A dynamic target high-precision real-time grid positioning method and device for a UAV

Abstract

The present application belongs to the field of unmanned aerial vehicle navigation and positioning technology, and provides a kind of unmanned aerial vehicle to dynamic target high-precision real-time grid positioning method and device, belongs to the field of unmanned aerial vehicle navigation and positioning technology.Method includes: constructing multi-level earth space three-dimensional grid model and initialization;Unmanned aerial vehicle initial position is mapped into grid model to obtain first grid code;Real-time acquisition unmanned aerial vehicle position, attitude and linear distance to dynamic target, based on space forward intersection principle calculates the theoretical coordinates of dynamic target;Theoretical coordinates are mapped into grid model to obtain second grid code;Whether the grid unit size corresponding to second grid code is greater than precision threshold value is judged, if greater than, then trigger grid refinement, until the accuracy meets the requirements, output second grid code as the real-time positioning result of dynamic target.The present application solves the problem of unmanned aerial vehicle high-precision real-time positioning of dynamic target in complex environment.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) navigation and positioning technology, specifically to a method and apparatus for high-precision real-time grid positioning of dynamic targets by UAVs. Background Technology

[0002] With the widespread application of drone technology in military reconnaissance, disaster monitoring, and traffic management, high-precision real-time positioning of dynamic targets has become a critical requirement. Existing mainstream positioning methods each have their limitations: satellite navigation-based methods (such as GPS and BeiDou) experience a sharp drop in accuracy in environments with severe signal obstruction, such as urban canyons and forests, and have limited update frequencies; visual recognition-based methods are greatly affected by lighting, weather, and background complexity, consume high computational resources, and struggle to guarantee real-time performance; and wireless signal-based methods (such as Wi-Fi and Bluetooth) are susceptible to multipath effects and interference, resulting in large positioning errors and limited effective range.

[0003] Geospatial grid coding technology offers a new approach to geographic information management by dividing the Earth's space into regular grid units. However, there is currently no mature technology to combine this grid coding with UAV positioning of dynamic targets to fully leverage its standardized and structured advantages to improve positioning accuracy and real-time performance. Therefore, there is an urgent need for a new method and device that can overcome the shortcomings of existing technologies. Summary of the Invention

[0004] To achieve the above objectives, according to one aspect of this application, a high-precision real-time positioning method for dynamic targets of unmanned aerial vehicles (UAVs) based on geospatial grid partitioning is provided. This method constructs a multi-level geospatial three-dimensional grid model, maps the UAV's position onto the grid, calculates the theoretical coordinates of the target using the principle of forward intersection in space, and implements a dynamic grid refinement mechanism to locally refine the target area as needed, thereby achieving on-demand high-precision positioning. The specific solution is as follows: A method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles (UAVs), the method comprising the following steps: S1, construct a multi-level three-dimensional mesh model of Earth space and initialize it to the preset level; S2, map the initial spatial position coordinates of the UAV onto the three-dimensional mesh model of the Earth space to obtain the first mesh code, so as to determine the initial position of the UAV in the three-dimensional mesh model of the Earth space; S3 acquires the UAV's own spatial position, attitude information, and straight-line distance to the dynamic target in real time. Based on the principle of forward intersection in space, it calculates the theoretical coordinates of the dynamic target in the Earth coordinate system, thereby obtaining the theoretical coordinates of the dynamic target. S4, map the dynamic target theoretical coordinates calculated in step S3 onto the three-dimensional Earth space mesh model to obtain... Obtain the second grid code; S5, determine whether the size of the grid cell represented by the second grid code is greater than a preset precision threshold: If not, output the second grid code as the real-time positioning result of the dynamic target; If so, the grid adjustment algorithm is triggered to refine the grid where the dynamic target is located, and step S4 is re-executed until the size of the grid cell represented by the second grid code is less than or equal to the accuracy threshold. Then, the refined second grid code is output as the real-time positioning result of the dynamic target.

[0005] Preferably, in S5, the triggered mesh adjustment algorithm includes: initiating a local deeper level of mesh subdivision within the area corresponding to the second mesh code, and re-executing step S4, iteratively executing until the size of the current mesh cell is less than or equal to the accuracy threshold, or reaches the finest mesh level allowed by the system.

[0006] Preferably, it also includes the generation and updating of the trajectory of dynamic targets, specifically: The second target grid codes output from multiple consecutive time periods are concatenated in chronological order to form the motion trajectory of the dynamic target, and a filtering algorithm is used to smooth the motion trajectory of the dynamic target.

[0007] Preferably, the filtering algorithm is a Kalman filter, and its processing procedure is as follows: The target's motion in longitude, latitude, and elevation is modeled as a dynamic system. The calculated theoretical coordinates of the target or the coordinates of the grid center point calculated from the grid code are used as observations. The optimized estimate of the target's position is output through the prediction and update process. This optimized estimate is then mapped back to the Earth's spatial grid to obtain a smoothed second grid code sequence.

[0008] Preferably, in step S3, the execution process of the spatial forward rendezvous principle includes: The attitude angles obtained in the drone's body coordinate system are combined with the drone's own position information and converted into the azimuth and pitch angles from the drone to the dynamic target in the geographic coordinate system. Using the UAV's own position as the origin, the azimuth angle, pitch angle, and straight-line distance determine the ray, and the latitude, longitude, and altitude coordinates of the ray's endpoint in the Earth coordinate system are calculated through coordinate system transformation.

[0009] Preferably, the coordinate system transformation includes: Transform the direction vector in the UAV's body coordinate system to the North-East-Ground coordinate system; Transform the vectors in the North-East-Earth coordinate system to the geocentric map coordinate system; Convert the coordinates from the geocentric map coordinate system to the WGS84 geographic coordinate system.

[0010] Preferably, the method is applied in urban emergency pursuit, military target tracking and guided strikes, intelligent traffic monitoring, or ecological monitoring.

[0011] According to another aspect of this application, a high-precision real-time grid positioning device for dynamic targets by an unmanned aerial vehicle (UAV) is provided, the device comprising: The grid management module is used to build and manage multi-level geospatial grid models; The data acquisition and processing module is used to acquire and process drone sensor data; The positioning and calculation module is used to calculate the target coordinates based on the UAV's position and sensor data and map them to the grid. The trajectory management module is used to store and optimize target grid-coded trajectories.

[0012] Preferably, the mesh management module includes: an initialization unit for loading and constructing a multi-level mesh model, and a unit for positioning... Encoding and decoding units for converting between standard and grid codes; dynamic partitioning units for dynamic deep partitioning of specified regions; The data acquisition and processing module includes: a sensor interface unit and a coordinate transformation unit; the sensor interface unit Used to receive and analyze raw data streams from sensors; the coordinate transformation unit is used for attitude data conversion from the body coordinate system to the geographic coordinate system; The positioning and calculation module includes a forward intersection calculation unit and a grid mapping unit; the forward intersection calculation unit is used to perform spatial forward intersection calculation of the target's theoretical coordinates based on the UAV's pose and ranging data; the grid mapping unit is used to map the target coordinates to the current active grid level; The trajectory management module includes a sequence storage unit and a filtering optimization unit; the sequence storage unit is used to store the target grid codes with timestamps in chronological order, and the filtering optimization unit is used to generate an optimized trajectory by smoothing the grid code sequence through a filtering algorithm.

[0013] Compared with the prior art, the beneficial effects of this application are as follows: This application combines a multi-level three-dimensional geospatial mesh model with the principle of forward intersection in space to achieve target coordinates calculation without relying on external signals, using only the UAV's own position, attitude, and ranging data. It can still output stable positioning results in complex environments, overcoming the accuracy fluctuation problem caused by environmental interference, thus solving the environmental adaptability defects of traditional positioning methods and achieving high-precision positioning of dynamic targets.

[0014] This application employs a dynamic mesh refinement mechanism to perform local deep-level subdivision of the target region based on a preset accuracy threshold, rather than constructing a high-precision mesh globally. This enables dynamic adjustment of accuracy as needed, satisfying the accuracy requirements of different task stages while avoiding resource waste caused by over-computation.

[0015] This application reduces the amount of data computation and storage requirements through local partitioning, supports high-frequency output of positioning results, meets the high-speed motion tracking requirements of dynamic targets, and avoids the problem of insufficient real-time performance.

[0016] In this application, the grid encoding adopts an "integer or string" format, and the data volume is much smaller than the floating-point data of traditional latitude and longitude coordinates. In the collaborative transmission between UAVs and ground stations and multiple systems, it can reduce bandwidth consumption and further ensure the real-time synchronization of positioning results. It is especially suitable for military, emergency and other scenarios that are sensitive to transmission delays, while improving data transmission efficiency. In this application, grid coding is used as a unified spatial language. All systems (UAVs, ground control stations, fire units, and emergency command centers) calculate their positions based on the same grid model. Without coordinate transformation, the geographical range can be obtained directly through the encoding and decoding units. For example, in military scenarios, the response cycle can be compressed to the second level to avoid missing opportunities.

[0017] This application uses a trajectory management module to concatenate continuous grid codes over time and combines this with Kalman filtering to smooth the trajectory, forming a structured grid code sequence. Compared to traditional discrete latitude and longitude elevation data, the structured sequence is more convenient for spatiotemporal querying, historical trajectory backtracking, and target motion prediction. Even if the target is temporarily lost (e.g., entering woodland for occlusion), a reliable position estimate can still be output through the filtering optimization unit, solving the problems of "unstructured and difficult prediction" in traditional trajectories.

[0018] This application ensures the accuracy of the target's theoretical coordinate calculation through a "coordinate system transformation chain"; it further corrects position deviations through dynamic mesh refinement iteration; it filters sensor noise through Kalman filtering and outputs an optimized mesh coding sequence. The collaboration of multiple links reduces the impact of errors in a single link on the final result. Furthermore, the conversion between coordinates and mesh codes adopts a standardized mapping function to ensure that the mesh code mapped to the same coordinate at the same level is unique, avoiding position offsets caused by algorithm differences in traditional positioning and improving the consistency and reliability of positioning results. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the process of a high-precision real-time grid positioning method for dynamic targets by an unmanned aerial vehicle (UAV) according to the present invention. Figure 2 This is a schematic diagram illustrating the overall concept of a high-precision real-time grid positioning method for dynamic targets using an unmanned aerial vehicle (UAV) as described in this invention. Figure 3 This is a schematic diagram of the GeoSOT-3D global subdivision in this invention; Figure 4 This is a schematic diagram of the target space coordinate calculation algorithm in this invention; Figure 5 This is a schematic diagram of the dynamic mesh generation process of the present invention; Figure 6 This is a schematic diagram illustrating the implementation process of the present invention in a drone-guided precision strike scenario. Detailed Implementation

[0020] like Figure 1 As shown, a method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles (UAVs) is provided, the method comprising the following steps: S1. Construct a multi-level three-dimensional mesh model of Earth space and initialize it to a preset level. The specific process is as follows: A virtual 3D Earth space grid (such as the GeoSOT-3D grid system) corresponding to the real world is constructed in the computer; it is initialized by subdivision at level 32 or higher, so that the constructed 3D Earth space grid model is a multi-level global discrete grid system from coarse to fine, providing a basic framework for subsequent coordinate mapping.

[0021] S2, map the initial spatial position coordinates of the UAV to the corresponding hierarchical unit in the Earth space 3D mesh model to obtain the first mesh code, so as to determine the initial position of the UAV in the virtual Earth space 3D mesh model (i.e., digital clone). The specific process is as follows: The drone uses its onboard sensors, such as GNSS / INS and lidar, to obtain its own position, attitude, and straight-line distance to the target in real time. Based on the UAV's own positioning accuracy, its latitude, longitude, and altitude coordinates are mapped to the corresponding level of the grid model to obtain the first grid code, so as to determine the initial position of the UAV in the virtual three-dimensional grid model of the Earth (i.e., digital clone).

[0022] S3. The UAV's spatial position, attitude information, and straight-line distance to the dynamic target are acquired in real time. Based on the principle of forward intersection in space, the UAV's real-time spatial position is used as the origin of the ray, the azimuth and pitch angles determined by the attitude information are used as the ray direction, and the straight-line distance is used as the ray length. A ray pointing towards the dynamic target is constructed in the three-dimensional mesh model of Earth space, and the theoretical coordinates of the ray's endpoint in the Earth coordinate system are calculated. Specifically: In a virtual three-dimensional Earth space grid model, a spatial ray is simulated starting from the "digital clone" of the drone. The grid cell of the three-dimensional Earth space grid model into which the other end of this spatial ray falls is identified as the real-time location of the dynamic target. By analyzing the geographical range of this cell, the geographical coordinates and elevation of the target can be calculated.

[0023] The calculation process for the target's spatial coordinates is as follows: Based on the principle of forward intersection in space, the theoretical latitude, longitude, and altitude coordinates of the target in the WGS84 coordinate system are calculated through a coordinate system transformation chain (specifically, the aircraft's body coordinate system → NED coordinate system → ECEF coordinate system → WGS84 coordinate system).

[0024] S4, map the dynamic target theoretical coordinates obtained in step S3 onto the three-dimensional mesh model of Earth space to obtain... Second grid encoding; S5, the determination output result based on the judgment result of whether the size of the grid cell represented by the second grid code obtained in S4 is greater than the preset precision threshold.

[0025] Furthermore, in S5, the process for determining the output result is as follows: Determine whether the size of the grid cell represented by the second grid code is greater than the preset accuracy threshold. If not, use the second grid code as the output result of the real-time positioning of the dynamic target. If yes, trigger the grid adjustment basic algorithm to refine the grid where the dynamic target is located, and re-execute step S4 until the coordinate accuracy meets the requirements, and output the second grid code as the output result of the real-time positioning of the dynamic target.

[0026] Furthermore, in S5, the determination that the coordinate accuracy meets the requirements is as follows: within the area corresponding to the second grid encoding, a deeper level of local grid subdivision and remapping process are initiated and iteratively executed until the size of the current grid cell is less than or equal to the accuracy threshold, or the finest grid level allowed by the system is reached.

[0027] This approach, through the introduction of a closed-loop optimization mechanism, enables adaptive refinement of the grid region containing the target when the positioning accuracy does not meet requirements. The ray endpoints are then redefined using a higher-precision grid, thus achieving high-precision positioning of dynamic targets while ensuring real-time performance. The specific implementation process is shown in the attached figure. Figure 2 As shown, the specific process is as follows: First, the Earth's space is initially divided into grids to establish a basic grid framework.

[0028] Then, the drone's position / attitude / ranging data are acquired in real time: the drone's position information, attitude parameters and ranging data are continuously collected to provide raw data for subsequent positioning.

[0029] Next, the drone's real-time position is mapped to cell A in the initial grid to determine the drone's position in the grid.

[0030] By using the spatial forward rendezvous method and combining relevant data from the UAV, the theoretical coordinates of the target are calculated.

[0031] The calculated target theoretical coordinates are mapped to mesh cell B.

[0032] Determine whether the accuracy of the divided grid cells meets the requirements, and assess the accuracy of target positioning: If yes, output the target mesh code and timestamp, and then merge the historical mesh sequence to generate a smooth trajectory; if no: start dynamic mesh subdivision to improve the accuracy of local meshes, and then return to the "Map target coordinates to mesh cell B" step to re-perform the mesh mapping and accuracy judgment of the target coordinates.

[0033] The specific implementation process of the UAV high-precision real-time grid positioning method for dynamic targets is as follows: Step S101, Mesh initialization: As attached Figure 3 The GeoSOT-3D global mesh diagram shown demonstrates how a multi-level three-dimensional Earth space mesh model can be constructed in a computer, such as a GeoSOT-3D-based mesh system, initialized to levels 20 to 32, forming a global discrete mesh system from coarse to fine to provide a basic framework for subsequent coordinate mapping.

[0034] Step S102, Data Acquisition: By using the positioning module (such as GPS), attitude sensor (such as IMU), and ranging module (such as lidar and radar) onboard the drone, it can obtain its own spatial position (latitude, longitude, and altitude), attitude information (yaw angle, pitch angle, and roll angle), and straight-line distance to dynamic targets in real time.

[0035] Step S103, Calculation of target theoretical coordinates: Based on the principle of forward rendezvous in space, a ray pointing towards the target is constructed, with the UAV's position as the origin of the ray, the azimuth and pitch angles determined by the attitude angles as the ray direction, and the straight-line distance as the ray length. Through coordinate system transformation (including body coordinate system → NED coordinate system → ECEF coordinate system → WGS84 coordinate system), the latitude, longitude, and altitude coordinates of the ray's endpoint in the WGS84 coordinate system are calculated. The specific process is as follows: (1) Perform UAV position grid mapping. The specific process is as follows: The real-time latitude, longitude and altitude coordinates of the UAV are obtained and mapped to the corresponding hierarchical unit in the initialized three-dimensional geospace grid model according to its positioning accuracy (e.g., RTK provides centimeter-level accuracy). For example, the coordinates of the UAV are mapped to a level 30 grid unit, and the unique coded grid of this grid unit in the three-dimensional geospace grid model is recorded as the first grid code (grid code A).

[0036] (2) Mapping function, specifically: call the GetCode(lat,lon,height,level) function of the selected grid system (GeoSOT); (3) Determine the mapping level: Determine the mapping level based on the accuracy of the UAV's own positioning. For example, if the UAV uses RTK positioning with an accuracy of centimeters, it can be directly mapped to a grid of level 30 or finer.

[0037] (4) The algorithm based on the level is: its principle is that the target level is equal to min (the maximum available level, the level corresponding to the UAV positioning accuracy).

[0038] Finally, based on the determined algorithm, a grid code B of integer or string type is returned through the selected mapping function, such as "G203E152N3032...".

[0039] Based on the principle of forward intersection in space, the instantaneous absolute coordinates (theoretical coordinates) of a dynamic target in a three-dimensional mesh model of Earth space are calculated. The algorithm principle is as follows: Figure 4 As shown: (1) The attitude angles obtained by the UAV IMU in the body coordinate system are combined with the UAV's own position information and transformed into the North-East-Earth (NED) geographic coordinate system to obtain the true north azimuth and pitch angles from the UAV to the target.

[0040] Coordinate system transformation chain: Body coordinate system → North-East-Earth (NED) coordinate system → Geocentric map coordinate system (ECEF) → WGS84 geographic coordinate system.

[0041] (2) Using the UAV's own position (as the origin of the ray), azimuth and pitch angles (to determine the direction of the ray), and laser range measurement values ​​(to determine the length of the ray), a ray pointing towards the target is constructed in three-dimensional space.

[0042] Direction vector calculation: In the body coordinate system, the direction vector of the laser beam is V. b =[R,0,0] indicates that the laser was emitted a distance R along the X-axis of the machine.

[0043] Using the azimuth (ψ), pitch (θ), and roll (φ) provided by the IMU, construct the rotation matrix R_b^n of the aircraft to the NED coordinate system.

[0044] After applying the rotation matrix in the NED coordinate system, the vector V in the NED coordinate system is obtained. n =[N,E,D]=R_b^n·V b .

[0045] This V n The vector contains the target's northward, eastward, and groundward displacements relative to the UAV. The azimuth (α) and pitch (β) angles are derived from this vector V. n The result of the reverse calculation: Azimuth α = arctan(E / N) The azimuth angle β = arcsin(-D / R) (since the D axis is positive downwards, it is negative).

[0046] (3) The latitude, longitude and altitude coordinates of the ray terminus in the WGS84 Earth coordinate system are obtained by geometric calculation.

[0047] Target absolute coordinates calculation: The WGS84 coordinates of the drone (lat) u ,lon u ,alt u Convert to geocentric map coordinate system (X) u ,Y u Z u ).

[0048] In the NED coordinate system, the target's position is: [N,E,D]=[R•cos(β)•cos(α),R•cos(β)•sin(α),-R•sin(β)], (D-axis is positive downwards, so the elevation is negative).

[0049] Using the NED to ECEF transformation matrix, transform [N,E,D] to the ECEF coordinate system to obtain the target's ECEF coordinates (X). t ,Y t Z t ).

[0050] Finally, (X) t ,Y t Z t ) Reverse calculate to WGS84 geographic coordinates (lat t ,lon t ,alt t ).

[0051] Step S104, Target Mesh Mapping: Map the target theoretical coordinates to the current level of the mesh model to obtain the second mesh encoding. Specifically, step S104... The calculated theoretical coordinates of the target are mapped onto a geospatial grid cell adapted to the current accuracy to obtain the real-time position second grid code of the dynamic target (denoted as grid code B), and together with the timestamp, are output as the basic positioning result for this period.

[0052] Call GetCode(lat) t ,lon t ,alt t level current )function; levell here current The initial values ​​are the same as the levels in S103; Output grid code B and high-precision timestamp to form a (timestamp, GridCode) strip. B (Basic positioning records)

[0053] Step S105, Accuracy judgment and mesh refinement: Determine if the grid cell size corresponding to the second grid code is greater than a preset precision threshold. If it is, then in the current grid... A deeper level of subdivision is initiated within the region, the target coordinates are remapped, and the iteration continues until the accuracy requirements are met. Specifically: To achieve high-precision positioning, a dynamic grid partitioning mechanism is introduced during the final confirmation of the second grid encoding. When the grid size exceeds the accuracy threshold, the current grid is partitioned into quadtrees to generate finer-level subgrids, and the target coordinates are remapped. Specifically: The system determines whether the accuracy of the current target's grid meets the preset requirements. If not, the system automatically initiates a finer-level local mesh subdivision (e.g., from level 30 to level 31) within the area represented by the second mesh code (i.e., mesh code B). Subsequently, the target coordinates are mapped again to the newly generated, finer mesh, and a new, more accurate mesh code is output.

[0054] The process for determining the accuracy of the grid where the current target is located is as follows: a preset accuracy threshold P is used. threshold (For example, 0.5 meters). Calculate the size S of the grid represented by the current grid code B. current (For example, the side length of a level 30 grid is approximately 0.9 meters).

[0055] Decision: If S current >P threshold This triggers dynamic partitioning.

[0056] Mesh generation: Instead of actually generating all the fine meshes globally, it dynamically and in memory performs the next level of mesh generation on the region represented by mesh code B.

[0057] As attached Figure 5 As shown, the level 30 grid "3221..." is divided into four level 31 subgrids: "32210", "32211", "32212", and "32213".

[0058] Remapping: Remapping the target coordinates (lat) calculated in S104 t ,lon t ,alt t Then, the mapping operation of S105 is executed again, but this time the level parameter of the mapping is level_current+1.

[0059] Iteration: Repeat steps 1-4 until S current ≤P threshold Or reach the finest level allowed by the system (such as level 32).

[0060] This enables the system to achieve high-precision positioning of the area of ​​interest without increasing the overall computational load.

[0061] Step S106, Trajectory Generation and Optimization: The second grid codes with continuous time periods are concatenated in chronological order to form the target motion trajectory, and Kalman filtering is then applied. The trajectory is smoothed, and the optimized grid-coded sequence is output.

[0062] The second grid codes (B1, B2, B3, ...) of the target across multiple consecutive time periods are concatenated in chronological order to form the target's trajectory. A Kalman filter is then used to smooth this grid sequence, reducing the impact of sensor noise.

[0063] Track storage: Use a first-in-first-out (FIFO) queue or time-series database to store a recent series of timestamps, grid codes, etc. B ,(lat t ,lon t ,alt t )).

[0064] Filtering algorithm: Kalman filtering: Models the motion of a target in each dimension (longitude, latitude, elevation) as a dynamic system (such as a uniform velocity model CV or a uniform acceleration model CA). The state vector is [position, velocity, (acceleration)].

[0065] Observations: The target coordinates calculated from S104 (lat)t ,lon t ,alt t Alternatively, the coordinates of the grid center point calculated from the grid code B can be used as the observation value.

[0066] Prediction and Update: The filter predicts the target position at the current time step based on the state at the previous time step, and then performs a weighted fusion with the current time step's observation to output an optimal estimate. This estimate is smoother than the original observation and can predict the short-term future position.

[0067] Step S107, Grid trajectory output: The optimal estimated coordinates after S106 filtering are mapped back to the Earth space grid to obtain a smoothed and more reasonable second grid encoding sequence, which serves as the final dynamic target trajectory.

[0068] The high-precision real-time grid positioning method for dynamic targets provided in this application can be applied to urban emergency pursuit, military target tracking and guidance strikes, as well as intelligent traffic monitoring or ecological monitoring.

[0069] A high-precision real-time grid positioning device for dynamic targets on unmanned aerial vehicles (UAVs) is disclosed. This device can be integrated into the UAV flight control system, ground control station, or cloud processing platform. The device includes: 1. Grid Management Module Initialization Unit: Responsible for loading and building a multi-level geospatial grid model when the system starts up.

[0070] Encoding and Decoding Unit: Provides functions for converting latitude, longitude, and altitude coordinates into grid codes (forward calculation) and restoring grid codes to geographic extents (reverse calculation).

[0071] Dynamic meshing unit: Responsible for dynamically and at a deeper level meshing of a specified area according to accuracy instructions.

[0072] 2. Data Acquisition and Processing Module Sensor Interface Unit: Responsible for receiving and parsing raw data streams from various sensors of the UAV.

[0073] Coordinate transformation unit: responsible for converting the attitude data of the body coordinate system into the azimuth and pitch angles of the geographic coordinate system.

[0074] 3. Location Solving Module Forward rendezvous calculation unit: The core calculation unit performs spatial forward rendezvous calculations based on the UAV's position, attitude, and ranging data to obtain the target's theoretical coordinates.

[0075] Mesh mapping unit: Responsible for quickly mapping the calculated target coordinates to the currently active mesh level provided by the mesh management module.

[0076] 4. Track Management Module Sequence storage unit: Stores target grid codes with timestamps in chronological order.

[0077] Filtering and optimization unit: Runs the filtering algorithm to smooth the stored grid-coded sequence and generate an optimized target motion trajectory.

[0078] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to explain the present invention and do not constitute a limitation on the scope of protection of the present invention.

[0079] Example This embodiment uses a UAV-guided precision strike scenario targeting time-sensitive maneuvering targets as an example. Specific implementation processes of the method and apparatus described in this invention are explained.

[0080] On the modern battlefield, high-value targets such as enemy mobile rocket artillery and air defense systems often utilize their high mobility to carry out "hit-and-run" tactics in complex environments (such as open fields and woodlands). This poses a severe challenge to the speed and accuracy of the closed-loop response from reconnaissance to strike.

[0081] The existing technology mainly suffers from the following pain points: Inconsistent information formats lead to low collaboration efficiency: the target positions transmitted by the forward UAV reconnaissance units are in latitude, longitude, and altitude coordinates, while the artillery units and command systems at the rear may use different coordinate systems (such as UTM or local grids). Multi-source heterogeneous data requires frequent conversion before fusion, which not only introduces errors but also causes a delay of up to several minutes in the critical "sensor-to-shooter" link, making it extremely easy to miss opportunities.

[0082] Fixed positioning accuracy is unsuitable for different mission phases: Traditional positioning methods output coordinates with fixed precision. In the wide-area search phase, excessively high precision data increases unnecessary communication bandwidth burden; while in the precision strike phase, insufficient precision may lead to mission failure. Current technology lacks the ability to dynamically adjust positioning accuracy according to mission phases.

[0083] Target trajectory management is unstructured and prediction is difficult: continuous target locations exist as a series of discrete, unstructured latitude and longitude points, making efficient spatial querying and analysis difficult. Especially when the target is briefly lost, trajectory smoothing and prediction calculations are complex and unreliable.

[0084] The implementation of this embodiment is a progressive process from "coarse" to "refined," from "perception" to "action," and its process can be found in the appendix. Figure 6 The device is integrated into the mission system of a tactical unmanned aerial vehicle, and the specific execution steps are as follows: 1) Gridded wide-area search and initial target indication: The high-altitude long-endurance UAV performs a wide-range scan of the mission area, and the grid management module of the device has been initialized into a multi-level three-dimensional Earth space grid model. The drone's own location and scanning area are managed and indexed using a Level 25 (100-meter level) geospatial grid.

[0085] Upon detecting a suspected target, the system initially delineates its location within one or more grids of level 25. The data acquisition and processing module and the positioning calculation module in the device immediately generate the grid code, and distribute the grid code as preliminary intelligence to the tactical UAV responsible for precision tracking.

[0086] 2) Refined tracking and dynamic positioning: Once the tactical drone arrives over the target area, high-precision sensors (such as lidar and IMU) are used.

[0087] The data acquisition and processing module in the device acquires the precise position and attitude information of the UAV itself, as well as the straight-line distance to the dynamic target in real time.

[0088] Its positioning and calculation module first maps the UAV's high-precision position to the first grid code at level 30 (approximately meter-level). Then, the forward intersection calculation unit of the positioning and calculation module, based on the principle of spatial forward intersection, calculates the theoretical coordinates of the target in the WGS84 coordinate system in real time through spatial forward intersection. The grid mapping unit then maps these theoretical coordinates to the same precision level (level 30), obtaining the target's second grid code, and outputs it at a high frequency (e.g., 10 times per second), forming the initial trajectory point sequence of the target's motion.

[0089] When the target enters woodland or other areas requiring precise identification... The system's preset accuracy threshold is triggered. Since the current level 30 grid cell size is larger than this threshold, the dynamic subdivision unit in the grid management module of the device is invoked to perform dynamic grid refinement in the local area where the target is located (e.g., from level 30 to level 31 or 32). At this time, the positioning and solving module remaps the target's theoretical coordinates to the newly generated, finer grid, and finally outputs a second grid code that meets the accuracy requirements (e.g., level 32, centimeter level), thereby clearly identifying the precise location of the target vehicle in the complex environment.

[0090] 3) Structured intelligence distribution and fire guidance: The trajectory management module in the device receives a continuous second grid-encoded sequence. Its filtering and optimization unit uses a Kalman filter algorithm to smooth the sequence, generating a stable and optimized target motion trajectory.

[0091] The drone will send the final high-precision second grid code and its smoothed trajectory sequence directly to the rear fire unit (such as rocket artillery unit, loitering munition) via data link.

[0092] The fire control system of the rear fire unit incorporates the same mesh model as the UAV. Therefore, the received mesh code can be directly decoded into firing parameters or guidance commands by the encoding and decoding unit of the mesh management module without any coordinate transformation. The same mesh code can be distributed in parallel to multiple attack units, enabling coordinated attacks based on a unified spatial reference.

[0093] This embodiment achieves a shift from a "chain-like process" to "networked collaboration." By using the "Earth space grid" as a universal spatial language, reconnaissance, command, and strike units can achieve "plug-and-play" and "second-level synchronization" of information, greatly compressing the OODA (Observe-Adjust-Decision-Action) cycle from detection to strike. It also achieves a shift from "fixed precision" to "on-demand precision": through a dynamic grid partitioning mechanism, the grid level of positioning output can be intelligently adjusted according to the mission phase (search, track, strike) and environmental requirements, achieving globally optimal allocation of computing resources, communication bandwidth, and combat effectiveness.

[0094] It has achieved a transformation from "data piling up" to "knowledge generation": by outputting a structured grid-coded sequence, this data is inherently suitable for efficient spatiotemporal query and analysis, making advanced applications such as target behavior prediction and intelligent path planning possible, and improving the level of intelligence in command and decision-making.

[0095] In summary, this embodiment fully demonstrates that by introducing the Earth spatial grid as a unified spatial information "base," the present invention not only solves the pain points of existing positioning methods technically, but also optimizes the reconnaissance-strike process at the system level, significantly improving the ability to perform high-precision, real-time positioning and coordinated strikes against dynamic targets in complex environments.

Claims

1. A method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles (UAVs), characterized in that, The method includes the following steps: S1, construct a multi-level three-dimensional mesh model of Earth space and initialize it to the preset level; S2, map the initial spatial position coordinates of the UAV onto the three-dimensional mesh model of the Earth space to obtain the first mesh code, so as to determine the initial position of the UAV in the three-dimensional mesh model of the Earth space; S3 acquires the UAV's own spatial position, attitude information, and straight-line distance to the dynamic target in real time. Based on the principle of spatial forward intersection, it calculates the theoretical coordinates of the dynamic target in the Earth coordinate system. Specifically, it converts the UAV's body coordinate system into the North-East-Earth coordinate system, the geocentric map coordinate system, and the WGS84 geographic coordinate system in sequence. S4, map the dynamic target theoretical coordinates calculated in step S3 onto the three-dimensional Earth space mesh model to obtain... Obtain the second grid code; S5, determine whether the size of the grid cell represented by the second grid code is greater than a preset precision threshold: If not, output the second grid code as the real-time positioning result of the dynamic target; If so, the grid adjustment algorithm is triggered to refine the grid where the dynamic target is located, and step S4 is re-executed until the size of the grid cell represented by the second grid code is less than or equal to the accuracy threshold. Then, the refined second grid code is output as the real-time positioning result of the dynamic target.

2. The method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles according to claim 1, characterized in that, In S5, the triggered mesh adjustment algorithm includes: initiating a local deeper level mesh subdivision within the area corresponding to the second mesh encoding, and re-executing step S4, iteratively executing until the size of the current mesh cell is less than or equal to the accuracy threshold, or reaches the finest mesh level allowed by the system.

3. The method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles according to claim 1, characterized in that, It also includes the trajectory generation and updating of dynamic targets, specifically: The second grid codes output from multiple consecutive time periods are concatenated in chronological order to form the motion trajectory of the dynamic target, and a filtering algorithm is used to smooth the motion trajectory of the dynamic target.

4. The method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles according to claim 3, characterized in that, The filtering algorithm is a Kalman filter, and its processing procedure is as follows: The target's motion in longitude, latitude, and elevation is modeled as a dynamic system. The calculated theoretical coordinates of the target or the coordinates of the grid center point calculated from the grid code are used as observations. The optimized estimate of the target's position is output through the prediction and update process. This optimized estimate is then mapped back to the Earth's spatial grid to obtain a smoothed second grid code sequence.

5. The method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles according to claim 1, characterized in that, In step S3, the execution process of the spatial forward rendezvous principle includes: The attitude angles obtained in the drone's body coordinate system are combined with the drone's own position information and converted into the azimuth and pitch angles from the drone to the dynamic target in the geographic coordinate system. Using the UAV's own position as the origin, the azimuth angle, pitch angle, and straight-line distance determine the ray, and the latitude, longitude, and altitude coordinates of the ray's endpoint in the Earth coordinate system are calculated through coordinate system transformation.

6. A method for high-precision real-time grid positioning of dynamic targets by unmanned aerial vehicles (UAVs) according to claim 5, characterized in that, The coordinate system transformation includes: Transform the direction vector in the UAV's body coordinate system to the North-East-Ground coordinate system; Transform the vectors in the North-East-Earth coordinate system to the geocentric map coordinate system; Convert the coordinates from the geocentric map coordinate system to the WGS84 geographic coordinate system.

7. A method for high-precision real-time grid positioning of dynamic targets by an unmanned aerial vehicle (UAV) according to claim 1, characterized in that, The method described has applications in urban emergency pursuit and capture, military target tracking and guided strikes, intelligent traffic monitoring, or ecological monitoring.

8. An apparatus employing the high-precision real-time grid positioning method for dynamic targets by an unmanned aerial vehicle (UAV) according to any one of claims 1 to 7, characterized in that, The device includes: The grid management module is used to build and manage multi-level geospatial grid models; The data acquisition and processing module is used to acquire and process drone sensor data; The positioning and calculation module is used to calculate the target coordinates based on the UAV's position and sensor data and map them to the grid. The trajectory management module is used to store and optimize target grid-coded trajectories.

9. A high-precision real-time grid positioning device for dynamic targets by an unmanned aerial vehicle (UAV) according to claim 8, characterized in that, The grid management module includes: an initialization unit for loading and constructing a multi-level grid model, an encoding and decoding unit for converting coordinates to grid codes, and a dynamic subdivision unit for dynamic deep subdivision of a specified region. The data acquisition and processing module includes a sensor interface unit and a coordinate transformation unit; the sensor interface unit is used to receive and parse the raw data stream from the sensor; the coordinate transformation unit is used to convert the attitude data from the body coordinate system to the geographic coordinate system. The positioning and calculation module includes a forward intersection calculation unit and a grid mapping unit; the forward intersection calculation unit is used to perform spatial forward intersection calculation of the target's theoretical coordinates based on the UAV's pose and ranging data; the grid mapping unit is used to map the target coordinates to the current active grid level; The trajectory management module includes a sequence storage unit and a filtering optimization unit; the sequence storage unit is used to store the target grid codes with timestamps in chronological order; the filtering optimization unit is used to generate an optimized trajectory by smoothing the grid code sequence through a filtering algorithm.

Images