Methods, devices, equipment and media for regional navigation enhancement positioning of unmanned aerial vehicle (UAV) swarms

Abstract

This application relates to a method, apparatus, device, and medium for enhancing regional navigation positioning of an unmanned aerial vehicle (UAV) swarm. The method includes: acquiring first observation data received by a receiving UAV from satellite signals; acquiring second observation data broadcast by a relaying UAV based on the first observation data and collected by a cooperative target formation, wherein the receiving UAV and the relaying UAV are different nodes, and the cooperative target formation has a known fixed geometric configuration; characterizing the positions of the UAV swarm and the cooperative target formation as functions of aerial formation pose parameters and target formation pose parameters, respectively; constructing a multi-layer observation equation, establishing a nonlinear least-squares objective function with the two types of pose parameters as joint unknowns, and obtaining the target formation pose estimate through alternating optimization and iterative solution to determine the positioning results of each cooperative target. This method can reduce the scale of unknowns and achieve high-precision stable positioning under conditions of insufficient observation.

Description

Technical Field

[0001] This application relates to the field of UAV cooperative positioning technology, and in particular to a method, apparatus, device and medium for regional navigation enhancement positioning of UAV swarms. Background Technology

[0002] Global Navigation Satellite Systems (GNSS) are the mainstream positioning and timing infrastructure, widely used in scenarios such as transportation, unmanned system collaboration, and emergency rescue. Conventional satellite navigation relies on receivers to directly receive signals from multiple satellites for pseudorange calculation. In environments with obstructions, complex terrain, or poor satellite distribution, the observation quality, positioning continuity, and accuracy can easily degrade significantly, and it can also affect formation coordination and mission execution.

[0003] Existing emergency navigation and augmented positioning methods are mainly divided into two categories: internal sensor fusion and external information assistance. The former combines inertial, visual, radar and other sensors with satellite navigation to maintain positioning for a short time, but errors are prone to accumulation and have high requirements for environment and cost. The latter uses terrestrial communication, low-orbit satellites, UWB and other means to provide augmented information, which is more flexible, but is limited by coverage and infrastructure dependence, and has limited applicability in temporary deployment and air-ground integrated missions.

[0004] Constructing regional navigation augmentation networks using aerial platforms has emerged as a new direction. These platforms offer flexibility, good line-of-sight conditions, and rapid deployment to form formations, providing users with auxiliary navigation information. However, existing methods still have shortcomings: independently modeling aerial nodes can lead to an excessive number of parameters to be estimated and insufficient observations; integrating transceiver functions with nodes raises electromagnetic compatibility issues; there is limited utilization of both aerial and target formation constraints; node division of labor and information distribution lack systematic design; and the coupling relationships between deployment, signal mapping, and joint solution are unclear. Summary of the Invention

[0005] Therefore, it is necessary to provide a regional navigation augmentation positioning method, device, equipment, and medium for UAV swarms that can achieve stable positioning of multiple targets under limited observation, while taking into account deployment optimization and performance evaluation, in order to address the above-mentioned technical problems.

[0006] A method for enhancing regional navigation and positioning in a drone swarm, the method comprising: Acquire the first observation data, which is obtained by the receiving drone in the drone cluster receiving satellite signals; The second observation data is obtained based on the first observation data. After being transmitted through the information link within the UAV cluster, the second observation data is broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each of the forwarding UAVs are respectively represented as a first function with the aerial formation pose parameters as independent variables; Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables; Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, a multi-layered observation equation is constructed, which is transmitted from the satellite to the cooperative target via the receiving drone and the relaying drone. Using the aerial formation pose parameters and the target formation pose parameters as joint unknowns, a nonlinear least squares objective function is established based on the multi-layer observation equation; Alternating optimization is performed on the objective function to obtain the target formation pose estimate by iteratively updating the joint unknowns; Based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation, the positioning result of each cooperative target is determined.

[0007] In one embodiment, the drone swarm includes three receiving drones and three relaying drones; and the satellite signal of each visible satellite is received by one receiving drone and broadcast by one relaying drone, forming a mapping relationship between satellite signal reception and broadcasting.

[0008] In one embodiment, the target formation pose parameters include the rotation matrix and translation vector of the target formation in the global coordinate system, and the second function includes: The global coordinates of each cooperative target are determined based on the known coordinates of each cooperative target in the local coordinate system of the cooperative target formation, the rotation matrix, and the translation vector.

[0009] In one embodiment, the rotation matrix is ​​represented using Rodrigues vector parameterization.

[0010] In one embodiment, the multi-layer observation equation is a combined observation consisting of a first distance observation term from the satellite to the receiving drone, an information link transmission correction term from the receiving drone to the relaying drone, and a second distance observation term from the relaying drone to the cooperative target.

[0011] In one embodiment, alternating optimization is performed on the objective function, including: With the aerial formation pose parameters fixed, solve the target formation pose parameter subproblem with boundary constraints, and update the target formation pose parameters. The target formation pose parameters are fixed and updated. The relay formation pose parameter subproblem with regularization term is solved to update the air formation pose parameters. The target formation pose parameters and the aerial formation pose parameters are alternately updated until the convergence condition is met, and the target formation pose estimate is obtained.

[0012] In one embodiment, before establishing the objective function, the method further includes: A multi-starting-point initialization strategy is adopted to generate multiple sets of initial values ​​for the aerial formation pose parameters and the target formation pose parameters. The alternating optimization solution is then performed on each set, and the solution with the smallest residual is selected as the target formation pose estimate.

[0013] This application also proposes a regional navigation enhancement positioning device for unmanned aerial vehicle (UAV) swarms, the device comprising: The first acquisition module is used to acquire the first observation data, which is obtained by receiving satellite signals from the receiving drone in the drone cluster; The second acquisition module is used to acquire second observation data. The second observation data is based on the first observation data and is transmitted through the information link within the UAV cluster. It is then broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. The first characterization module is used to characterize the positions of the receiving drone and each of the forwarding drones as a first function with the aerial formation pose parameters as independent variables, based on the known formation topology of the drone swarm. The second characterization module is used to characterize the position of each cooperative target as a second function with the target formation pose parameters as independent variables, based on the known fixed geometric configuration of the cooperative target formation. The equation construction module is used to construct a multi-layered observation equation from the satellite to the cooperative target via the receiving drone and the relaying drone, based on the satellite position, the first function, the second function, the first observation data, and the second observation data. The objective function construction module is used to establish a nonlinear least squares objective function based on the multi-layer observation equation, using the aerial formation pose parameters and the target formation pose parameters as joint unknowns. The optimization solution module is used to perform alternating optimization solutions on the objective function and obtain the target formation pose estimate by iteratively updating the joint unknowns. The positioning determination module is used to determine the positioning result of each cooperative target based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation.

[0014] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps in the above-described method for enhancing the regional navigation and positioning of a drone swarm.

[0015] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method for enhancing the regional navigation and positioning of a drone swarm.

[0016] The aforementioned method, apparatus, equipment, and medium for enhancing regional navigation and positioning of UAV swarms acquire first observation data obtained by receiving satellite signals from a receiving UAV within the swarm, and second observation data obtained by transmitting the first observation data via an internal information link within the UAV swarm, broadcasting it to a cooperative target formation to be located by a relaying UAV, and then collecting it from the cooperative target formation. Here, the receiving UAV and the relaying UAV are different UAV nodes within the UAV swarm, and the cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each relaying UAV are respectively represented as first functions with aerial formation pose parameters as independent variables. Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables. Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, a multi-layer observation equation is constructed from the satellite to the cooperative target via the receiving UAV and the relaying UAV. With the aerial formation pose parameters and the target formation pose parameters as joint unknowns, a nonlinear least square objective function is established based on the multi-layer observation equation. Alternating optimization is performed on the objective function, and the target formation pose estimate is obtained by iteratively updating the joint unknowns. Based on the target formation pose estimate and the known fixed geometric configuration of each cooperative target in the cooperative target formation, the positioning result of each cooperative target is determined.

[0017] Beneficial effects: This method models the known formation topology of the UAV swarm and the known fixed geometric configuration of the cooperative target formation separately. It uniformly represents the position of each node as a function with the aerial formation pose parameters and the target formation pose parameters as independent variables. This compresses the large number of independent coordinate unknowns that originally required point-by-point estimation into a small number of pose parameters, effectively solving the underdetermined problem under insufficient observation conditions and improving the solvability and stability of the positioning. Simultaneously, the first and second observation data acquired collaboratively by the receiving and relaying UAVs separate the satellite signal receiving function from the regional broadcasting function at the data acquisition level, avoiding the electromagnetic compatibility problems caused by the co-transmitter / receiver. Based on this, a multi-layered observation equation is constructed from the satellite to the cooperative target via the receiving and relaying UAVs. A nonlinear least-squares objective function is established with the aerial formation pose parameters and the target formation pose parameters as joint unknowns. An alternating optimization strategy is used to iteratively update and solve the joint unknowns, achieving high-precision positioning of the cooperative target formation under complex geometric configurations. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a method for enhancing the regional navigation and positioning of a drone swarm in one embodiment. Figure 2 This is a schematic diagram of the air and ground formation composition in one embodiment; Figure 3 This is a schematic diagram of signal delay in one embodiment; Figure 4 This is a schematic diagram of the overall processing flow architecture of the method in one embodiment; Figure 5 This is a schematic diagram of the joint localization visualization results in one embodiment, wherein, Figure 5 (a) shows a schematic diagram of the 3D positioning results. Figure 5 (b) shows a schematic diagram of relay error. Figure 5 (c) shows a schematic diagram of the target error. Figure 5 (d) shows a diagram comparing relative distances; Figure 6 This is a structural block diagram of a regional navigation enhancement and positioning device for a drone swarm in one embodiment; Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] To address the problems in existing technologies, such as modeling airborne nodes independently leading to too many parameters to be estimated and insufficient observations, integrating transmission and reception functions with the same node causing electromagnetic compatibility issues, and the limited solvability and accuracy of positioning due to the infrequent use of both airborne and target formation constraints, such as... Figure 1 As shown, this application provides a regional navigation enhancement positioning method for unmanned aerial vehicle (UAV) swarms, which specifically includes the following steps: Step S100: Obtain the first observation data, which is obtained by the receiving drone in the drone cluster receiving satellite signals.

[0021] Step S110: Obtain second observation data. The second observation data is based on the first observation data and is transmitted through the information link within the UAV cluster. It is then broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations.

[0022] Step S120: Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each forwarding UAV are represented as first functions with the aerial formation pose parameters as independent variables.

[0023] Step S130: Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables.

[0024] Step S140: Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, construct a multi-layered observation equation from the satellite to the receiving drone and the relaying drone to the cooperative target.

[0025] Step S150: Using the aerial formation pose parameters and the target formation pose parameters as joint unknowns, establish a nonlinear least squares objective function based on the multi-layer observation equation.

[0026] Step S160: Perform alternating optimization on the objective function and obtain the target formation pose estimate by iteratively updating the joint unknowns.

[0027] Step S170: Determine the positioning results of each cooperative target based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation.

[0028] This application provides a regional navigation enhancement and positioning method for aerial formations, i.e., UAV swarms. It is applicable to scenarios where cooperative targets within a region are distributed in a fixed formation, aerial UAV nodes can maintain a stable relative configuration, some UAVs have satellite signal receiving capabilities, and there are information links between UAVs. This method uses a small number of receiving UAVs to acquire satellite navigation signals, and then maps the received navigation signals or equivalent observations to other UAVs through internal formation information links. The latter then undertake the tasks of regional signal enhancement and observation geometric expansion, forming a multi-layered observation relationship between satellites, receiving UAVs, relaying UAVs, and fixed formation targets. Combined with formation constraints, multi-target joint positioning is achieved.

[0029] In the above method, the aerial UAV formation acts as the provider of navigation enhancement services, utilizing its known formation topology and stable relative configuration to provide a spatial reference for positioning calculations. The ground cooperative target formation acts as the receiver of positioning services, with each target having a known fixed geometric configuration. By modeling the aerial formation and the target formation as rigid bodies respectively, and representing their positions uniformly with their corresponding pose parameters, a large number of independent node coordinate unknowns are compressed into a small number of pose parameters, achieving high-precision joint positioning of the target formation under limited observation conditions.

[0030] In this embodiment, the aerial drone formation is composed of It consists of several drones, denoted as .

[0031] The drone swarm comprises three receiving drones and three relaying drones. The primary receiving drone receives satellite signals, while the secondary relaying drones do not directly receive satellite signals. Instead, they obtain corresponding navigation signal data, equivalent ranging information, or correction information from the primary receiving drone via inter-drone information links, and then broadcast or relay this information to cooperative targets within the area. This design separates the receiving function from the large-scale transmission function at physical nodes, significantly reducing the isolation pressure when transmitting and receiving are integrated, and lowering the complexity of platform hardware design. To balance system complexity and observation capabilities, this method employs a mapping method where one receiving drone receives signals from two satellites. Each visible satellite's signal is received by one receiving drone and broadcast by one relaying drone, establishing a mapping relationship between satellite signal reception and broadcasting. Simultaneously, the targets to be located within the area are arranged in a fixed formation, allowing this invention to apply rigid body constraints not only to the aerial drone formation but also to the target formation, thereby further compressing the dimensionality of unknown parameters and improving overall positioning solvability. A schematic diagram of the system model of this method is shown below. Figure 2 As shown.

[0032] Specifically, in step S100, the receiving drone in the drone swarm uses its onboard satellite navigation receiving module to capture and track visible satellite signals under the current spatiotemporal conditions, acquiring first observation data including pseudorange, carrier phase, or Doppler shift. The receiving drone refers to a drone node in the swarm that has satellite signal receiving capabilities but does not undertake the function of broadcasting enhanced information to the cooperative target formation.

[0033] Specifically, in step S110, after acquiring the first observation data, the receiving UAV transmits the first observation data or equivalent observations generated based on the first observation data to the relaying UAV via the information link within the UAV swarm. The relaying UAV generates enhanced navigation information based on the received data and broadcasts it to the cooperative target formation to be located. Each cooperative target in the formation receives and collects this information, forming the second observation data. Here, the relaying UAV and the receiving UAV are different nodes in the UAV swarm, achieving separation of satellite signal reception and regional broadcasting functions on the physical platform. The cooperative target formation is a rigid formation composed of multiple cooperative targets, each with a known fixed geometric configuration.

[0034] Specifically, no. The satellite signal was received by the drone. The drone is responsible for receiving and relaying the message. If the broadcaster is responsible for broadcasting, the relationship between satellite signal reception and relay can be determined by the mapping pair. This indicates that in aerial formations, satellite signal receiving nodes and regional broadcasting nodes are located on different UAV platforms, and the receiving and transmitting functions are not coupled within the same node. This helps reduce the difficulty of transmission and reception isolation on small UAV platforms and improves the flexibility of the system structure.

[0035] In step S120, taking advantage of the stable and pre-calibrated relative geometric relationships within the aerial UAV formation, the entire UAV cluster is regarded as a rigid body, and the global position of all UAV nodes is uniformly expressed by pose parameters with six degrees of freedom (three-dimensional rotation and three-dimensional translation), thereby significantly reducing the number of parameters to be estimated.

[0036] Let the first The known coordinates of the drones in the aerial formation coordinate system are: The rotation matrix of the aerial formation in the global coordinate system is: The translation vector is Then the first The global coordinates of the drone are: (1) Therefore, the global coordinates of all receiving and relaying drones are expressed as aerial formation pose parameters. , This is a function of the independent variable, i.e., the first function. At this point, the positions of all drones in the aerial formation are no longer considered as... Each unknown is treated as an independent unknown, while the pose parameters of the six-dimensional rigid body are used for unified characterization.

[0037] To facilitate subsequent nonlinear optimization solutions, the rotation matrix... Rodrigues vectors are used for parameterization, thus representing the rotation components as three-dimensional vectors and avoiding the solution difficulties caused by orthogonal matrix constraints. Let the aerial formation rotation vector matrix be... Then the first function can be expressed as: (2) In formula (2), A 3×3 third-order identity matrix, with diagonal elements all being 1 and all other elements being 0, can be represented as: .

[0038] In step S130, similar to the aerial drone formation, this step also treats the cooperative target group to be located as a rigid formation, and uses the known fixed geometric relationships within it to compress the independent coordinate unknowns of multiple targets into a unified six-dimensional pose parameter.

[0039] Assuming the configuration of the target fixed formation in its local coordinate system is known, the first... The local coordinates of the target are The rotation matrix of the target formation in the global coordinate system is: The translation vector is Then the first The global coordinates of the target are: (3) At this point, the global coordinates of all cooperative targets are expressed as target formation pose parameters. , The function of the independent variable is the second function. The positions of all targets do not need to be estimated independently point-by-point, but are represented solely by the pose parameters of the six-dimensional rigid body. In this way, the joint localization problem of multiple cooperative targets within the region is further dimensionality-reduced.

[0040] Furthermore, if individual adjustments to a few targets are required during the mission, local offsets can be superimposed on the rigid body formation model to balance the overall constraints of the formation with the flexible adjustment of individual targets.

[0041] Similarly, the rotation matrix of the target formation Rodriguez vectors are also used for parameterization to facilitate joint optimization. Let the target formation rotation matrix be... Then the second function can be expressed as: (4) Through the dual-rigid-body modeling in steps S120 and S130, the estimation that was originally required... The independent position coordinates were compressed into 12 pose parameters for two rigid bodies, effectively overcoming the underdetermined problem caused by insufficient observations, and laying the foundation for subsequent construction of a unified multi-layer observation equation and joint optimization solution.

[0042] In step S140, the various types of information acquired and constructed in the preceding steps are integrated to establish a full-link observation mathematical model from satellite to target. For example... Figure 3 As shown, for the first The satellite signal is acquired by a receiving drone and mapped to an auxiliary relay drone through an information link, and then broadcast to target nodes within the coverage area, thus establishing a combined observation model.

[0043] Specifically, in Figure 3 In the diagram, the signal from the nth satellite is acquired by the receiving UAV node m, transmitted to the auxiliary relay node r via the inter-UAV information link, and then broadcast to the kth user by node r. : The clock difference of the nth satellite clock relative to the system reference time, in seconds. : The clock difference between the user receiver clock and the system reference time, in seconds. : The launch timestamp of the navigation signal written by the nth satellite. : The timestamp of signal reception recorded by the k-th user receiver. n,m r: The spatial propagation distance from the nth satellite to the receiving UAV node m, in meters. m,r Δlink: The broadcast propagation distance from the auxiliary forwarding node r to the receiving drone node m, in meters. n The link transmission delay, in seconds, is the time it takes for the receiving node m to transmit signals or equivalent observation information to the forwarding node r.

[0044] In this embodiment, the multi-layer observation equation is a combined observation consisting of the first distance observation term from the satellite to the receiving UAV, the information link transmission correction term from the receiving UAV to the relaying UAV, and the second distance observation term from the relaying UAV to the cooperative target.

[0045] Specifically, let the person in charge of receiving the first The location of the drone receiving the satellite signal is The location of the auxiliary relay drone responsible for broadcasting the satellite's corresponding augmentation information is... The receiving drone location and the assisting relay drone location are given by the first function in step S120, and the target location. Given the second function in step S130, its corresponding equivalent observation can be expressed as: (5) In formula (5), This represents the equivalent transmission correction amount of the information link from the receiving UAV to the assisting relay UAV. This correction amount can be compensated for through system calibration, link timing, or synchronization protocols. To measure noise. When the transmission delay of the information link between UAVs is much smaller than the overall ranging error tolerance, and the link synchronization error has been pre-corrected, This can be considered as a known correction and incorporated into the system constant term. In this case, the combined observation model can be simplified to: (6) The multi-layered observation equation shown in Equation (6) indicates that an effective augmentation observation consists of a spatial propagation term from the satellite to the receiving UAV, an information mapping term from the receiving UAV to the auxiliary relay UAV, and a broadcast propagation term from the auxiliary relay UAV to the target node. The receiving position and the broadcast position can be modeled separately, which can more accurately reflect the engineering implementation mechanism of the transmit-receive separation navigation augmentation system.

[0046] In step S150, based on the multi-layer observation equations constructed in step S140, the regional navigation enhancement positioning problem based on the transmit-receive separation air formation is transformed into a constrained nonlinear least squares optimization problem for unified modeling.

[0047] Specifically, the pose parameters of the relay formation (aerial drone swarm) are defined as follows: The pose parameters of the target formation are The joint unknown parameters are: Definition of the first The satellite signal arrived at the first The measurement residuals for each target are: (7) In formula (7), Indicates the first The location of the target, affected by constraint.

[0048] Furthermore, by introducing a regularization term into the residual vector that is equivalent to the two sets of rotation vectors and the distance between the Earth's centers, a joint optimization objective function is constructed as follows: (8) In formula (8), For rotational regularization weights, The distance to the geocentric regularization weight is used to constrain the physical feasibility of the solution. The objective function shown in Equation (8) jointly models and optimizes the pose parameters of the two rigid bodies, the aerial formation and the target formation. The positions of all UAV nodes and target nodes are indirectly expressed through their respective pose parameters, and the observation residuals directly act on the pose space.

[0049] Because the objective function has high-dimensional non-convex properties, direct global solution is prone to getting trapped in local minima. Therefore, in step S160, an alternating optimization strategy is needed for iterative solution. The basic process of alternating optimization is as follows: First, fix the aerial formation pose parameters, solve the target formation pose parameter subproblem with boundary constraints, update the target formation pose parameters, fix the updated target formation pose parameters, solve the relay formation pose parameter subproblem with regularization terms, and update the aerial formation pose parameters. Alternately update the target formation pose parameters and the aerial formation pose parameters until the convergence condition is met, and obtain the target formation pose estimate.

[0050] Specifically, the target pose and relay pose are updated using an alternating iterative method: the relay pose is fixed. Solving the target pose subproblem with boundary constraints: (9) In formula (9), This represents the total number of target nodes in the target fixed formation, with the target index being... , This indicates the total number of valid satellite signals or equivalent augmentation observation links participating in the joint positioning, with the satellite / observation index being... , The first term determined by the target formation pose parameters Target locations.

[0051] Fixed target pose Solve the relay pose problem with regularization terms: (10) However, relying solely on the aforementioned basic alternating optimization process may still lead to convergence to a local minimum or a mirrored fuzzy solution due to improper initial value selection. To address this, this method introduces an initialization phase after establishing the objective function and before executing the alternating optimization solution, and constructs a complete four-stage solution framework around alternating optimization, progressively advancing from initial value generation, mirror correction, accuracy improvement to anomaly recovery.

[0052] In this embodiment, after establishing the objective function and before performing alternating optimization, an initialization phase is included. First, a candidate origin mechanism is used for baseline selection. By translating the origin of the coordinate system to a reasonable reference position, the numerical ill-conditioned problem in the subsequent optimization process is alleviated. Based on this, the initial values ​​of the aerial formation pose parameters are fixed. The satellite-to-ground link distance from the satellite to the receiving UAV is subtracted from the observation data, simplifying the problem to a single-segment observation from the UAV to the target. Then, a multi-point positioning method combined with the Procrustes algorithm is used to obtain the target formation pose parameters. The initial estimates are obtained by introducing random perturbations or sampling at different scales. Multiple initial values ​​for the aerial formation pose parameters and the target formation pose parameters are generated from the single initialization process described above. These initial values ​​are then substituted into the alternating optimization solution process for iteration, yielding multiple candidate solutions. Finally, the solution with the smallest residual is selected as the preliminary estimate. This multi-starting-point initialization strategy effectively reduces the sensitivity of alternating optimization to initial values ​​and improves the reachability of the global optimum.

[0053] To address the complex nonlinear cooperative localization model (Equation 8) constructed above, and to achieve a balance between computational efficiency and global optimality, this method, based on the initialization described above, designs a four-stage solution framework: First, based on the selected origin, multiple random starting points of different scales are generated, and the above alternating optimization is performed in parallel. The solution with the smallest residual is selected as the initial estimate to overcome the problem of easily getting trapped in local minima with a single initial value. Second, using the plane formed by the relay UAV formation, the mirror pose of the target rigid body is calculated. The cost and elevation error of the original configuration and the mirror configuration are compared to identify and correct potential errors. The mirror blur that occurs ensures the correct spatial orientation of the solution result; in the third stage, a strategy of gradually reducing the regularization weight is adopted to perform joint least squares nonlinear optimization. In the early stage of iteration, a larger regularization weight is used to ensure the stability of the solution. As convergence occurs, the regularization weight is gradually reduced to improve the data fitting accuracy, thereby achieving a coarse-to-fine pose estimation refinement; in the fourth stage, it is evaluated whether the final residual is significantly higher than the theoretical noise basis. If a preset threshold is triggered, the spatial search is re-performed by combining the mirror solution and local perturbations to escape the possible local minima traps and ensure the robustness of the solution and the reliability of the positioning in complex environments.

[0054] Specifically, in the first stage mentioned above, based on the selected origin, multiple random starting points of different scales are generated, and alternating optimization is performed in parallel. The solution with the smallest residual is selected as the initial estimate to overcome the problem that a single initial value is prone to getting trapped in local minima.

[0055] Specifically, in the second stage described above, under the dual-rigid-body modeling framework, the target formation may have a mirror-symmetric solution with respect to the plane where the relay UAV formation is located. That is, the optimization solution may converge to a mirror configuration with the spatial orientation opposite to the actual position. In this stage, after obtaining the preliminary estimated solution from the first stage, the plane formed by the relay UAV formation is used as a reference datum to calculate the mirror pose of the target rigid body with respect to this plane. The objective function cost and elevation error corresponding to the original configuration and the mirror configuration are calculated separately and compared. If both the cost and elevation error indicate that the mirror configuration is more physically plausible, the original solution is determined to have mirror ambiguity, and the solution is corrected to the pose parameters corresponding to the mirror configuration; otherwise, the original solution is retained. This stage effectively eliminates the spatial orientation ambiguity caused by rigid-body symmetry, ensuring that the spatial orientation of the solution result is consistent with physical reality.

[0056] Specifically, in the third stage described above, the pose estimation solutions obtained in the first and second stages still have room for improvement in convergence accuracy. This stage, supported by the regularization term in the objective function, employs a strategy of gradually reducing the regularization weight for joint least-squares nonlinear optimization. In the early stages of iteration, a larger regularization weight is set to ensure stable convergence of the optimization process through strong constraints from the regularization term, avoiding oscillations in regions far from the true value. As the convergence progresses, the regularization weight is gradually reduced to a smaller value, allowing the observation residual term to dominate the objective function, thereby fully fitting the observation data and improving the estimation accuracy of the pose parameters. Through this gradual refinement process from coarse to fine, the pose estimation results gradually approach a high-precision solution while maintaining stability.

[0057] Specifically, in the fourth stage described above, to achieve robust localization of the system in complex environments, this stage verifies the reliability of the final solution output from the third stage. First, a theoretical noise floor is determined based on the statistical characteristics of the observed noise, and a preset threshold for residual anomaly detection is set accordingly. If the residual of the final objective function is significantly higher than this threshold, it indicates that the current solution may have large unresolved biases or is still trapped in local minima. In this case, by combining the mirror solution calculated in the second stage with the local small perturbations of the current solution, candidate starting points for supplementary searches are generated, and spatial search and alternating optimization are performed again to escape the remaining local minima traps. If the residual is within the threshold range, the current solution is directly output as the final result. This stage, as the final safeguard of the solution framework, ensures the robustness of the solution and the reliability of localization in complex scenarios such as poor observation conditions and high noise levels.

[0058] Through the above four-stage solution framework, this method organically integrates multi-startpoint global search, mirror deblurring, stepwise refinement and anomaly recovery, and achieves high-precision and robust estimation of the pose parameters of cooperative target formations under limited observation conditions.

[0059] like Figure 4 The diagram shows the overall processing architecture of the UAV swarm area navigation enhancement positioning method. This diagram illustrates the complete processing chain in a modular fashion, from observation data acquisition and formation constraint modeling to joint optimization solution and result output. It mainly includes an observation receiving module, an airborne information forwarding module, a target observation and constraint modeling module, and a joint optimization solution and result output module.

[0060] Specifically, in the observation and reception module, the receiving UAV first selects a visible satellite and receives GNSS signals to obtain the first observation data. In the airborne information forwarding module, the first observation data is transmitted to each forwarding UAV (UAV 1, UAV 2, UAV 3) via the inter-UAV link. The forwarding UAVs generate equivalent observation information and broadcast enhanced navigation information; simultaneously, the cooperative target receives the enhanced navigation information to form the second observation data. In the target observation and constraint modeling module, double-rigid-body constraint modeling is performed based on the relative geometric configuration of the airborne formation and the target formation, and a unified nonlinear least-squares objective function is constructed by combining combined observations. In the joint optimization solution and result output module, multi-starting-point initialization, alternating optimization solution, and joint refinement are executed sequentially to obtain the airborne formation pose estimate and the target formation pose estimate. If the residuals meet the preset threshold, the target formation position result and the airborne UAV formation pose result within the region are output; otherwise, after mirror correction and additional constraint filtering, the solution is returned until the conditions are met. Finally, relevant visualization charts can be output. This processing flow embodies the organic unity of transmission and reception separation, double-rigid-body dimensionality reduction modeling, and multi-stage optimization solution.

[0061] In the aforementioned regional navigation enhancement positioning method for UAV swarms, a transmit-receive separation architecture is used. A small number of UAVs are dedicated to receiving satellite signals and transmitting them to relaying UAVs via inter-UAV links for regional broadcasting. This physically avoids the electromagnetic compatibility problem of high-sensitivity reception and high-power transmission on a single platform, lowers the hardware design threshold, and successfully achieves collaborative positioning with "3 receivers + 3 relays." Furthermore, both the UAV formation and the target group are modeled as rigid bodies. Known geometric constraints within the formation are used to compress a large number of independent coordinate unknowns into a small number of pose parameters, effectively solving the underdetermined problem under limited observation conditions. Figure 5 As shown, under complex geometric configurations, the estimated relative distance to the target is highly consistent with the true value, and the average positioning error of both the relay UAV and the target group remains at a low level. Meanwhile, a multi-layer observation model of "satellite-receiver-transponder-target" is constructed, and a multi-starting point initialization alternating optimization strategy is adopted. Mirror correction and residual verification mechanisms are introduced to effectively overcome the problems of local optima and geometric flipping, achieving accurate matching between the true and estimated positions in three-dimensional space, ensuring the robustness and positioning reliability of the system in complex environments.

[0062] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0063] In one embodiment, such as Figure 6 As shown, a regional navigation enhancement positioning device for a drone swarm is provided, comprising: a first acquisition module 200, a second acquisition module 210, a first characterization module 220, a second characterization module 230, an equation construction module 240, an objective function construction module 250, an optimization solution module 260, and a positioning determination module 270, wherein: The first acquisition module 200 is used to acquire first observation data, which is obtained by receiving satellite signals from a receiving drone in the drone cluster. The second acquisition module 210 is used to acquire second observation data. The second observation data is based on the first observation data and is transmitted through the information link within the UAV cluster. It is then broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. The first characterization module 220 is used to characterize the positions of the receiving drone and each of the forwarding drones as a first function with the aerial formation pose parameters as independent variables, based on the known formation topology of the drone cluster. The second characterization module 230 is used to characterize the position of each cooperative target as a second function with the target formation pose parameters as independent variables, based on the known fixed geometric configuration of the cooperative target formation. The equation construction module 240 is used to construct a multi-layered observation equation from the satellite to the cooperative target via the receiving UAV and the relaying UAV, based on the satellite position, the first function, the second function, the first observation data, and the second observation data. The objective function construction module 250 is used to establish a nonlinear least squares objective function based on the multi-layer observation equation, using the aerial formation pose parameters and the target formation pose parameters as joint unknowns. The optimization solution module 260 is used to perform alternating optimization solution on the objective function and obtain the target formation pose estimate by iteratively updating the joint unknowns; The positioning determination module 270 is used to determine the positioning result of each cooperative target based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation.

[0064] Specific limitations regarding the area navigation enhancement and positioning device for UAV swarms can be found in the limitations of the area navigation enhancement and positioning method for UAV swarms mentioned above, and will not be repeated here. Each module in the aforementioned area navigation enhancement and positioning device for UAV swarms can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0065] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a regional navigation augmentation positioning method for a drone swarm. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0066] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0067] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: Acquire the first observation data, which is obtained by the receiving drone in the drone cluster receiving satellite signals; The second observation data is obtained based on the first observation data. After being transmitted through the information link within the UAV cluster, the second observation data is broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each of the forwarding UAVs are respectively represented as a first function with the aerial formation pose parameters as independent variables; Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables; Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, a multi-layered observation equation is constructed, which is transmitted from the satellite to the cooperative target via the receiving drone and the relaying drone. Using the aerial formation pose parameters and the target formation pose parameters as joint unknowns, a nonlinear least squares objective function is established based on the multi-layer observation equation; Alternating optimization is performed on the objective function to obtain the target formation pose estimate by iteratively updating the joint unknowns; Based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation, the positioning result of each cooperative target is determined.

[0068] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Acquire the first observation data, which is obtained by the receiving drone in the drone cluster receiving satellite signals; The second observation data is obtained based on the first observation data. After being transmitted through the information link within the UAV cluster, the second observation data is broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each of the forwarding UAVs are respectively represented as a first function with the aerial formation pose parameters as independent variables; Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables; Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, a multi-layered observation equation is constructed, which is transmitted from the satellite to the cooperative target via the receiving drone and the relaying drone. Using the aerial formation pose parameters and the target formation pose parameters as joint unknowns, a nonlinear least squares objective function is established based on the multi-layer observation equation; Alternating optimization is performed on the objective function to obtain the target formation pose estimate by iteratively updating the joint unknowns; Based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation, the positioning result of each cooperative target is determined.

[0069] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0070] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for enhancing regional navigation and positioning in an unmanned aerial vehicle (UAV) swarm, characterized in that, The method includes: Acquire the first observation data, which is obtained by the receiving drone in the drone cluster receiving satellite signals; The second observation data is obtained based on the first observation data. After being transmitted through the information link within the UAV cluster, the second observation data is broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. Based on the known formation topology of the UAV swarm, the positions of the receiving UAV and each of the forwarding UAVs are respectively represented as a first function with the aerial formation pose parameters as independent variables; Based on the known fixed geometric configuration of the cooperative target formation, the position of each cooperative target is represented as a second function with the target formation pose parameters as independent variables; Based on the satellite position, the first function, the second function, the first observation data, and the second observation data, a multi-layered observation equation is constructed, which is transmitted from the satellite to the cooperative target via the receiving drone and the relaying drone. Using the aerial formation pose parameters and the target formation pose parameters as joint unknowns, a nonlinear least squares objective function is established based on the multi-layer observation equation; Alternating optimization is performed on the objective function to obtain the target formation pose estimate by iteratively updating the joint unknowns; Based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation, the positioning result of each cooperative target is determined.

2. The regional navigation enhancement positioning method for UAV swarms according to claim 1, characterized in that, The drone swarm consists of three receiving drones and three relaying drones; and the satellite signal of each visible satellite is received by one receiving drone and broadcast by one relaying drone, forming a mapping relationship between satellite signal reception and broadcasting.

3. The regional navigation enhancement positioning method for UAV swarms according to claim 1, characterized in that, The target formation pose parameters include the rotation matrix and translation vector of the target formation in the global coordinate system, and the second function includes: The global coordinates of each cooperative target are determined based on the known coordinates of each cooperative target in the local coordinate system of the cooperative target formation, the rotation matrix, and the translation vector.

4. The regional navigation enhancement positioning method for UAV swarms according to claim 3, characterized in that, The rotation matrix is ​​represented using Rodrigues vector parameterization.

5. The regional navigation enhancement positioning method for unmanned aerial vehicle (UAV) swarms according to claim 1, characterized in that, The multi-layered observation equation is a combined observation consisting of the first distance observation term from the satellite to the receiving UAV, the information link transmission correction term from the receiving UAV to the relaying UAV, and the second distance observation term from the relaying UAV to the cooperative target.

6. The regional navigation enhancement positioning method for unmanned aerial vehicle (UAV) swarms according to claim 1, characterized in that, The objective function is solved by alternating optimization, including: With the aerial formation pose parameters fixed, solve the target formation pose parameter subproblem with boundary constraints, and update the target formation pose parameters. The target formation pose parameters are fixed and updated. The relay formation pose parameter subproblem with regularization term is solved to update the air formation pose parameters. The target formation pose parameters and the aerial formation pose parameters are alternately updated until the convergence condition is met, and the target formation pose estimate is obtained.

7. The regional navigation enhancement positioning method for unmanned aerial vehicle (UAV) swarms according to any one of claims 1-6, characterized in that, Before establishing the objective function, the following is also included: A multi-starting-point initialization strategy is adopted to generate multiple sets of initial values ​​for the aerial formation pose parameters and the target formation pose parameters. The alternating optimization solution is then performed on each set, and the solution with the smallest residual is selected as the target formation pose estimate.

8. A regional navigation enhancement and positioning device for a drone swarm, characterized in that, The device includes: The first acquisition module is used to acquire the first observation data, which is obtained by receiving satellite signals from the receiving drone in the drone cluster; The second acquisition module is used to acquire second observation data. The second observation data is based on the first observation data and is transmitted through the information link within the UAV cluster. It is then broadcast by the forwarding UAV to the cooperative target formation to be located and collected by the cooperative target formation. The receiving UAV and the forwarding UAV are different UAV nodes in the UAV cluster. The cooperative target formation is a rigid formation composed of multiple cooperative targets with known fixed geometric configurations. The first characterization module is used to characterize the positions of the receiving drone and each of the forwarding drones as a first function with the aerial formation pose parameters as independent variables, based on the known formation topology of the drone swarm. The second characterization module is used to characterize the position of each cooperative target as a second function with the target formation pose parameters as independent variables, based on the known fixed geometric configuration of the cooperative target formation. The equation construction module is used to construct a multi-layered observation equation from the satellite to the cooperative target via the receiving drone and the relaying drone, based on the satellite position, the first function, the second function, the first observation data, and the second observation data. The objective function construction module is used to establish a nonlinear least squares objective function based on the multi-layer observation equation, using the aerial formation pose parameters and the target formation pose parameters as joint unknowns. The optimization solution module is used to perform alternating optimization solutions on the objective function and obtain the target formation pose estimate by iteratively updating the joint unknowns. The positioning determination module is used to determine the positioning result of each cooperative target based on the target formation pose estimation value and the known fixed geometric configuration of each cooperative target in the cooperative target formation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images