method for step S5 is as follows: Based on the location, location uncertainty, orientation information, or sampling points in the restricted area of ​​the protected object, generate equivalent leakage representations of the protected object corresponding to the locations of the multiple candidate UAVs; Construct the target user direction subspace projection matrix based on the statistically robust second hop matrix representation; Based on the projection intensity of the equivalent leakage characterization of the protected object onto the target user direction subspace, the predicted leakage cost corresponding to multiple candidate UAV locations is generated. The method for filtering and determining the target communication location from multiple candidate UAV locations is one of the following: Based on communication availability metrics and predicted leakage costs, the target communication location is selected from multiple candidate drone locations: ； in, Indicates the first Candidate drone positions at each optimization moment For the first Predicted leakage cost for a protected object A set of protected objects; Indicates the first Leakage risk threshold for each protected object; Determine the target communication location using a penalty-based approach: ； in, This indicates the penalty weight.

10. The method for optimizing the deployment location of UAV relays based on user location uncertainty according to claim 1, characterized in that, In step S5, the calculation method for the physical indicators during the communication phase after the UAV stops moving is as follows: Calculate the base station's transmitted signal as follows: ； in, Represents the base station precoding matrix. Represents the target user data symbol vector; Calculate the amplified relay received signal of the UAV. as follows: ； in, Indicates the target communication location The corresponding base station to the first hop channel of the drone, This indicates the noise received by the drone; Calculate the amplified and relayed signal of the UAV as follows: ； in, This represents the drone amplification relay matrix; No. End-to-end signal-to-interference-plus-noise ratio for each target user as follows: ； in, Indicates drones up to the The second-hop instantaneous channel for the target user Indicates the first The actual location of the target user Indicates the corresponding number in the base station precoding matrix Precoded vectors of data streams for each target user Indicates the received noise power of the drone. Indicates the first The target user receives noise power. For the target user set.

Images