A Multi-Source Fusion-Based Positioning Method for Maritime Unmanned Aerial Vehicles

Abstract

This invention proposes a multi-source fusion-based maritime UAV positioning method, comprising the following steps: Step 1, calculating the UAV navigation state estimate based on the IMU sensor on the UAV and constructing IMU factors; Step 2, performing fault detection on GNSS data based on the GNSS receiver and IMU sensor on the UAV, calculating the GNSS positioning solution, and constructing GNSS factors; Step 3, calculating the UAV position based on the LEO receiver and BA sensor on the UAV and constructing LEO / BA factors; Step 4, constructing a factor graph based on the IMU factors, GNSS factors, and LEO / BA factors, and calculating the weight matrix of each factor using a sliding window; Step 5, performing information fusion based on the weight matrix and solving to obtain the final navigation and positioning result.

Description

Technical Field

[0001] This invention relates to a method for locating unmanned aerial vehicles (UAVs) at sea, and more particularly to a method for locating unmanned aerial vehicles at sea based on multi-source fusion. Background Technology

[0002] This section provides only background information relevant to this disclosure and is not necessarily prior art.

[0003] With the booming development of the marine economy, unmanned aerial vehicles (UAVs), as automated intelligent transport vehicles, have broad application prospects in marine activities. As a core module of UAVs, navigation and positioning provide crucial position, speed, and attitude information for their flight missions. Due to the challenging flight conditions at sea and the lack of temporary landing sites, large-scale maritime UAV applications place stringent demands on the accuracy and reliability of navigation and positioning. Currently, UAV positioning mainly relies on GNSS to provide absolute positioning information, combined with the assistance of other sensors. However, in the marine environment, the abundant and variable water vapor in the troposphere, along with the strong reflection of navigation satellite signals from the sea surface, results in poor GNSS observation data quality, further contributing to the unreliability of UAV positioning.

[0004] GNSS / IMU / visual integrated navigation is the mainstream solution for navigation and positioning of land-based UAVs. It primarily achieves high-precision navigation and positioning information by fusing information from these three sensors to complement each other. GNSS positioning mainly determines the user's absolute position by measuring the distance from satellites to the receiving device and using resection. However, GNSS signals are susceptible to environmental interference, and single-GNSS positioning is discontinuous. Inertial navigation, on the other hand, measures the vehicle's acceleration and angular velocity, integrating to obtain relative motion, and is almost unaffected by the environment, but suffers from error accumulation. Based on the inherent complementary advantages of GNSS and inertial navigation, GNSS / inertial integrated navigation is widely used in applications with lower precision requirements, such as manned and vehicle-mounted systems. However, if the GNSS signal quality is poor for a long period, continuous independent operation of inertial navigation will lead to filter divergence, resulting in unreliable positioning or even failure. Therefore, visual positioning based on environmental perception is also used to assist UAV navigation in urban and other scenarios. At the GNSS / IMU / visual information fusion level, loose combination utilizes multi-source information in the positioning domain to estimate navigation state, which has the advantages of modularity and low complexity. Therefore, UAV navigation generally adopts loose combination strategies based on Extended Kalman Filter (EKF) or factor graph.

[0005] However, the aforementioned prior art has the following drawbacks:

[0006] a) Complex water vapor and multipath effects at sea cause significant errors in GNSS observation data, meaning faulty data exists, and multiple faults occur frequently at the same time. Existing integrated navigation algorithms based on GNSS / IMU / visual multi-source fusion typically use the odd-even vector method for GNSS fault detection, but this method is only applicable to single fault cases, resulting in poor fault detection performance and further contributing to unreliable positioning.

[0007] b) In the marine environment, the sparse and monotonous visual features cause the performance of visual navigation sources to degrade or even become unusable, making it impossible to effectively compensate for inertial navigation errors when the GNSS signal quality is poor. This further leads to the accumulation or even divergence of navigation errors based on GNSS / IMU / visual multi-source fusion.

[0008] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0009] Purpose of the invention: The technical problem to be solved by the present invention is to provide a maritime unmanned aerial vehicle (UAV) positioning method based on multi-sensor multi-source fusion, which addresses the shortcomings of the existing technology.

[0010] To address the aforementioned technical problems, this invention discloses a maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion, comprising the following steps:

[0011] Step 1: Calculate the drone navigation state estimate based on the IMU sensor on the drone. Construct IMU factors;

[0012] Step 2: Based on the GNSS receiver and IMU sensor on the UAV, perform fault detection on the GNSS data and calculate the GNSS positioning solution. Construct GNSS factors;

[0013] Step 3: Calculate the drone's position based on the LEO receiver and BA sensor on the drone. Construct the LEO / BA factor;

[0014] Step 4: Construct a factor graph based on the IMU factor, GNSS factor, and LEO / BA factor, and calculate the weight matrix of each factor using a sliding window.

[0015] Step 5: Based on the weight matrix, perform information fusion and solve to obtain the final navigation and positioning result.

[0016] Furthermore, the construction of the IMU factor described in step 1 includes:

[0017] The specific force and angular velocity of the drone measured by IMU sensors are used to calculate the current force through mechanical programming. Epichronous-based inertial navigation-based estimation of UAV navigation state And it is expressed as an IMU factor as follows:

[0018]

[0019] in, , and These represent the three-dimensional position estimation, velocity estimation, and carrier vector estimation of the UAV, respectively; superscript Represents the matrix transpose operator; subscript and These represent the IMU and the epoch number, respectively.

[0020] Furthermore, the construction of GNSS factors described in step 2 includes:

[0021] Step 2-1: Perform preliminary fault detection, mark the faulty satellites, and obtain the number of GNSS satellites marked as faulty. and the number of visible GNSS satellites Specifically, it includes the following steps:

[0022] Step 2-1-1: Calculate the 3D position estimate based on inertial navigation. GNSS inversion pseudorange , means as follows:

[0023]

[0024] in, Indicates GNSS satellite Location; Indicates GNSS satellite The geometric distance between the inertial-estimated UAV position and the position of the UAV; Represents the speed of light; and These represent the estimated values ​​of receiver clock bias and satellite clock bias, respectively. and These represent the estimated values ​​for tropospheric delay and ionospheric delay, respectively; subscripts. and These represent GNSS and GNSS satellite serial numbers, respectively.

[0025] Step 2-1-2, Velocity estimation based on inertial navigation Calculate GNSS satellites Inversion Doppler frequency shift observations , means as follows:

[0026]

[0027] in, Indicates GNSS satellite The center frequency of the carrier wave; Indicates GNSS satellite The unit direction vector to the receiver; Indicates GNSS satellite The velocity vector; Indicates receiver clock drift;

[0028] Step 2-1-3, Construct GNSS satellites Fault detection statistics , means as follows:

[0029]

[0030] in, and These represent GNSS satellites. Measured pseudorange and measured Doppler frequency shift; and These represent the standard deviations of GNSS pseudorange and Doppler shift observation noise, respectively.

[0031] Step 2-1-4, Set up fault detection statistics threshold ,like Then determine GNSS satellite If the observed values ​​are normal, then mark it as a faulty satellite.

[0032] Step 2-2, Set empirical parameters The number of GNSS satellites marked as faulty and the number of visible GNSS satellites Perform a check; if it does not meet the requirements...

[0033]

[0034] The current epoch is then calculated using the pseudorange of GNSS satellites with normal observations after initial fault detection. GNSS positioning solution based on least squares method As a GNSS factor, otherwise, fault detection based on deconstruction is performed and the current epoch is calculated accordingly. Final GNSS positioning solution , as a GNSS factor.

[0035] Furthermore, step 2-2 involves performing fault detection based on deconstruction and calculating the current epoch accordingly. Final GNSS positioning solution ,include:

[0036] Step 2-2-1: Using the pseudoranges of currently visible GNSS satellites, construct a subset of all possible locationable pseudoranges. , means as follows:

[0037]

[0038] in, The minimum number of pseudoranges included is the number of constellations used plus 3; This represents the total number of pseudorange subsets of all currently locatable GNSS satellites;

[0039] Step 2-2-2: Calculate the localization solution corresponding to each pseudo-range subset. ;

[0040] Step 2-2-3: Traverse each localization solution in three-dimensional space, using the localization solution as the geometric center, and... Construct a sphere with radius, and count the number of other localized solutions that fall within the sphere. ; and use This indicates that the solution falls within the localized solution. The number of other locating solutions within the sphere at the center of the sphere;

[0041] Step 2-2-4, Quantity maximum value The pseudoranges within the corresponding pseudorange subset are marked as normal pseudoranges, and the other pseudoranges are marked as fault pseudoranges. maximum value The corresponding positioning solution is used as the final GNSS positioning solution for the current epoch. , as a GNSS factor.

[0042] Furthermore, the construction of the LEO / BA factor described in step 3 includes:

[0043] Step 3-1, construct the previous epoch UAV navigation state vector based on LEO / BA fusion , means as follows:

[0044]

[0045] in, , and These are the UAV position, velocity, and LEO receiver clock drift based on LEO / BA integrated navigation in the previous epoch;

[0046] Step 3-2: Using currently visible LEO satellites, construct a subset of all measurable Doppler shift observations. , means as follows:

[0047]

[0048] in, The minimum number of Doppler shift observations included is the number of constellations used plus 3; This represents the total number of all measurable Doppler observations in the subset.

[0049] Step 3-3: Calculate the velocity solution for each Doppler frequency shift subset. ;

[0050] Steps 3-4: In three-dimensional space, traverse each velocity solution, using that velocity solution as the geometric center, and... Construct a sphere with radius, and count the number of other velocity solutions falling inside the sphere. ; and use This indicates that the solution falls on the velocity. The number of other velocity solutions within the sphere at the center of the sphere;

[0051] Steps 3-5, Quantity Maximum value The pseudorange within the corresponding subset of Doppler frequency shifts is labeled as normal Doppler frequency shift, and other Doppler frequency shifts are labeled as faulty Doppler frequency shifts; if there is more than one identical maximum value... Then, the pseudo-ranges within the subset of Doppler frequency shifts corresponding to these maximum values ​​are all marked as normal Doppler frequency shifts, and other Doppler frequency shifts are marked as faulty Doppler frequency shifts;

[0052] Steps 3-6: Construct observation vectors based on normal LEO Doppler frequency shift and BA measurements after elevation transformation. , means as follows:

[0053]

[0054] in, They represent LEO satellites. The center frequency of the carrier wave; They represent LEO satellites. Measured Doppler frequency shift; This represents the current measured value of BA after elevation conversion;

[0055] Steps 3-6: Configure the UAV navigation state vector Perform state prediction based on the constant velocity assumption and based on observation vectors After the measurement update, the UAV navigation estimation result based on LEO / BA fusion is obtained for the current epoch. and the location of the drones within. The location of the drone As a LEO / BA factor.

[0056] Furthermore, step 4, calculating the weight matrix of each factor, includes:

[0057] Step 4-1, define the multi-source fusion navigation state vector based on factor graph as follows: , means as follows:

[0058]

[0059] Among them, subscript Representative factor plot; , and These represent the three-dimensional position, velocity, and carrier vector estimates of the UAV after factor graph information fusion;

[0060] Step 4-2, IMU factor As a state transition node, the GNSS factor and LEO / BA factor As an observation correction node, and using the previous epochs. to Historical data from different eras is used as a sliding window, in which The length of the factor graph sliding window is used to construct the UAV navigation state vector for estimating the current epoch. Factor plot;

[0061] Step 4-3, make the drone fly in a straight line. Seconds, by performing least-squares calculations on GNSS data, the initial sliding window sequence is determined. This completes the initialization of the factor graph;

[0062] Step 4-4: Calculate the IMU factor in real time using historical data within the sliding window. GNSS factor and LEO / BA factor Error covariance estimate , and They are represented as follows:

[0063]

[0064]

[0065]

[0066] in, These are IMU factors GNSS factor LEO / BA factor The estimated value of the error covariance;

[0067] Steps 4-5 yield the IMU factor. GNSS factor and LEO / BA factor The weight matrices are respectively , and .

[0068] Furthermore, step 5, which involves fusing information based on the weight matrix and solving to obtain the final navigation and positioning result, includes:

[0069] Step 5-1: Based on historical data within the sliding window, calculate the optimal navigation state of the current UAV using a factor graph, and construct a nonlinear optimization problem, as follows:

[0070]

[0071] in, Indicates the use of and Calculate the IMU edge factor residual; Indicated by is the square of the weighted Euclidean norm of the weight matrix;

[0072] Step 5-2: Solve the nonlinear optimization problem constructed in Step 5-1 to obtain the optimal estimated navigation state vector that minimizes the right-hand side of the equation. Among them, the optimal estimated navigation state vector Includes drone position estimation Speed ​​estimation and attitude estimation This is the final navigation and positioning result.

[0073] Beneficial effects:

[0074] 1. This invention proposes a GNSS fault detection method that combines the measurement and positioning domains with IMU / LEO joint assistance, achieving rapid and accurate detection and elimination of multiple GNSS faults. This solves the problem of poor GNSS fault detection performance in navigation and positioning algorithms based on GNSS / IMU / visual multi-source fusion.

[0075] 2. This invention proposes to construct a sub-filter based on EKF for LEO / BA to provide auxiliary information for IMU error compensation, and to construct a multi-source adaptive fusion factor map of GNSS / IMU / LEO / BA to output high-precision and reliable navigation and positioning information, thus solving the problem of error accumulation and even divergence in the navigation and positioning scheme of maritime UAVs. Attached Figure Description

[0076] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments, and the advantages of the present invention in the above and / or other aspects will become clearer.

[0077] Figure 1 This is a flowchart of the present invention.

[0078] Figure 2 This is a schematic diagram illustrating the positioning errors of the two methods in the horizontal and vertical directions in the embodiment.

[0079] Figure 3 A factor graph for multi-source fusion navigation of UAVs. Detailed Implementation

[0080] This invention provides a "Navigation and Positioning Algorithm for Maritime UAVs Based on GNSS / IMU / LEO / BA Multi-Source Fusion." Addressing the poor GNSS fault detection performance of existing navigation and positioning algorithms based on GNSS / IMU / visual multi-source fusion, this invention proposes a GNSS fault detection method combining measurement and positioning domains with IMU / LEO joint assistance, achieving rapid and accurate detection and elimination of multiple GNSS faults. Furthermore, to address the issue of error accumulation and even divergence in existing maritime UAV navigation and positioning schemes, this invention proposes constructing an EKF-based LEO / BA sub-filter to provide auxiliary information for IMU error compensation, and constructing a GNSS / IMU / LEO / BA multi-source adaptive fusion factor map to output high-precision and reliable navigation and positioning information.

[0081] The core content of this invention is a navigation and positioning method for maritime unmanned aerial vehicles (UAVs), the overall process of which is as follows: Figure 1 As shown. The selected sensors are a GNSS (Global Navigation Satellite System) receiver, a LEO (Low Earth Orbit) satellite receiver, an IMU (Inertial Measurement Unit), and a BA (Barometric Altimeter).

[0082] Step 1) Begin by using the specific force and angular velocity of the UAV measured by the IMU, and through mechanical arrangement, calculate the current UAV navigation state estimate based on inertial navigation. As an IMU factor, it is represented as follows:

[0083]

[0084] in, , and These represent the three-dimensional position estimation, velocity estimation, and carrier vector estimation of the UAV, respectively; superscript Represents the matrix transpose operator; subscript and These represent the IMU and the epoch number, respectively.

[0085] Step 2) Perform fault detection on the GNSS data and obtain the GNSS factor using the fault-detected data.

[0086] First, perform preliminary fault detection. Then, calculate a three-dimensional position estimate based on inertial navigation. GNSS inversion pseudorange :

[0087]

[0088] in, Indicates GNSS satellite Location; Indicates GNSS satellite The geometric distance between the inertial-estimated UAV position and the position of the UAV; Represents the speed of light; and These represent the estimated values ​​of receiver clock bias and satellite clock bias, respectively. and These represent the estimated values ​​for tropospheric delay and ionospheric delay, respectively; subscripts. and These represent GNSS and GNSS satellite serial numbers, respectively.

[0089] Velocity estimation based on inertial navigation Calculate GNSS satellites Inversion Doppler frequency shift observations :

[0090]

[0091] in, Indicates GNSS satellite The center frequency of the carrier wave; Indicates GNSS satellite The unit direction vector to the receiver; Indicates GNSS satellite The velocity vector; This indicates receiver clock drift.

[0092] Building GNSS satellites Fault detection statistics :

[0093]

[0094] in, and These represent GNSS satellites. Measured pseudorange and measured Doppler frequency shift; and These represent the standard deviations of GNSS pseudorange and Doppler shift observation noise, respectively.

[0095] threshold Take an experience value of 3 to 6. If Determine GNSS satellites If the observed value is normal, it is marked as a faulty satellite.

[0096] The preliminary GNSS fault detection described above is verified. If the following formula is not satisfied, the GNSS positioning solution for the current epoch is calculated using the pseudorange after the preliminary fault detection, based on the least squares method. As a GNSS factor, proceed directly to step 3); if the following equation is satisfied, it is determined that due to the large deviation in the inertial navigation state estimation, too many satellites are incorrectly marked as faults, and further fault detection based on decoupling is required:

[0097]

[0098] In the formula, This indicates the number of GNSS satellites marked as faulty; Indicates the number of visible GNSS satellites; The empirical parameters are taken as 0.3 to 0.6.

[0099] The following is a GNSS fault detection method based on deconstruction: Using the pseudoranges of currently visible GNSS satellites, a subset of all locatable pseudoranges is constructed:

[0100]

[0101] in, The minimum number of pseudoranges included is the number of constellations used plus 3; This represents the total number of pseudorange subsets of all currently locatable GNSS satellites.

[0102] Calculate the localization solution for each pseudorange subset:

[0103]

[0104] In three-dimensional space, traverse each localization solution, taking that localization solution as the geometric center, and... Construct a sphere for the radius, and count the number of other localized solutions that fall within the sphere:

[0105]

[0106] in, This indicates that the solution falls within the localized solution. The number of other locating solutions within the sphere at the center of the sphere.

[0107] Will Maximum value The pseudoranges within the corresponding pseudorange subset are marked as normal pseudoranges, and the other pseudoranges are marked as fault pseudoranges. maximum value The corresponding localization solution is taken as the final GNSS localization solution for the current epoch. , as a GNSS factor.

[0108] Step 3) Construct the UAV navigation state vector based on LEO / BA fusion for the previous epoch. :

[0109]

[0110] in, , , These represent the UAV's position, velocity, and LEO receiver clock drift based on LEO / BA integrated navigation in the previous epoch.

[0111] Using currently visible LEO satellites, a subset of all measurable Doppler shift observations was constructed:

[0112]

[0113] in, The minimum number of Doppler shift observations included is the number of constellations used plus 3; This represents the total number of all current subsets of measurable Doppler observations.

[0114] Calculate the velocity solution for each Doppler frequency shift subset:

[0115]

[0116] In three-dimensional space, traverse each velocity solution, taking that velocity solution as the geometric center, and... Construct a sphere with radius, and count the number of other velocity solutions falling inside the sphere:

[0117]

[0118] in, This indicates that the solution falls on the velocity. The number of other velocity solutions within the sphere at the center of the sphere.

[0119] Will Maximum value The pseudorange within the corresponding Doppler frequency shift subset is labeled as normal Doppler frequency shift, while other Doppler frequency shifts are labeled as faulty Doppler frequency shifts. If there are more than one identical pseudorange... Then, the pseudo-ranges within the subset of Doppler frequency shifts corresponding to these maximum values ​​are all marked as normal Doppler frequency shifts, and other Doppler frequency shifts are marked as faulty Doppler frequency shifts.

[0120] Construct observation vectors based on normal LEO Doppler frequency shift and BA measurements after elevation transformation. :

[0121]

[0122] in, They represent LEO satellites. The center frequency of the carrier wave; They represent LEO satellites. Measured Doppler frequency shift; This represents the current BA measurement after elevation conversion.

[0123] right Perform state prediction based on the constant velocity assumption and based on After the measurement update, the UAV navigation estimation result based on LEO / BA fusion is obtained for the current epoch. and the location of the drones within. and will As a LEO / BA factor.

[0124] Step 4) Define the GNSS / IMU / LEO / BA multi-source fusion navigation state vector based on the factor graph as follows:

[0125]

[0126] Among them, subscript Representative factor plot; , and These represent the estimated three-dimensional position, velocity, and carrier vector of the UAV after factor graph information fusion.

[0127] like Figure 3 As shown, by using the IMU factor As a state transition node, the GNSS factor and LEO / BA factor As an observation correction node, and using the previous epochs. to Historical data of an epoch is used as a sliding window ( (representing the length of the factor graph sliding window), constructing a vector for estimating the UAV navigation state in the current epoch. Factor plot.

[0128] Make the drone fly in a straight line Second( To ensure sufficient initialization time, the initial sliding window sequence is determined by performing least-squares calculations on the GNSS data. This completes the initialization of the factor graph.

[0129] Calculate the IMU factor in real time using historical data within a sliding window. GNSS factor LEO / BA factor The estimated value of the error covariance:

[0130]

[0131]

[0132]

[0133] in, These are IMU factors GNSS factor LEO / BA factor The estimated value of the error covariance.

[0134] IMU factor GNSS factor LEO / BA factor The weight matrices are respectively .

[0135] Step 5) Based on the historical data within the sliding window, calculate the optimal navigation state of the current UAV using a factor graph, i.e., solve the following nonlinear optimization problem to obtain the optimal estimated navigation state vector that minimizes the right side of the equation. To achieve weighted multi-source fusion navigation and positioning using GNSS / IMU / LEO / BA:

[0136]

[0137] in, Indicates the use of and Calculate the IMU edge factor residual; Indicated by It is the square of the weighted Euclidean norm of the weight matrix.

[0138] Algorithm output UAV position estimation Speed ​​estimation Attitude estimation .

[0139] Example:

[0140] This invention implements and verifies the proposed algorithm through simulation experiments. First, a reference flight trajectory is generated based on the UAV's maritime flight mission plan. The reference trajectory includes the UAV's true position, velocity, and attitude information in the geographic coordinate system. Based on this reference trajectory, observation models for GNSS, IMU, LEO, BA, and VO are established respectively. Measurement noise and error terms conforming to actual engineering characteristics are introduced into each observation model to simulate the actual observation output of each sensor in the maritime environment. The sampling frequencies for GNSS, IMU, LEO, BA, and VO are set to 1Hz, 200Hz, 1Hz, 10Hz, and 5Hz, respectively. GNSS multipath interference at sea is simulated during the periods of 170s to 220s and 1310s to 1390s. The control method selected for the simulation experiment is the GNSS / IMU / VO integrated navigation algorithm. Under the above simulation environment, the observation data from each sensor are input into the multi-source fusion positioning algorithm and the control algorithm proposed in this invention, and the UAV's position and attitude estimation results are output in real time. By comparing and analyzing the fused positioning results with the reference trajectory, the positioning accuracy of the method of the present invention in complex maritime environments is evaluated, thereby verifying the effectiveness of the method of the present invention. Figure 2 As shown, the positioning errors of the two methods in both horizontal and vertical directions were compared. During the periods of 170s to 220s and 1310s to 1390s, due to the influence of GNSS multipath interference, the positioning error of the control algorithm increased significantly, with maximum horizontal and vertical positioning errors reaching approximately 80 meters and 110 meters, respectively. During the same period, the positioning error of the proposed method did not increase significantly, and the horizontal and vertical positioning errors of the proposed method remained at approximately 1 meter throughout the experiment. The experimental results demonstrate that the proposed method can maintain a positioning accuracy better than 1 meter in the complex multipath environment at sea by effectively eliminating faulty data and using multi-source information weighted fusion based on factor graphs, thus meeting the needs of maritime UAV applications.

[0141] In its specific implementation, this application provides a computer storage medium and a corresponding data processing unit. The computer storage medium is capable of storing a computer program, which, when executed by the data processing unit, can run the invention's content regarding a multi-source fusion-based maritime UAV positioning method, as well as some or all of the steps in various embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0142] Those skilled in the art will clearly understand that the technical solutions in the embodiments of the present invention can be implemented using computer programs and their corresponding general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of computer programs, i.e., software products. These computer program software products can be stored in a storage medium and include several instructions to cause a device containing a data processing unit (which may be a personal computer, server, microcontroller, MCU, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0143] This invention provides a concept and method for maritime unmanned aerial vehicle (UAV) positioning based on multi-source fusion. Many methods and approaches exist for implementing this technical solution; the above description is merely a preferred embodiment. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this invention, and these improvements and modifications should also be considered within the scope of protection of this invention. All components not explicitly stated in this embodiment can be implemented using existing technologies.

Claims

1. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion, characterized in that, Includes the following steps: Step 1: Calculate the drone navigation state estimate based on the IMU sensor on the drone. Construct IMU factors; Step 2: Based on the GNSS receiver and IMU sensor on the UAV, perform fault detection on the GNSS data and calculate the GNSS positioning solution. Construct GNSS factors; Step 3: Calculate the drone's position based on the LEO receiver and BA sensor on the drone. Construct the LEO / BA factor; Step 4: Construct a factor graph based on the IMU factor, GNSS factor, and LEO / BA factor, and calculate the weight matrix of each factor using a sliding window. Step 5: Based on the weight matrix, perform information fusion and solve to obtain the final navigation and positioning result, including: Step 5-1: Based on the historical data within the sliding window, use the factor graph to calculate the current optimal navigation state of the UAV and construct a nonlinear optimization problem; Step 5-2: Solve the nonlinear optimization problem constructed in Step 5-1 to obtain the optimal estimated navigation state vector. Among them, the optimal estimated navigation state vector Includes drone position estimation Speed ​​estimation and attitude estimation This is the final navigation and positioning result.

2. The maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 1, characterized in that, The construction of the IMU factor described in step 1 includes: The specific force and angular velocity of the drone measured by IMU sensors are used to calculate the current force through mechanical programming. Epichronous-based inertial navigation-based estimation of UAV navigation state And it is expressed as an IMU factor as follows: ； in, , and These represent the three-dimensional position estimation, velocity estimation, and carrier vector estimation of the UAV, respectively; superscript Represents the matrix transpose operator; subscript and These represent the IMU and the epoch number, respectively.

3. The maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 2, characterized in that, The construction of GNSS factors described in step 2 includes: Step 2-1: Perform preliminary fault detection, mark the faulty satellites, and obtain the number of GNSS satellites marked as faulty. and the number of visible GNSS satellites ; Step 2-2, Set empirical parameters The number of GNSS satellites marked as faulty and the number of visible GNSS satellites Perform a check; if it does not meet the requirements... ； The current epoch is then calculated using the pseudorange of GNSS satellites with normal observations after initial fault detection. GNSS positioning solution based on least squares method As a GNSS factor, otherwise, fault detection based on deconstruction is performed and the current epoch is calculated accordingly. Final GNSS positioning solution As a GNSS factor, The value ranges from 0.3 to 0.

6.

4. The maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 3, characterized in that, Step 2-1 specifically includes the following steps: Step 2-1-1: Calculate the 3D position estimate based on inertial navigation. GNSS inversion pseudorange , means as follows: ； in, Indicates GNSS satellite Location; Indicates GNSS satellite The geometric distance between the inertial-estimated UAV position and the position of the UAV; Represents the speed of light; and These represent the estimated values ​​of receiver clock bias and satellite clock bias, respectively. and These represent the estimated values ​​for tropospheric delay and ionospheric delay, respectively; subscripts. and These represent GNSS and GNSS satellite serial numbers, respectively. Step 2-1-2, Velocity estimation based on inertial navigation Calculate GNSS satellites Inversion Doppler frequency shift observations , means as follows: ； in, Indicates GNSS satellite The center frequency of the carrier wave; Indicates GNSS satellite The unit direction vector to the receiver; Indicates GNSS satellite The velocity vector; Indicates receiver clock drift; Step 2-1-3, Construct GNSS satellites Fault detection statistics , means as follows: ； in, and These represent GNSS satellites. Measured pseudorange and measured Doppler frequency shift; and These represent the standard deviations of GNSS pseudorange and Doppler shift observation noise, respectively. Step 2-1-4, Set up fault detection statistics threshold ,like Then determine GNSS satellite If the observed values ​​are normal, then mark it as a faulty satellite.

5. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 4, characterized in that, Step 2-2 describes performing fault detection based on deconstruction and calculating the current epoch accordingly. Final GNSS positioning solution ,include: Step 2-2-1: Using the pseudoranges of currently visible GNSS satellites, construct a subset of all possible locationable pseudoranges. , means as follows: ； in, The minimum number of pseudoranges included is the number of constellations used plus 3; This represents the total number of pseudorange subsets of all currently locatable GNSS satellites; Step 2-2-2: Calculate the localization solution corresponding to each pseudo-range subset. ; Step 2-2-3: Traverse each localization solution in three-dimensional space, using the localization solution as the geometric center, and... Construct a sphere with radius, and count the number of other localized solutions that fall within the sphere. ; and use This indicates that the solution falls within the localized solution. The number of other locating solutions within the sphere at the center of the sphere; Step 2-2-4, Quantity maximum value The pseudoranges within the corresponding pseudorange subset are marked as normal pseudoranges, and the other pseudoranges are marked as fault pseudoranges. maximum value The corresponding positioning solution is used as the final GNSS positioning solution for the current epoch. , as a GNSS factor.

6. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 5, characterized in that, The construction of the LEO / BA factor described in step 3 includes: Step 3-1, construct the previous epoch UAV navigation state vector based on LEO / BA fusion , means as follows: ； in, , and These are the UAV position, velocity, and LEO receiver clock drift based on LEO / BA integrated navigation in the previous epoch; Step 3-2: Using currently visible LEO satellites, construct a subset of all measurable Doppler shift observations. , means as follows: ； in, The minimum number of Doppler shift observations included is the number of constellations used plus 3; This represents the total number of all measurable Doppler observations in the subset. Step 3-3: Calculate the velocity solution for each Doppler shift subset. ; Steps 3-4: In three-dimensional space, traverse each velocity solution, using that velocity solution as the geometric center, and... Construct a sphere with radius, and count the number of other velocity solutions falling inside the sphere. ; and use This indicates that the solution falls on the velocity. The number of other velocity solutions within the sphere at the center of the sphere; Steps 3-5, Quantity Maximum value The Doppler frequency shifts within the corresponding subset are labeled as normal Doppler frequency shifts, and other Doppler frequency shifts are labeled as faulty Doppler frequency shifts; if there is more than one identical maximum value... Then, the Doppler frequency shifts within the subset of Doppler frequency shifts corresponding to these maximum values ​​are all marked as normal Doppler frequency shifts, and other Doppler frequency shifts are marked as faulty Doppler frequency shifts; Steps 3-6: Construct observation vectors based on normal LEO Doppler frequency shift and BA measurements after elevation transformation. ; Steps 3-7: Configure the UAV navigation state vector Perform state prediction based on the constant velocity assumption and based on observation vectors After the measurement update, the UAV navigation estimation result based on LEO / BA fusion is obtained for the current epoch. and the location of the drones within. The location of the drone As a LEO / BA factor.

7. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 6, characterized in that, The observation vectors described in steps 3-6 , means as follows: ； in, They represent LEO satellites. The center frequency of the carrier wave; They represent LEO satellites. Measured Doppler frequency shift; This represents the current BA measurement after elevation conversion.

8. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 7, characterized in that, Step 4, which involves calculating the weight matrix of each factor, includes: Step 4-1, define the multi-source fusion navigation state vector based on factor graph as follows: , means as follows: ； Among them, subscript Representative factor plot; , and These represent the estimated three-dimensional position, velocity, and carrier vector of the UAV after factor graph information fusion; Step 4-2, IMU factor As a state transition node, the GNSS factor and LEO / BA factor As an observation correction node, and using the previous epochs. to Historical data from different eras is used as a sliding window, in which The length of the factor graph sliding window is used to construct the UAV navigation state vector for estimating the current epoch. Factor plot; Step 4-3: Make the drone fly in a straight line. Seconds, by performing least-squares calculations on GNSS data, the initial sliding window sequence is determined. This completes the initialization of the factor graph; Step 4-4: Calculate the IMU factor in real time using historical data within the sliding window. GNSS factor and LEO / BA factor Error covariance estimate , and They are represented as follows: ； ； ； in, These are IMU factors GNSS factor LEO / BA factor The estimated value of the error covariance; Steps 4-5 yield the IMU factor. GNSS factor and LEO / BA factor The weight matrices are respectively , and .

9. A maritime unmanned aerial vehicle (UAV) positioning method based on multi-source fusion according to claim 8, characterized in that, The nonlinear optimization problem described in step 5-2 is represented as follows: ； in, Indicates the use of and Calculate the IMU edge factor residual; Indicated by It is the square of the weighted Euclidean norm of the weight matrix.

Images