Unmanned aerial vehicle take-off and landing guide radar data correction method based on error map
By using an error map and Kalman filtering method, the navigation information of the UAV take-off and landing guidance radar is corrected in real time, which solves the problem of insufficient radar accuracy under satellite navigation interference and improves the safety and reliability of UAV take-off and landing.
Patent Information
- Application Number
- CN202511841675.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-09
- Publication Date
- 2026-03-20
AI Technical Summary
Under conditions of satellite navigation interference, traditional UAV take-off and landing guidance radars have low accuracy and are easily affected by obstruction, making it difficult to meet the take-off and landing safety requirements of high-value UAVs.
A radar data correction method for UAV take-off and landing guidance based on error maps is adopted. Combining Kalman filtering and aircraft kinematic models, radar error maps are obtained through multiple flight tests, and navigation information is corrected in real time using Kalman filters to improve radar tracking accuracy.
It improves the accuracy and safety of UAV take-off and landing guidance radar, ensures the reliability of UAV take-off and landing in complex environments, and meets the safety requirements of high-value UAVs.
Smart Images
Figure CN121703769A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of unmanned aerial vehicle landing guidance, and particularly relates to an unmanned aerial vehicle landing guidance radar data correction method based on an error map. BACKGROUND
[0002] With the continuous increase of the number of unmanned aerial vehicle equipment, as an important weapon on the battlefield, the number of unmanned aerial vehicle test flights increases exponentially, and the traditional unmanned aerial vehicle landing guidance method based on differential satellite navigation has more and more variable satellite interference and more and more security risks. In order to further improve the safety of unmanned aerial vehicle landing, landing guidance radar as a navigation method independent of satellite becomes a choice. Due to the low guidance accuracy of landing guidance radar, the use scene is harsh, such as no occlusion of the line of sight of the unmanned aerial vehicle, and other factors, the landing guidance radar is rarely applied to unmanned aerial vehicle landing. In order to improve the guidance accuracy and reduce the error caused by the traditional radar data processing using a simple filtering algorithm, the unmanned aerial vehicle landing guidance radar tracking accuracy data analysis and correction method based on the error map is adopted, and the Kalman filtering method is used in combination with the aircraft kinematic model to offset the system error of the unmanned aerial vehicle landing guidance radar and the external interference error caused by the unavoidable occlusion at the fixed position. SUMMARY
[0003] The technical problem solved by the application is to solve the problem of using landing guidance radar alone to complete the landing guidance of high-value unmanned aerial vehicles under the condition of satellite navigation interference. The application designs an unmanned aerial vehicle landing guidance radar data correction method based on an error map, obtains a time-domain and space-domain error map of the landing guidance radar based on multiple analysis results, and corrects the navigation information output by the landing guidance radar by using Kalman filtering in combination with an aircraft model, thereby improving the tracking accuracy of the landing guidance radar and ensuring that the single guidance mode of the landing guidance radar meets the safety requirements of high-value unmanned aerial vehicle landing, and improving the safety of unmanned aerial vehicle test flights.
[0004] The technical scheme of the application is as follows: An unmanned aerial vehicle landing guidance radar data correction method based on an error map, the method comprising: S1, acquiring reference data stream of an airborne RTK / INS integrated navigation system; acquiring unmanned aerial vehicle landing guidance radar data stream; and thereby acquiring error set of the reference data stream and the radar data stream; S2, establishing an error map according to the error set; and generating three-dimensional error distribution map of the unmanned aerial vehicle landing guidance radar in the airport airspace according to the error map; S3, integrating the three-dimensional error distribution map into an adaptive filter based on an aircraft kinematic model as a lookup table, obtaining an expected system bias in real time according to the current estimated position of the aircraft, and compensating in the update step.
[0005] Furthermore, In S1, a test UAV equipped with a high-precision RTK positioning system, an inertial navigation system (INS), a takeoff and landing guidance radar transponder, and an airborne data recorder is selected and conducts multiple standard takeoff and landing flight tests at the airport; a UAV takeoff and landing guidance radar to be calibrated is deployed on the ground and does not participate in the aircraft navigation control; during the flight test, the aircraft records the time-aligned reference data stream and radar data stream, including time, latitude and longitude coordinates; Baseline data stream , Radar data stream .
[0006] Furthermore, in S1, the error set between the reference data stream and the radar data stream is obtained: .
[0007] Furthermore, S2 specifically includes: S21 divides the airspace involved in aircraft takeoff and landing into discrete three-dimensional grids to form a data structure for an error map. Each grid is used to store the statistical results of all error data falling into that space. S22, Fill each grid with data according to the error set to obtain the error map; S23, perform post-smoothing processing on the error map; S24. Generate a three-dimensional error distribution map of the UAV take-off and landing guidance radar in the airport airspace based on the error map.
[0008] Furthermore, S21 divides the airspace involved in aircraft takeoff and landing into a discrete three-dimensional grid, specifically: Define the grid range: covering the airspace from the start to the finish line of the runway, and from the ground to a preset height; Define the grid resolution, including the resolution in the east-north direction and the resolution in the height direction; each grid Represents a small spatial unit, the coordinates of which are: .
[0009] Furthermore, S22 specifically includes: For each time point t, according to The specific grid where the location point is located ;in
[0010]
[0011]
[0012] in It is the origin of the airspace. This is the grid size, and floor is the floor function. Instantaneous error Accumulated to this grid From the statistical information, the following data was obtained: : The counter of valid data points falling into this grid; The sum of all eastward errors within this grid; The sum of all northward errors within this grid; The sum of all astronomical errors within this grid; Calculate the average error for each grid cell to create an error map: The average error includes:
[0013]
[0014] .
[0015] Furthermore, S23 specifically refers to: (a) Validity filtering: Filtering by the number of data points Grids with values below a certain threshold are marked as "invalid," and their deviation values are unreliable. (b) Smoothing filter: Apply a mean filter to the generated error map to eliminate statistical noise.
[0016] Furthermore, S3 specifically refers to: S31, State Prediction: Based on the takeoff and landing dynamics model, predict the aircraft's state at time k. ; S32, Calculate expected observations and biases: based on state predictions. The position components in the coordinate system are transformed to the airport coordinate system, and the prediction deviation at that location is obtained by querying the error map. Based on the predicted state values, using a standard observation model... The expected observation values of the error-free radar are calculated. S33, Construct a biased observation model: Establishing the standard observation equation is , where v is the observation noise, X is the system state equation, and H is the state transition matrix; Set a time-varying bias term b(X), where b(X) is the prediction bias found in the error map. Then the observation model with bias is: Z_radar_raw=H·X+b(X)+v; After processing by the adaptive filter described above, real-time guidance data is output for final takeoff and landing guidance.
[0017] The beneficial effects of this invention are as follows: A performance modeling and analysis method for UAV take-off and landing guidance radar based on error map was designed. It supports the marking of horizontal position deviation, altitude deviation, entry and exit points and obstruction points. The accuracy error of take-off and landing guidance radar under fixed deployment is obtained in a statistical manner. Furthermore, the performance model of each UAV take-off and landing guidance radar guidance device is obtained for subsequent data correction input.
[0018] A method for correcting UAV take-off and landing guidance radar data based on Kalman filtering algorithm is designed. The aircraft position information output by the aircraft kinematic model is used as the prediction step, and the position information output by the UAV take-off and landing guidance radar after correction based on the error map is used as the update step. More accurate navigation information is obtained through Kalman filtering, thereby improving the tracking accuracy of UAV take-off and landing guidance radar. Attached Figure Description
[0019] Figure 1 This is a flowchart illustrating a method for correcting radar data for UAV takeoff and landing guidance based on an error map. Detailed Implementation
[0020] The following is in conjunction with the appendix Figure 1 This document provides a detailed description of a preferred embodiment of the present invention to make the objectives, technical solutions, and advantages of the invention clearer. Those skilled in the art should understand that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of protection of the invention.
[0021] This embodiment provides a complete implementation process for a UAV takeoff and landing guidance radar guidance data correction system based on flight data fusion and model correction, such as... Figure 1 As shown, the main steps include data acquisition and preprocessing, performance modeling and analysis of UAV take-off and landing guidance radar, and design and online correction of Kalman filters based on aircraft kinematics models.
[0022] 1. Data Acquisition and Preprocessing The main function of the data acquisition and preprocessing steps is to obtain raw data for modeling and correction.
[0023] Data Acquisition: A test UAV equipped with a high-precision RTK positioning system, inertial navigation system (INS), takeoff and landing guidance radar transponder, and onboard data recorder was selected and conducted multiple standard takeoff and landing flight tests at the airport. Simultaneously, a UAV takeoff and landing guidance radar to be calibrated was deployed on the ground, but it was not initially involved in aircraft navigation control. During the flight tests, the following time-aligned data streams were recorded synchronously: a. Baseline Data Stream: Records high-precision data from the airborne RTK / INS integrated navigation system, including time, latitude, longitude, and altitude coordinates. The standard data stream is recorded as follows:
[0024] b. UAV takeoff and landing guidance radar data stream: Records navigation data from the UAV takeoff and landing guidance radar, including time, target position (distance, azimuth), and altitude information.
[0025] c. Obtain the error set as
[0026] 2. Performance Modeling and Analysis of UAV Takeoff and Landing Guidance Radar The performance modeling and data analysis steps of the UAV take-off and landing guidance radar are based on RTK data, and the error characteristics of the UAV take-off and landing guidance radar are quantified by establishing an error map.
[0027] 2.1 Map Grid Definition The airspace involved in aircraft takeoff and landing is divided into discrete three-dimensional grids to form an error map data structure. Each grid will be used to store the statistical results of all error data falling into that space.
[0028] Define the grid extent: covering the airspace from the start to the finish line of the runway, and from the ground to a certain altitude.
[0029] Define the grid resolution as follows: one grid every 10 meters in the east-north direction, and one grid every 5 meters in the vertical direction. (The interval can be further increased or decreased according to actual needs.) Each grid Represents a small spatial unit, the coordinates of which are: .
[0030] 2.2 Grid Filling and Data Association For each time point t, according to The specific grid where the location point is located .in
[0031]
[0032]
[0033] in It is the origin of the airspace. It is the grid size, and floor is the floor function.
[0034] Instantaneous error Accumulated to this grid From the statistical information, the following data was obtained: : The counter of valid data points falling into this grid.
[0035] : The sum of all eastward errors within this grid.
[0036] : The sum of all northward errors within this grid.
[0037] : The sum of all celestial errors within this grid.
[0038] 2.3 Error Map Calculation Calculate the average error (i.e., error map) and the error uncertainty (standard deviation, used as input for the subsequent Kalman filter uncertainty factor) for each grid: Calculate the average error
[0039]
[0040]
[0041] Calculate the standard deviation
[0042] 2.4 Post-smoothing Due to limitations in flight frequency and flight routes, the trajectory cannot uniformly cover each grid, resulting in some grids having fewer data points and thus unreliable statistics. Therefore, post-processing is required. a) Validity filtering: Filtering by the number of data points. Grids with fewer than a certain threshold (e.g., 5) are marked as "invalid" and their deviation values are unreliable.
[0043] b) Smoothing filter: Apply a mean filter to the generated error map to eliminate statistical noise and make the error field change smoother and more in line with physical reality (because radar error changes continuously in space).
[0044] 2.5 Time Domain and Spatial Domain Error Analysis Compare the UAV takeoff and landing guidance radar data at the same moment with RTK reference data, and calculate the deviations of various parameters (such as position deviation ΔP and altitude deviation ΔV). Plot the curves of deviation changes over time to qualitatively analyze the temporal correlation of radar error throughout the entire takeoff and landing process.
[0045] All data points are mapped to the airport's three-dimensional coordinate system with the runway threshold as the origin. For each small grid in space, the average deviation between all radar data points falling within that grid and the RTK reference data is calculated, thereby generating a three-dimensional error distribution map of the UAV takeoff and landing guidance radar in the airport airspace. This map can clearly identify which areas (such as near a hangar or terrain undulations) have fixed, systematic errors caused by multipath effects or obstruction.
[0046] 3 Design and Online Correction of Adaptive Filters Based on Aircraft Kinematics Model This step integrates the three-dimensional error distribution map generated in step two into the adaptive filter based on the aircraft kinematics model as a lookup table. Based on the aircraft's current estimated position, the table is looked up in real time to obtain an expected system deviation, which is then compensated for in the update step.
[0047] A fusion filter based on Error State Kalman Filtering (ESKF) is established using common methods. The filter's state variables include parameters such as aircraft position, velocity, and attitude. The system's "prediction step" is based on a pre-established takeoff and landing dynamics model of the aircraft; the system's "update step" uses real-time observation data from the UAV's takeoff and landing guidance radar. Specific steps: a) State Prediction: Similar to standard Kalman filtering, in the "prediction step," the system predicts the aircraft's state at time k based on the takeoff and landing dynamics model. .
[0048] b) Calculate expected observations and biases: based on state predictions The position components in the coordinate system are transformed to the airport coordinate system, and the prediction deviation for that position is obtained by querying the error map (through interpolation). Based on the predicted state values, using a standard observation model... The expected observation values of the "error-free radar" are calculated.
[0049] c) Construct a biased observation model: Establishing the standard observation equation is (where v is the observation noise).
[0050] Let a time-varying bias term be defined, b(X), where b(X) is the bias found in the error map. The new observation equation becomes: Z_radar_raw=H·X+b(X)+v d) Real-time data output: After processing by the above adaptive filter, the system outputs real-time guidance data that has been precisely corrected, smoothed, and conforms to the laws of aircraft motion, for use in the final takeoff and landing guidance.
[0051] In summary, this invention, through the approach of "offline modeling and online correction," creatively combines radar inherent error calibration, environmental interference suppression, and aircraft motion models, significantly improving the reliability and accuracy of UAV take-off and landing guidance radar guidance data in complex airport environments.
[0052] This invention addresses the issue of offsetting errors caused by the system errors of take-off and landing guidance radar and external interference caused by factors such as obstruction and multipath propagation within its effective range. It innovatively designs a data analysis and correction method for the tracking accuracy of UAV take-off and landing guidance radar based on an error map. Through RTK navigation flight test data acquisition, a set of reference positions and a set of UAV take-off and landing guidance radar errors are obtained. Based on the error map, a performance model of the UAV take-off and landing guidance radar is established. Furthermore, a Kalman filter is established using an aircraft kinematic model to reduce errors caused by radar system and external environmental obstruction and multipath effects, thereby improving the guidance accuracy of the UAV take-off and landing guidance radar. This enables it to meet the requirements of being a backup guidance method for differential take-off and landing guidance equipment, enhancing the safety of UAV take-off and landing in situations without satellite navigation or where navigation is interfered with.
Claims
1. A method for correcting UAV takeoff and landing guidance radar data based on an error map, characterized in that, The method includes: S1, acquire the reference data stream of the airborne RTK / INS integrated navigation system; acquire the UAV take-off and landing guidance radar data stream; thereby acquire the error set between the reference data stream and the radar data stream; S2, Based on the error set, establish an error map; based on the error map, generate a three-dimensional error distribution map of the UAV take-off and landing guidance radar in the airport airspace; S3 integrates the three-dimensional error distribution map as a lookup table into the adaptive filter based on the aircraft kinematics model. Based on the aircraft's current estimated position, the table is looked up in real time to obtain an expected system deviation, which is then compensated for in the update step.
2. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 1, characterized in that, In S1, a test UAV equipped with a high-precision RTK positioning system, an inertial navigation system (INS), a takeoff and landing guidance radar transponder, and an airborne data recorder is selected and conducts multiple standard takeoff and landing flight tests at the airport; a UAV takeoff and landing guidance radar to be calibrated is deployed on the ground and does not participate in the aircraft navigation control; during the flight test, the aircraft records the time-aligned reference data stream and radar data stream, including time, latitude and longitude coordinates; Baseline data stream , Radar data stream .
3. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 2, characterized in that, In S1, the error set between the reference data stream and the radar data stream is obtained: 。 4. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 3, characterized in that, S2 specifically includes: S21 divides the airspace involved in aircraft takeoff and landing into discrete three-dimensional grids to form a data structure for an error map. Each grid is used to store the statistical results of all error data falling into that space. S22, Fill each grid with data according to the error set to obtain the error map; S23, perform post-smoothing processing on the error map; S24. Generate a three-dimensional error distribution map of the UAV take-off and landing guidance radar in the airport airspace based on the error map.
5. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 4, characterized in that, S21 divides the airspace involved in aircraft takeoff and landing into a discrete three-dimensional grid, specifically: Define the grid range: covering the airspace from the start to the finish line of the runway, and from the ground to a preset height; Define the grid resolution, including the resolution in the east-north direction and the resolution in the height direction; each grid Represents a small spatial unit, the coordinates of which are: .
6. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 5, characterized in that, S22 specifically includes: For each time point t, according to The specific grid where the location point is located ;in in It is the origin of the airspace. This is the grid size, and floor is the floor function. Instantaneous error Accumulated to this grid From the statistical information, the following data was obtained: : The counter of valid data points falling into this grid; The sum of all eastward errors within this grid; The sum of all northward errors within this grid; The sum of all astronomical errors within this grid; Calculate the average error for each grid cell to create an error map: The average error includes: 。 7. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 6, characterized in that, S23 specifically refers to: (a) Validity filtering: Filtering by the number of data points Grids with values below a certain threshold are marked as "invalid," and their deviation values are unreliable. (b) Smoothing filter: Apply a mean filter to the generated error map to eliminate statistical noise.
8. The method for correcting UAV takeoff and landing guidance radar data based on an error map according to claim 7, characterized in that, S3 specifically refers to: S31, State Prediction: Based on the takeoff and landing dynamics model, predict the aircraft's state at time k. ; S32, Calculate expected observations and biases: based on state predictions. The position components in the coordinate system are transformed to the airport coordinate system, and the prediction deviation at that location is obtained by querying the error map. Based on the predicted state values, using a standard observation model... The expected observation values of the error-free radar are calculated. S33, Construct a biased observation model: Establishing the standard observation equation is , where v is the observation noise, X is the system state equation, and H is the state transition matrix; Set a time-varying bias term b(X), where b(X) is the prediction bias found in the error map. Then the observation model with bias is: Z_radar_raw=H·X+b(X)+v; After processing by the adaptive filter described above, real-time guidance data is output for final takeoff and landing guidance.