An RTK / INS integrated positioning method based on factor graph optimization

Abstract

This invention relates to an RTK / INS integrated positioning method based on factor graph optimization, comprising: setting the length of a sliding window; constructing a state vector of ambiguity parameters for each node at any given time based on IMU and GNSS data within the sliding window; marginalizing the states of ambiguity parameters that have moved out of the sliding window, and then combining this marginalized state with prior information to obtain marginalized prior information; constructing GNSS factors and IMU pre-integration factors based on the state vectors; constructing an objective function for the floating-point ambiguity solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors; and performing positioning based on the ambiguity calculated using the objective function; marginalizing the states that have moved out of the window using Schur complement, thus preserving the prior information of the nodes in the moved-out states, ensuring the accuracy of the floating-point ambiguity while reducing computational costs, and increasing the success rate and reliability of the RTK / INS integrated positioning results.

Description

Technical Field

[0001] This invention relates to the field of positioning and navigation technology, and in particular to an RTK / INS combined positioning method based on factor graph optimization. Background Technology

[0002] Existing positioning technologies can utilize measurement information from various sensors to estimate inter-frame motion, including visual sensors, lidar sensors, GNSS (Global Navigation Satellite System), and IMU (Inertial Measurement Unit).

[0003] GNSS's RTK (Real-Time Kinematic) carrier phase differential technology fully utilizes observation data from the rover and reference stations to construct double-difference observations, which can significantly reduce or eliminate some errors. Furthermore, the double-difference observations contain relatively small errors and have the ability to recover integer solutions from real-valued solutions, achieving centimeter-level positioning accuracy. However, GNSS receivers require at least four satellites to complete positioning, and the GNSS signal's susceptibility to environmental interference is amplified in environments with obstructions such as cities, canyons, tunnels, and overpasses, easily leading to receiver lock-off and affecting navigation accuracy, potentially causing navigation failure. Alternatively, there is another type of system on the market: INS (Inertial Navigation System). INS has the ability to independently calculate position, velocity, and attitude. Its signal is less susceptible to external environmental interference, features high short-term accuracy, and has a high data update rate, providing rich attitude information. Its applications are also very widespread. However, in practical use, because INS positioning is an integral process, its error accumulates over time; that is, the longer it is used, the worse its positioning accuracy becomes. Furthermore, it requires initial position, velocity, and attitude information before use, typically limiting its application to short-term positioning needs. Therefore, the industry often adopts RTK / INS integrated navigation systems to combine the advantages and compensate for the disadvantages of both RTK and INS, thereby improving the accuracy of the navigation system and providing users with more precise, continuous, and reliable navigation and positioning information.

[0004] The key technology of GNSS / INS fusion lies in using Kalman filtering at a single epoch. This avoids the risk of position error divergence caused by poor-quality GNSS updates by eliminating gross errors through system integrity monitoring. However, relying solely on the IMU output of the current epoch to calculate the navigation solution increment suffers from insufficient utilization of trajectory information from previous and subsequent time points. Introducing additional sensors for assistance leads to increased costs and narrowed applicability and application range. Existing RTK / INS tightly coupled algorithms based on sliding window smoothing methods often assume constant satellite visibility or perform post-processing fusion only in intervals with sufficient double-difference carrier observations and fixed ambiguity solutions. This cannot meet the requirements for long-distance, high-precision positioning with a certain degree of real-time performance. Summary of the Invention

[0005] This invention addresses the technical problems existing in the prior art by providing an RTK / INS combined localization method based on factor graph optimization. The method uses Schur complement to marginalize the state of the moved-out window, preserving the prior information of the nodes in the moved-out state. This ensures the accuracy of floating-point ambiguity while reducing computational costs, and increases the success rate and reliability of RTK / INS combined localization results.

[0006] According to a first aspect of the present invention, a factor graph-based RTK / INS combined localization method is provided, comprising:

[0007] Step 1: Set the length of the sliding window, and construct the state vector of the ambiguity parameters at each node time based on the IMU data and GNSS data within the sliding window;

[0008] Step 2: After marginalizing the state of the ambiguity parameter that is removed from the sliding window, the marginalization prior information is obtained by combining the prior information.

[0009] Step 3: Construct GNSS factors and IMU pre-integration factors based on the state vector; construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors; and perform localization based on the ambiguity calculated using the objective function.

[0010] Based on the above technical solution, the present invention can also be improved as follows.

[0011] Optionally, the state vector for:

[0012] ;

[0013] In the formula ;

[0014] Where n represents the length of the sliding window, and The state nodes of the nth epoch are respectively Three-dimensional position and velocity in the system; The carrier posture for the nth epoch; Add zero bias to the table for the nth epoch; The gyroscope has zero bias at the nth epoch; T is the matrix transpose. Let be the double difference ambiguity of the m-th station star.

[0015] Optionally, in step 2, the Schul complement theorem is used to marginalize the state of the ambiguity parameter that has been moved out of the sliding window.

[0016] Optionally, the prior information for marginalization obtained in step 2 is: ;

[0017] in, To marginalize residuals, It is a Jacobian matrix. This is the state vector.

[0018] Optionally, the GNSS factors include: a double-difference pseudorange factor and a double-difference carrier factor for common-view satellites at a set frequency;

[0019] The double difference pseudo-range factor and the double-difference carrier factor The double-difference pseudorange factor and double-difference carrier factor are:

[0020] ;

[0021] In the formula, Let f be the state vector to be optimized, and f be the frequency. and For common-view satellites, and They represent frequencies respectively. Up to satellite With reference satellite The double-difference pseudorange and carrier observations; This represents the double-difference satellite-to-ground distance calculated using approximate coordinates from the receiver; It represents double-difference ambiguity.

[0022] Optionally, the IMU pre-integration factor is:

[0023] ;

[0024] In the formula, ;

[0025] ;

[0026] in, Prior information for IMU-related variables; Let the state vector be the one to be optimized. and These are the pre-integrals of the gravity and Coriolis force terms with respect to position and velocity, respectively; This refers to the application of the gravity vector in the w-system; This is the projection of the Earth's rotation in the w-frame; , and These are the predicted components for position, velocity, and attitude, respectively. To add zero bias to the table; Zero bias for the gyroscope; , and These represent the position, velocity, and attitude in the w-frame, respectively. For the transformation quaternion from the b-system to the w-system; Let be the transformation matrix from the b-system to the w-system; k is the epoch; The time interval between adjacent epochs; For time intervals; It is a quaternion for the transformation from the k-1 epoch to the k epoch world coordinate system; It is a quaternion for the transformation from the b-system to the w-system.

[0027] Optionally, the objective function is:

[0028] ;

[0029] In the formula, Let be the state vector to be optimized. For marginalized prior information, As prior information, It is a Jacobian matrix; For IMU pre-integration, k is the epoch. Prior information for IMU-related variables; and Frequency The double-difference pseudorange factor and double-difference carrier factor.

[0030] According to a second aspect of the present invention, an RTK / INS combined localization system based on factor graph optimization is provided, comprising: a state vector construction module, an edge-forming factor construction unit, and a localization unit;

[0031] The state vector construction module is used to set the length of the sliding window and construct the state vector of the ambiguity parameters at each node time based on the IMU data and GNSS data within the sliding window.

[0032] The marginalization factor construction unit is used to marginalize the state of the ambiguity parameter that is moved out of the sliding window, and then combine it with prior information to obtain the prior information of marginalization.

[0033] The positioning unit is used to construct GNSS factors and IMU pre-integration factors based on the state vector, construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors and IMU pre-integration factors, and perform positioning based on the ambiguity calculated by the objective function.

[0034] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the processor is configured to implement the steps of a factor graph-optimized RTK / INS combined localization method when executing a computer management program stored in the memory.

[0035] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer management program stored thereon, which, when executed by a processor, implements the steps of a factor graph-optimized RTK / INS combined localization method.

[0036] This invention provides a factor graph-based RTK / INS combined localization method, system, electronic device, and storage medium. The method uses Schur complement to marginalize the state of the out-of-window node, preserving the prior information of the node in the out-of-window state. This ensures the accuracy of floating-point ambiguity while reducing computational cost and increasing AR success rate and reliability. During marginalization, FGO is used for global optimization, and linearization is performed again in each iteration, which reduces the nonlinearity caused by inaccurate initial approximate positions to some extent. Attached Figure Description

[0037] Figure 1 A flowchart of an RTK / INS combined localization method based on factor graph optimization provided by the present invention;

[0038] Figure 2 The diagram shown is a schematic diagram of constructing a state vector based on a sliding window according to an embodiment of the present invention;

[0039] Figure 3 A structural block diagram of an RTK / INS integrated positioning system based on factor graph optimization provided by the present invention;

[0040] Figure 4 A schematic diagram of the hardware structure of a possible electronic device provided by the present invention;

[0041] Figure 5 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided by the present invention. Detailed Implementation

[0042] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0043] Figure 1 A flowchart of an RTK / INS combined localization method based on factor graph optimization provided by this invention is shown below. Figure 1 As shown, the positioning method includes:

[0044] Step 1: Set the length of the sliding window, and construct the state vector of the ambiguity parameters of each node at any given time based on the IMU data and GNSS data within the sliding window.

[0045] Step 2: After marginalizing the state of the ambiguity parameter that is moved out of the sliding window, the marginalization prior information is obtained by combining it with the prior information.

[0046] Step 3: Construct GNSS factors and IMU pre-integration factors based on the state vector. Construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors. Perform localization based on the ambiguity calculated using the objective function.

[0047] This invention provides an RTK / INS combined localization method based on factor graph optimization. The method uses Schur complement to marginalize the state of the moved-out window, which preserves the prior information of the nodes in the moved-out state. This ensures the accuracy of floating-point ambiguity while reducing computational cost, and increases the success rate and reliability of RTK / INS combined localization results.

[0048] Example 1

[0049] Embodiment 1 provided by this invention is an embodiment of the RTK / INS combined localization method based on factor graph optimization provided by this invention, combined with... Figure 1 It can be seen that embodiments of this positioning method include:

[0050] Step 1: Set the length of the sliding window, and construct the state vector of the ambiguity parameters of each node at any given time based on the IMU data and GNSS data within the sliding window.

[0051] In one possible implementation, the state vector for:

[0052] .

[0053] In the formula .

[0054] Where n represents the length of the sliding window, and The state nodes of the nth epoch are respectively Three-dimensional position and velocity in the system; The carrier posture for the nth epoch; Add zero bias to the table for the nth epoch; The gyroscope has zero bias at the nth epoch; T is the matrix transpose. Let be the double difference ambiguity of the m-th station star.

[0055] Step 2: After marginalizing the state of the ambiguity parameter that is moved out of the sliding window, the marginalization prior information is obtained by combining it with the prior information.

[0056] In one possible embodiment, step 2 uses the Schul complement theorem to marginalize the state of the ambiguity parameter that has moved out of the sliding window.

[0057] In one possible embodiment, the marginalization prior information obtained in step 2 is: .

[0058] in, To marginalize residuals, It is a Jacobian matrix. This is the state vector.

[0059] like Figure 2 The diagram illustrates the construction of a state vector based on a sliding window according to an embodiment of the present invention. When no cycle slip occurs, the ambiguity is estimated as a constant, and continuously observed carriers are connected to the same ambiguity factor. Only when a cycle slip occurs (e.g., ...) In Cycle slip occurred, thus increasing ) or when new satellites are added (such as In This will increase the number of ambiguity factors. Current techniques use large sliding window lengths to fully utilize the characteristics of continuously observed carriers, but this significantly increases the computational burden: for high-elevation satellites, the number of consecutive observation epochs may exceed the window length. If this factor is not considered, some historical information will be lost, reducing the accuracy of floating-point ambiguity and thus affecting the success rate and reliability of AR. Therefore, for states that have moved out of the window, Shure complement is used to marginalize them, such as... Figure 2 As shown Ambiguity, after being marginalized, is equivalent to retaining its original state. If the prior information of a node is not edge-processed, then Equivalent to in The accuracy of the solution is inevitably affected by newly introduced factors at each node. FGO is a global optimization method that re-linearizes the solution in each iteration, which reduces the nonlinearity caused by inaccurate initial approximate positions to some extent.

[0060] When the sliding window length exceeds a threshold, the oldest state vector in the window will be marginalized, thereby reducing the solution scale of the optimization problem. Simultaneously, the corresponding pre-integration factor and GNSS factor will be converted into prior factors and added to the graph optimization objective function.

[0061] Step 3: Construct GNSS factors and IMU pre-integration factors based on the state vector. Construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors. Perform localization based on the ambiguity calculated using the objective function.

[0062] In one possible embodiment, the GNSS factors include: a double-difference pseudorange factor and a double-difference carrier factor for common-view satellites at a set frequency.

[0063] Double difference pseudo-distance factor and double-difference carrier factor The double-difference pseudorange factor and the double-difference carrier factor are:

[0064] .

[0065] In the formula, Let f be the state vector to be optimized, and f be the frequency. and For common-view satellites, and They represent frequencies respectively. Up to satellite With reference satellite The double-difference pseudorange and carrier observations; This represents the double-difference satellite-to-ground distance calculated using approximate coordinates from the receiver; It represents double-difference ambiguity.

[0066] In one possible embodiment, the IMU pre-integration factor is:

[0067] .

[0068] In the formula, .

[0069] .

[0070] The second term in the above formula represents the effect of Earth's rotation on the pre-integral.

[0071] in, Prior information for IMU-related variables, including position, velocity, attitude, gyroscope, and zero bias of the added table; Let the state vector be the one to be optimized. and These are the pre-integrals of the gravity and Coriolis force terms with respect to position and velocity, respectively; This refers to the application of the gravity vector in the w-system; This is the projection of the Earth's rotation in the w-frame; , and These are the predicted components for position, velocity, and attitude, respectively. To add zero bias to the table; Zero bias for the gyroscope; , and These represent the position, velocity, and attitude in the w-frame, respectively. For the transformation quaternion from the b-system to the w-system; Let be the transformation matrix from the b-system to the w-system; k is the epoch; The time interval between adjacent epochs; For time intervals; It is a quaternion for the transformation from the k-1 epoch to the k epoch world coordinate system; It is a quaternion for the transformation from the b-system to the w-system.

[0072] In one possible implementation, the objective function is:

[0073] .

[0074] In the formula, Let be the state vector to be optimized. For marginalized prior information, As prior information, It is a Jacobian matrix; For IMU pre-integration, k is the epoch. Prior information for IMU-related variables; and Frequency The double-difference pseudorange factor and double-difference carrier factor.

[0075] Example 2

[0076] Embodiment 2 provided by this invention is an embodiment of an RTK / INS combined localization system based on factor graph optimization provided by this invention. Figure 3 This invention provides a structural diagram of an RTK / INS integrated localization system based on factor graph optimization, combined with... Figure 3 It can be seen that the implementation of the positioning system includes: a state vector construction module, an edge-forming factor construction unit, and a positioning unit.

[0077] The state vector construction module is used to set the length of the sliding window and construct the state vector of the ambiguity parameters at each node time based on the IMU data and GNSS data within the sliding window.

[0078] The marginalization factor construction unit is used to marginalize the state of the ambiguity parameter that has been removed from the sliding window, and then combine it with prior information to obtain the prior information of marginalization.

[0079] The positioning unit is used to construct GNSS factors and IMU pre-integration factors based on the state vector, construct the objective function of the ambiguity floating-point solution based on the marginalized prior information, GNSS factors and IMU pre-integration factors, and perform positioning based on the ambiguity calculated by the objective function.

[0080] It is understood that the factor graph-based RTK / INS combined localization system provided by this invention corresponds to the factor graph-based RTK / INS combined localization method provided in the foregoing embodiments. The relevant technical features of the factor graph-based RTK / INS combined localization system can be referred to the relevant technical features of the factor graph-based RTK / INS combined localization method, and will not be repeated here.

[0081] Please see Figure 4 , Figure 4 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 4 As shown, this embodiment of the invention provides an electronic device, including a memory 1310, a processor 1320, and a computer program 1311 stored in the memory 1310 and executable on the processor 1320. When the processor 1320 executes the computer program 1311, it performs the following steps: setting the length of a sliding window; constructing a state vector of ambiguity parameters at each node time based on IMU data and GNSS data within the sliding window; marginalizing the states of ambiguity parameters that have moved out of the sliding window, and then combining them with prior information to obtain marginalized prior information; constructing GNSS factors and IMU pre-integration factors based on the state vectors; constructing an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors; and performing localization based on the ambiguity calculated using the objective function.

[0082] Please see Figure 5 , Figure 5 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 5As shown, this embodiment provides a computer-readable storage medium 1400, on which a computer program 1411 is stored. When the computer program 1411 is executed by a processor, it performs the following steps: setting the length of a sliding window; constructing a state vector of ambiguity parameters at each node time based on IMU data and GNSS data within the sliding window; marginalizing the states of ambiguity parameters that have moved out of the sliding window, and then combining them with prior information to obtain marginalized prior information; constructing GNSS factors and IMU pre-integration factors based on the state vectors; constructing an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors; and performing localization based on the ambiguity calculated using the objective function.

[0083] In existing technologies, for high-elevation satellites, the number of consecutive observation epochs may exceed the sliding window length. If this factor is not considered, some historical information will be lost, reducing the accuracy of floating-point ambiguity. When using a larger sliding window length to make full use of the characteristics of continuous observation carriers, the computational burden is significantly increased, which in turn affects the success rate and reliability of AR.

[0084] This invention provides an RTK / INS fusion positioning method, system, electronic device, and storage medium based on factor graph optimization. The specific RTK / INS fusion positioning process based on graph optimization is as follows: Preprocessing of reference station and rover data includes common-view satellite selection, cycle slip detection, and double-difference observation construction; merging ambiguity parameters of continuously tracked data within a window, deleting ambiguity parameters from out-of-window data, and adding ambiguity parameters for newly added and cycle-slip carrier observations; constructing double-difference pseudorange and carrier factors, as well as IMU pre-integration factors and marginalization prior factors; obtaining floating-point ambiguity and the corresponding variance-covariance matrix by solving the maximum a posteriori probability problem of the objective function; fixing the ambiguity and outputting the solution results. The state of the out-of-window is marginalized using Schur complement, preserving the prior information of the nodes in the out-of-window state, ensuring the accuracy of the floating-point ambiguity while reducing computational cost and increasing AR success rate and reliability; global optimization using FGO during marginalization, with re-linearization in each iteration, to some extent reducing the nonlinearity problem caused by inaccurate initial approximate positions.

[0085] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0086] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0091] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A factor graph-based RTK / INS combined localization method, characterized in that, The positioning method includes: Step 1: Set the length of the sliding window, and construct the state vector of the ambiguity parameters at each node time based on the IMU data and GNSS data within the sliding window; Step 2: After marginalizing the state of the ambiguity parameter that is removed from the sliding window, the marginalization prior information is obtained by combining the prior information. Step 3: Construct GNSS factors and IMU pre-integration factors based on the state vector; construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors, and IMU pre-integration factors; and perform localization based on the ambiguity calculated using the objective function.

2. The combined positioning method according to claim 1, characterized in that, The state vector for: ； In the formula ; Where n represents the length of the sliding window, and The state nodes of the nth epoch are respectively Three-dimensional position and velocity in the system; The carrier posture for the nth epoch; Add zero bias to the table for the nth epoch; The gyroscope has zero bias at the nth epoch; T is the matrix transpose. Let be the double-difference ambiguity of the m-th station star.

3. The combined positioning method according to claim 1, characterized in that, In step 2, the Schul complement theorem is used to marginalize the state of the ambiguity parameter that has been moved out of the sliding window.

4. The combined positioning method according to claim 1, characterized in that, The prior information on marginalization obtained in step 2 is: ; in, To marginalize residuals, It is a Jacobian matrix. This is the state vector.

5. The combined positioning method according to claim 1, characterized in that, The GNSS factors include: the double-difference pseudorange factor and the double-difference carrier factor for common-view satellites at a set frequency; The double difference pseudo-range factor and the double-difference carrier factor The double-difference pseudorange factor and double-difference carrier factor are: ； In the formula, Let f be the state vector to be optimized, and f be the frequency. and For common-view satellites, and They represent frequencies respectively. Up to satellite With reference satellite The double-difference pseudorange and carrier observations; This represents the double-difference satellite-to-ground distance calculated using approximate coordinates from the receiver; It represents double-difference ambiguity.

6. The combined positioning method according to claim 1, characterized in that, The IMU pre-integration factor is: ； In the formula, ; ； in, Prior information for IMU-related variables; Let the state vector be the one to be optimized. and These are the pre-integrals of the gravity and Coriolis force terms with respect to position and velocity, respectively; This refers to the application of the gravity vector in the w-system; This is the projection of the Earth's rotation in the w-frame; , and These are the predicted components for position, velocity, and attitude, respectively. To add zero bias to the table; Zero bias for the gyroscope; , and These represent the position, velocity, and attitude in the w-frame, respectively. For the transformation quaternion from the b-system to the w-system; Let be the transformation matrix from the b-system to the w-system; k is the epoch; The time interval between adjacent epochs; For time intervals; It is a quaternion for the transformation from the k-1 epoch to the k epoch world coordinate system; It is a quaternion for the transformation from the b-system to the w-system.

7. The combined positioning method according to claim 1, characterized in that, The objective function is: ； In the formula, Let be the state vector to be optimized. For marginalized prior information, As prior information, It is a Jacobian matrix; For IMU pre-integration, k is the epoch. Prior information for IMU-related variables; and Frequency The double-difference pseudorange factor and double-difference carrier factor.

8. A factor graph-based RTK / INS integrated positioning system, characterized in that, The positioning system includes: a state vector construction module, an edge-forming factor construction unit, and a positioning unit; The state vector construction module is used to set the length of the sliding window and construct the state vector of the ambiguity parameters at each node time based on the IMU data and GNSS data within the sliding window. The marginalization factor construction unit is used to marginalize the state of the ambiguity parameter that is moved out of the sliding window, and then combine it with prior information to obtain the prior information of marginalization. The positioning unit is used to construct GNSS factors and IMU pre-integration factors based on the state vector, construct an objective function for the ambiguity floating-point solution based on the marginalized prior information, GNSS factors and IMU pre-integration factors, and perform positioning based on the ambiguity calculated by the objective function.

9. An electronic device, characterized in that, The system includes a memory and a processor, wherein the processor is used to execute computer management programs stored in the memory to implement the steps of the factor graph-based RTK / INS combined localization method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer management program, which, when executed by a processor, implements the steps of the factor graph-based RTK / INS combined localization method as described in any one of claims 1-7.

Images