Satellite navigation enhancement method and system based on inertial navigation data

Abstract

The application discloses a satellite navigation enhancement method and system based on inertial navigation data, and aims at solving the problems of slow convergence, frequent cycle slip, easy loss of fixed solution, discontinuous or jump of navigation output of real-time dynamic differential positioning and precise point positioning calculation under high dynamic and complex shielding environment. The application performs time alignment on inertial measurement unit data, carrier phase and Doppler observation data, constructs continuous time state trajectory and establishes Gaussian process prior. The trajectory state and carrier phase integer ambiguity are tightly coupled and jointly estimated in a sliding time window. Change point detection is performed based on carrier phase residuals to determine cycle slip, and ambiguity reset and reinitialization constraint based on inertial prediction state and uncertainty are triggered to complete update estimation. The technical effects of robust cycle slip detection and rapid recovery, improved fixed solution retention capability and switching smoothness, enhanced navigation calculation continuity and stability are realized.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation and positioning and inertial navigation, and in particular to a satellite navigation enhancement method and system based on inertial navigation data. Background Technology

[0002] Satellite navigation and positioning technology can achieve continuous positioning based on pseudorange, carrier phase, and other observations. Real-time dynamic differential positioning and precise single-point positioning can provide high-precision results ranging from centimeters to decimeters in open environments and have been widely applied in scenarios such as drones, autonomous driving, surveying, and mobile robots. To further improve usability in dynamic environments, the industry typically combines inertial measurement units (IMUs) to construct integrated navigation schemes. The short-term stability of IMUs compensates for observation discontinuities caused by satellite signal obstruction and geometric degradation. The fusion framework has also evolved from early loose coupling to tight coupling and even deep coupling, and solution forms more suitable for high-frequency inertial data and asynchronous observations have emerged, such as sliding time window optimization, factor graphs, and continuous time trajectory modeling. Meanwhile, to ensure the continuity of carrier phase observations, existing technologies generally employ methods such as geometrically independent combination, time difference, or residual thresholding for cycle slip detection and repair, and reinitialize ambiguity and filtering states after lock-out or cycle slip.

[0003] However, existing technologies still have shortcomings in highly dynamic and complex occlusion environments, mainly in the following aspects: 1. When the number of visible satellites decreases, obstruction occurs frequently, or the motion acceleration is large, the convergence speed of differential positioning and precise single-point positioning slows down, the fixed solution of carrier phase integer ambiguity is difficult to maintain, and the fixed solution is prone to loss and frequent switching between floating-point solution and fixed solution, resulting in unstable positioning output.

[0004] 2. Under conditions of multipath, signal loss, and rapid changes in observation quality, cycle slip detection is susceptible to noise and model errors, increasing the probability of missed or false detections, which in turn leads to erroneous ambiguity processing and degraded solution.

[0005] 3. The reinitialization of the state after cycle slip or loss of lock often relies on simple reset strategies or weak constraint processing, which can easily cause the solution result to jump abruptly in the reinitialization epoch, making it difficult to guarantee the continuity and smoothness of the output.

[0006] Therefore, a satellite navigation enhancement method and system that can overcome the shortcomings of the existing technology is a problem that needs to be solved by those skilled in the art. Summary of the Invention

[0007] One objective of this invention is to propose a satellite navigation enhancement method based on inertial navigation data. Addressing the problems of slow convergence, frequent carrier phase cycle slips, loss of fixed integer ambiguity solutions, and discontinuities or jumps in positioning output caused by switching between floating-point and fixed solutions in high-dynamic and complex obstruction environments, this invention proposes a method that constructs a continuous-time state trajectory based on inertial measurement unit data and establishes a Gaussian process prior. Within a sliding time window, the trajectory prior is tightly coupled with the carrier phase observation equation and the Doppler observation equation to form a joint estimation to obtain the fused solution result and observation residual sequence. Furthermore, change point detection is performed on the carrier phase residual to determine cycle slips, generating cycle slip markers and ambiguity reset commands. When a cycle slip is triggered, the carrier phase integer ambiguity of the corresponding satellite is reset. Simultaneously, re-initialization constraints are constructed based on the inertial prediction state and its uncertainty to complete the update estimation and output the enhanced navigation solution result. This invention possesses the technical effects of robust cycle slip detection, rapid ambiguity recovery, improved fixed solution preservation capability, and smooth suppression of jumps during the re-initialization process, thereby improving the continuity and stability of navigation solution.

[0008] On the one hand, the present invention provides a satellite navigation enhancement method based on inertial navigation data, comprising: S1. Collect inertial measurement unit (IMU) data and satellite navigation observation data, and perform time alignment processing. The satellite navigation observation data includes carrier phase observations. S2. Construct a continuous-time state trajectory based on the IMU data, establish a Gaussian process prior model for the continuous-time state trajectory, generate the inertial prediction state and its uncertainty, and obtain trajectory prior information. S3. Construct a tightly coupled observation model based on the trajectory prior information and satellite navigation observation data. Within a sliding time window, construct and solve a joint estimation problem for the continuous-time state trajectory and carrier phase integer ambiguity to obtain the initial fusion solution result and calculate the observation residual sequence. S4. Detect change points in the residuals corresponding to the carrier phase observations based on the observation residual sequence. S5. Based on the cycle slip determination, the carrier phase cycle slip detection result is output, and cycle slip processing information is generated, including cycle slip markers and corresponding ambiguity reset instructions; S6. Based on the cycle slip processing information, the carrier phase integer ambiguity corresponding to the satellite with the cycle slip marker is reset according to the ambiguity reset instructions, and re-initialization constraints are constructed based on the initial fusion solution results and the inertial prediction state and its uncertainty, generating re-initialization solution initial values ​​and re-initialization constraint information; S7. Within the sliding time window, based on the trajectory prior information, satellite navigation observation data, re-initialization solution initial values, and re-initialization constraint information, the continuous time state trajectory and carrier phase integer ambiguity are updated and estimated, and the enhanced navigation solution results are output.

[0009] Optionally, S1 includes: Record corresponding time labels for angular velocity data, acceleration data, and satellite navigation observation data respectively; A unified alignment time series is determined, and the angular velocity data and acceleration data are resampled according to the alignment time series to obtain angular velocity aligned data and acceleration aligned data; Time alignment processing is performed on satellite navigation observation data to ensure that the time labels of carrier phase observations and Doppler observations are consistent with the aligned time series. The angular velocity aligned data and the acceleration aligned data are combined to form synchronous inertial measurement unit data, and the time-aligned satellite navigation observation data is determined as synchronous satellite navigation observation data.

[0010] Optionally, S2 includes: Multiple trajectory moments are set within the time range corresponding to the synchronous inertial measurement unit data, and state variables of the continuous time state trajectory are defined at each trajectory moment. The state variables include position, velocity, attitude, and inertial measurement unit error parameters. Based on the data from the synchronous inertial measurement unit, inertial recursion is performed between adjacent trajectory times to obtain the state correlation between adjacent trajectory times; Based on the state correlation, a Gaussian process prior model of the continuous-time state trajectory is constructed, and the inertial prediction state and its uncertainty corresponding to each trajectory time are calculated by the Gaussian process prior model. The inertial prediction state and its uncertainty are determined as trajectory prior information.

[0011] Optionally, S3 includes: Based on the prior trajectory information, the predicted value of the continuous time state trajectory at each moment within the sliding time window is determined, and a tightly coupled observation model is constructed by combining the ephemeris information of the corresponding satellite in the synchronous satellite navigation observation data. The tightly coupled observation model includes a carrier phase observation equation corresponding to the carrier phase observation value and a Doppler observation equation corresponding to the Doppler observation value. The unknowns of the carrier phase observation equation include at least the state variables of the continuous-time state trajectory and the carrier phase integer ambiguity. The unknowns of the Doppler observation equation include at least the state variables of the continuous-time state trajectory. Within the sliding time window, the trajectory prior information, the carrier phase observation equation, and the Doppler observation equation are combined to form a joint estimation problem, and the initial fusion solution is obtained by iterative solution. Based on the initial fusion solution results, the observation residuals corresponding to the carrier phase observations and the Doppler observations are calculated respectively, and the observation residual sequence is formed in chronological order.

[0012] Optionally, S4 includes: Extract the carrier phase residual sequence corresponding to the carrier phase observation value from the observation residual sequence; The carrier phase residual sequence is subjected to change point determination. The change point determination adopts the Bayesian online change point detection algorithm or the generalized likelihood ratio test algorithm, and the cycle slip determination quantity is calculated at each epoch. When the cycle slip determination quantity meets the preset determination conditions, the corresponding epoch and the corresponding satellite are marked as cycle slip marks and the carrier phase cycle slip detection result is output. Cycle slip processing information is generated based on the cycle slip marker, and the cycle slip processing information includes the cycle slip marker and the ambiguity reset instruction corresponding to the cycle slip marker.

[0013] Optionally, S5 includes: The satellite corresponding to the cycle slip mark is read according to the cycle slip processing information, and the carrier phase integer ambiguity of the satellite corresponding to the cycle slip mark is reset according to the ambiguity reset instruction; The initial range of the reset carrier phase integer ambiguity is determined based on the initial fusion solution results; Based on the inertial prediction state and its uncertainty, a reinitialization constraint is constructed. The reinitialization constraint is used to constrain the deviation between the state variables of the continuous-time state trajectory and the inertial prediction state at the reinitialization epoch. The initial values ​​for reinitialization and the reinitialization constraint information are generated based on the initial value range and the reinitialization constraint conditions.

[0014] Furthermore, the weights or covariance of the reinitialization constraints are determined by the inertial prediction state and its uncertainty, and the constraint weights on the inertial prediction state are increased when a cycle slip or a decrease in the number of available satellites is detected, in order to suppress the solution jump caused by reinitialization.

[0015] Optionally, S6 includes: Within the sliding time window, the re-initialized initial value is used as the initial state for iterative solution, and the re-initialization constraint information is added to the joint estimation problem; Gaussian process prior constraints are applied to the continuous time state trajectory based on the trajectory prior information, and carrier phase observation equation constraints and Doppler observation equation constraints are applied to the continuous time state trajectory and carrier phase integer ambiguity based on the synchronous satellite navigation observation data. The updated continuous-time state trajectory and the updated carrier phase integer ambiguity are obtained by iterative solution. The updated continuous-time state trajectory is then evaluated at the output epoch to obtain the enhanced navigation solution result. The enhanced navigation solution includes the position information, the velocity information, and the attitude information.

[0016] Optionally, the preset determination condition for the change point determination is an adaptive threshold, which is updated online based on at least one of the following: the uncertainty of the inertial prediction state and the variance of the carrier phase residual sequence, so that the robustness of the cycle slip determination is improved when the uncertainty of the inertial prediction state increases.

[0017] On the other hand, the present invention also provides a satellite navigation enhancement system based on inertial navigation data, comprising: The data alignment module is used to collect inertial measurement unit data and satellite navigation observation data, and perform time alignment. The trajectory prior module is used to construct a continuous-time state trajectory based on the data from the inertial measurement unit and establish a Gaussian process prior model, outputting the inertial prediction state and its uncertainty as trajectory prior information. The tightly coupled solution module is used to construct a joint estimation problem based on the trajectory prior information and the satellite navigation observation data within a sliding time window, solve the ambiguity of the continuous time state trajectory and carrier phase integer, obtain the initial fusion solution result and generate the observation residual sequence; The cycle slip detection module is used to detect change points in the carrier phase residual based on the observation residual sequence to determine cycle slips and generate cycle slip processing information including cycle slip markers and ambiguity reset instructions. The reinitialization module is used to reset the carrier phase integer ambiguity corresponding to the satellite that has a cycle slip mark according to the cycle slip processing information, and to generate reinitialization initial values ​​and reinitialization constraint information based on the initial fusion solution results and the inertial prediction state and its uncertainty. The update output module is used to update the estimated continuous-time state trajectory and carrier phase integer ambiguity within the sliding time window based on the trajectory prior information, the satellite navigation observation data, the re-initialization solution initial value, and the re-initialization constraint information, and output the enhanced navigation solution result.

[0018] The beneficial effects of this invention are: 1. By constructing a continuous-time state trajectory based on inertial measurement unit data and introducing Gaussian process priors, and performing tight-coupled joint estimation with carrier phase and Doppler observations within a sliding time window, stable prior support can be provided for the state solution under high dynamic and satellite geometric degradation conditions. This improves the solution convergence speed and the ability to maintain integer ambiguity fixed solutions, and reduces the probability of fixed solution loss and frequent switching between floating-point solutions and fixed solutions.

[0019] 2. Cycle slip determination is performed by performing change point detection based on the carrier phase residual sequence formed by tightly coupled solution. This can improve the robustness of cycle slip detection, reduce missed detections and false detections, and reduce solution degradation caused by cycle slips, even when observation noise, occlusion and multipath cause changes in residual statistical characteristics.

[0020] 3. After detecting cycle slip, the integer ambiguity of the corresponding satellite is reset, and a re-initialization constraint is constructed in combination with the inertial prediction state and its uncertainty. This makes the re-initialization have a controllable prior constraint strength, thereby suppressing the solution jump of the re-initialization epoch and improving the output continuity and switching smoothness in the process of occlusion recovery or cycle slip repair. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a satellite navigation enhancement method based on inertial navigation data proposed in this invention. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0023] refer to Figure 1 A satellite navigation enhancement method based on inertial navigation data includes: S1. Collect inertial measurement unit (IMU) data and satellite navigation observation data, and perform time alignment processing. The satellite navigation observation data includes carrier phase observations. S2. Construct a continuous-time state trajectory based on the IMU data, establish a Gaussian process prior model for the continuous-time state trajectory, generate the inertial prediction state and its uncertainty, and obtain trajectory prior information. S3. Construct a tightly coupled observation model based on the trajectory prior information and satellite navigation observation data. Within a sliding time window, construct and solve a joint estimation problem for the continuous-time state trajectory and carrier phase integer ambiguity to obtain the initial fusion solution result and calculate the observation residual sequence. S4. Detect change points in the residuals corresponding to the carrier phase observations based on the observation residual sequence. S5. Based on the cycle slip determination, the carrier phase cycle slip detection result is output, and cycle slip processing information is generated, including cycle slip markers and corresponding ambiguity reset instructions; S6. Based on the cycle slip processing information, the carrier phase integer ambiguity corresponding to the satellite with the cycle slip marker is reset according to the ambiguity reset instructions, and re-initialization constraints are constructed based on the initial fusion solution results and the inertial prediction state and its uncertainty, generating re-initialization solution initial values ​​and re-initialization constraint information; S7. Within the sliding time window, based on the trajectory prior information, satellite navigation observation data, re-initialization solution initial values, and re-initialization constraint information, the continuous time state trajectory and carrier phase integer ambiguity are updated and estimated, and the enhanced navigation solution results are output.

[0024] In this specific embodiment, S1 includes: The inertial measurement unit data and satellite navigation observation data were time-aligned according to a unified time reference to form a synchronized dataset; Angular velocity and acceleration data output from the inertial measurement unit are collected separately, and a corresponding time stamp is recorded when each angular velocity and acceleration data is generated. The time stamp uses the satellite navigation system time consistent with the satellite navigation receiver and is expressed in seconds. Each sample of angular velocity data is recorded as a three-axis angular velocity vector. And its time stamp is recorded as Each sample of acceleration data is denoted as a triaxial acceleration vector. And its time stamp is recorded as ,in The unit of measurement is , The unit of measurement is And compensate for the fixed time delay between the internal clock of the inertial measurement unit and the time of the satellite navigation system to make and All are within the same satellite navigation system time coordinate system; Subsequently, satellite navigation observation data was collected, and for each epoch and each available satellite, carrier phase observation values ​​and Doppler observation values, along with their time labels, were recorded. The time label of each epoch is denoted as The satellites in this epoch are identified as The carrier phase observation value is denoted as Furthermore, its unit is a week, and the satellites in that epoch are identified as... Doppler observations are denoted as Its unit is Hz and it represents the rate of change of the carrier phase over time; Next, a unified alignment time series is determined, and the alignment time interval is set. To align the start time The largest not greater than the first satellite navigation epoch time label Integer multiples of time, and construct an aligned time series. satisfy and It is a non-negative integer; Then, the angular velocity and acceleration data are resampled according to the aligned time series, at each aligned time point. Search for each The angular velocity aligned data is obtained by calculating a pair of adjacent angular velocity samples using piecewise linear interpolation. At the same time, the search satisfies Acceleration alignment data is obtained by calculating a pair of adjacent acceleration samples using piecewise linear interpolation. and will and Index by Time Combined to form synchronous inertial measurement unit data ; Finally, time alignment processing is performed on the satellite navigation observation data, and time labels for each satellite navigation epoch are generated. Seeking satisfaction The only alignment moment The carrier phase observations and Doppler observations of all satellites within this epoch are uniformly labeled as time stamps. (1. In step S1 of the instruction manual, a unified alignment time series is first defined.) ,in

[0025] Therefore, the symbol The definition is always the time value aligned with the time grid. 2. In the satellite navigation epoch time label When performing alignment, "find a match" The only alignment moment The meaning of "" is: in aligning time series Selecting from and The corresponding alignment time value. Because the alignment time series is constructed at fixed intervals and uses... The neighborhood criterion, therefore... Under normal sampling conditions, it is uniquely determined. 3. "The carrier phase observations and Doppler observations of all satellites within this epoch are uniformly labeled as time stamps." "Not bestowed" The new meaning is that the time labels of the aligned observation data are defined to satisfy:

[0026] That is, the aligned observation time label is the selected alignment time. .therefore, The terms "alignment time" and "aligned time stamp" are used interchangeably, differing only in their descriptive perspective; their numerical values ​​are consistent with their definitions and do not constitute a misuse or ambiguity of symbols. Simultaneously, Doppler observations are used to perform intraepoch time shift compensation on carrier phase observations to ensure that the carrier phase observations and Doppler observations are aligned. The compensation satisfies the following conditions: ; in Indicates that the time stamps are aligned to The satellite after The carrier phase observations, in units of cycles. Indicates time label Time Satellite The carrier phase observations, in units of cycles. Indicates time label Time Satellite Doppler observations and units Indicates the first in the aligned time series Each aligned time point is in seconds. Indicates the first The time stamp for each satellite navigation epoch is in seconds. Indicates satellite identifier, Indicates an aligned time series index. Represents the satellite navigation epoch index and the aligned carrier phase observations. Aligned Doppler observations according to The organization synchronizes satellite navigation observation data for subsequent tightly coupled calculations.

[0027] In this specific embodiment, S2 includes: Based on synchronous inertial measurement unit data Construct a continuous-time state trajectory and establish a Gaussian process prior model to output the inertial predicted state and its uncertainty; Set the trajectory time sequence within the time range of the synchronous inertial measurement unit data. ,in and And fix the time interval between adjacent trajectories to 0.1 s, where Indicates the first Each trajectory is measured in seconds. This represents the alignment time constructed in step S1, in seconds. Indicates and The corresponding alignment time index, Indicates the time index of the trajectory; Subsequently at each trajectory time Define the state variables of the continuous-time state trajectory ,in By location ,speed ,attitude And the composition of the inertial measurement unit error parameters, Indicates the carrier in the navigation coordinate system The position vector below, with units in meters. Indicates the carrier in the navigation coordinate system The velocity vector below and its unit is , Indicates from the carrier coordinate system To navigation coordinate system The unit quaternion attitude, the error parameter of the inertial measurement unit is determined by the gyroscope zero bias. Zero bias of accelerometer constitute, The unit is , The unit is and the navigation coordinate system Defined as a local northeast-sky coordinate system with its gravitational acceleration vector fixed as And the unit is ; Next, based on the data from the synchronous inertial measurement unit, inertial recursion is performed between adjacent trajectory times to obtain the state correlation between adjacent trajectory times. The inertial recursion is performed in each alignment time interval. Inner Perform median integral updates for the integral step size, where the angular velocity data is aligned. Data aligned with acceleration First, subtract the current zero-biased estimate separately. and The deflection angular velocity and deflection force are obtained, and then the attitude increment is calculated from the deflection angular velocity to update the quaternion. And the deflection force will be determined by attitude. Rotate to navigation coordinate system Later and The acceleration of the navigation system is obtained by summing the values ​​to update the velocity. Then by speed Update location At the same time and The model is a random walk process, and its mean is kept constant between adjacent trajectory times during the recursion. Based on the aforementioned state correlations, a Gaussian process prior model of continuous-time state trajectories is constructed, which will... arrive The inertial recursive result is defined as the inertial deterministic propagation function. And introduce process noise. To characterize the uncertainties caused by inertial noise and zero-biased random walks, a Gaussian process prior relationship is established between adjacent trajectory moments: ; in Representing trajectory time The state variable at that point, Representing trajectory time The state variable at that point, Indicates from the aligned index arrive Angular velocity aligned data and acceleration aligned data sequences, This represents the deterministic propagation function of inertia achieved according to the above-mentioned mean integral and coordinate rotation rules. Indicates from arrive Discrete process noise vector, This indicates that the mean is zero and the covariance is... Gaussian distribution, Indicates and A zero vector with consistent dimensions. This represents the process noise covariance matrix obtained by discretizing the inertial noise parameters; The According to the alignment sampling interval Discretize and construct the noise power spectral density in the interval The power spectral density of the gyroscope white noise is obtained by gradual accumulation within the internal system, where the power spectral density of the gyroscope white noise is set to... and The accelerometer white noise power spectral density is set to and The power spectral density of the gyroscope's zero-bias random walk is set to and The accelerometer zero-bias random walk power spectral density is set to and and make In attitude, speed, position, The corresponding sub-blocks respectively reflect the cumulative variance of white noise after integral propagation and zero-bias random walk; Then, given the initial state and its initial covariance Starting from the trajectory time sequence Forward propagation of the prior relations of the above Gaussian process yields the inertial prediction state at each trajectory time step. and the corresponding uncertainty ,in This represents the state mean obtained solely from data recursively derived from the synchronous inertial measurement unit (SMU) data. This represents the uncertainty covariance matrix of the state mean, and... Output as prior trajectory information.

[0028] In this specific embodiment, S3 includes: Based on prior trajectory information With synchronous satellite navigation observation data Construct a tightly coupled observation model and formulate a joint estimation problem within a sliding time window; First, set the length of the sliding time window. And record the alignment time corresponding to the current output epoch as Select the time that satisfies the conditions within the sliding time window. Furthermore, there exists a set of epoch indices for synchronous satellite navigation observation data. ,in Indicates the alignment time in seconds. Indicates the alignment time index. This indicates the current epoch-aligned time in seconds. Indicates the current output epoch index. This indicates the duration of the sliding time window, in seconds. This represents the set of epoch indices within the sliding time window; Then select the coverage within the sliding time window. The set of trajectory moments ,in The trajectory time is expressed in seconds. For trajectory time index, The set of trajectory time indices within the sliding time window is defined, and the variable to be estimated is defined as the state variable at each trajectory time within the sliding time window. Carrier phase integer ambiguity of each available satellite within the sliding time window ,in Consistent with step S2 and includes location ,speed ,attitude gyroscope zero bias Zero bias of accelerometer For satellite identification, The ambiguity is expressed in whole weeks and remains constant within the sliding time window. Next, based on the prior trajectory information, continuous time state trajectory prediction values ​​are constructed for each epoch within the sliding time window, and each observation epoch... Locate the time of its adjacent trajectories And according to the median integral propagation rule consistent with inertial recursion in step S2, by The propagation of the era State prediction at the location ,in Represents the continuous-time state trajectory at time t. The evaluation result at that epoch includes the predicted position for that epoch. Prediction speed With predicted posture The discrete summation result of the process noise covariance is used as the uncertainty of the propagation for subsequent weighting. Then, combining the data from each satellite in the geostationary satellite navigation observation data... The ephemeris information is used to construct a tightly coupled observation model, specifically in each epoch. The satellite's position and velocity at the time of reception are calculated using ephemeris data, and corrections are made for Earth rotation, relativistic effects, and satellite clock errors. Calculate geometric distance and by Calculate the relative radial velocity along the line of sight, and then combine the carrier phase observations. Compared with Doppler observations Each is compared with its corresponding prediction to form the carrier phase residual. With Doppler residuals ,in The carrier phase observations after time alignment in step S1 are in cycles. The Doppler observations after time alignment in step S1 are in units of This represents the carrier phase observation residual, expressed in cycles. This represents the Doppler observation residual in Hz, and explicitly incorporates the corresponding satellite's carrier phase integer ambiguity into the carrier phase prediction. To ensure that the unknowns in the carrier phase observation equation include at least the state variables of the continuous-time state trajectory and the carrier phase integer ambiguity, and to ensure that the unknowns in the Doppler prediction include at least the state variables of the continuous-time state trajectory. Subsequently, within the sliding time window, the trajectory prior constraints, carrier phase observation equation constraints, and Doppler observation equation constraints are combined to form a joint estimation problem, which is then solved using Gauss-Newton iteration. The initial values ​​for the iteration are set as follows: and The number of iterations is fixed at 5, and each iteration constructs a normal equation by linearizing the residual function and solves it using Cholesky decomposition for incremental updates. and ,in Representing trajectory time In the state of inertial prediction; The joint estimation problem aims to minimize the objective function. Implemented in the form and written as: ; in This represents the minimization operation with the variable to be estimated as the independent variable. This indicates the joint estimation of the objective function. Represents the epoch The available satellite set, Let the prior residual vector of the trajectory be defined as follows: and Representing the time intervals of adjacent trajectories respectively and The state variable at that point, This represents the inertial deterministic propagation function recursively obtained from the data of the synchronous inertial measurement unit in step S2. and These represent angular velocity aligned data and acceleration aligned data, respectively. Indicates from arrive The process noise covariance matrix, This indicates the transpose operation. This represents the matrix inversion operation. and Representing epochs satellite The carrier phase observation residual and the Doppler observation residual, This represents the standard deviation of carrier phase observation noise, taken as 0.01 cycles. This represents the standard deviation of Doppler observation noise, taken as 0.2 Hz. After completing the iterative solution, the last epoch within the sliding time window will be... place With all As the initial fusion solution output, and based on this initial fusion solution, the results are applied to each epoch within the sliding time window. and each available satellite Recalculate and And generate the observation residual sequence in chronological order.

[0029] In this specific embodiment, S4 includes: Based on the observed residual sequence, change point detection is performed on the carrier phase residual to determine cycle slips and output cycle slip processing information; First, sort the observed residual sequence by satellite identifier. Extracting carrier phase residuals And indexed by alignment time The carrier phase residual sequence of the satellite is formed, in which This indicates that after the sliding time window tightly coupled joint estimation is completed, at the epoch... Satellites calculated at the location Carrier phase observation residuals in cycles. This indicates the alignment time in step S1, in seconds. Indicates the alignment time index. Indicates satellite identifier; Subsequently, each one in the epoch Available satellites At each epoch, a change point determination based on the generalized likelihood ratio test is performed. The change point determination models "no cycle slip" as the carrier phase residual satisfies a zero-mean Gaussian distribution and has stable statistical characteristics within a short time window, and models "cycle slip" as the carrier phase residual experiences a sudden change in mean at the current epoch, causing the residual to deviate significantly from the existing statistical center. The normalized residual change amount is used as the cycle slip determination amount. To ensure that the decision quantity can be calculated online and is consistent with the observed noise level, in each epoch... For each satellite Maintenance length is Historical residual cache, cache content is And all of them are from before the current epoch. One valid epoch, when the historical residual cache length is insufficient. When the residual standard deviation is fixed, take The historical residual mean was fixed at 0 weeks. When the historical residual cache length reaches At that time, calculate the historical residual mean. Standard deviation of historical residuals ,in Indicates by The calculated mean, in weeks. Indicates by The calculated standard deviation is given in weeks and is in epochs. Calculate the cycle slip determination value: ; in Represents the epoch satellite The cycle slip determination quantity is dimensionless. Represents the epoch satellite The carrier phase residual, in units of cycles. Represents the epoch Previously based on recent The mean value calculated from historical residuals, with units of weeks. Represents the epoch Previously based on recent The standard deviation is calculated from the historical residuals, with units of weeks. Indicates the alignment time index. Indicates satellite identifier, This indicates the length of the historical residual cache and is set to 10. If and only if and At the same time, the satellite was determined. In the calendar Cycle slips occur and cycle slip markers are generated, where This indicates the cycle slip threshold, which is set to 25. The cycle slip marker includes the satellite identifier. Epicycle jump ; Simultaneously with generating cycle slip markers, a fuzziness reset instruction corresponding to each cycle slip marker is generated. The fuzziness reset instruction specifies that the object to be reset is a satellite. Carrier phase integer ambiguity And reset the trigger epoch to The set of cycle slip markers generated by all satellites at each epoch and the set of ambiguity reset instructions are encapsulated together as cycle slip processing information and the carrier phase cycle slip detection results are output.

[0030] In this specific embodiment, S5 includes: Based on the cycle slip processing information, the carrier phase integer ambiguity of the satellite with cycle slip marking is reset and the initial values ​​for reinitialization solution and reinitialization constraint information are generated. First, the cycle slip marker set is read from the cycle slip processing information. The cycle slip markers are identified by satellites. Alignment time corresponding to the epoch of the cycle slip Composition, in which Satellite identifiers indicating cycle slips. Indicates the alignment time of the cycle slip, in seconds. Indicates the alignment time index of the cycle jump epoch; The cycle jump epoch was then located in the initial fusion solution results. The corresponding continuous-time state trajectory evaluation results and locate the contents trajectory moments Make ,in This indicates that step S3 occurs at time [time]. The result of calculating the state trajectory at that point. The time interval of the trajectory is represented in seconds. Represents the trajectory time index associated with the cycle jump epoch; Then, based on the ambiguity reset instructions in the cycle slip processing information, each satellite that has a cycle slip marker is processed. Perform carrier phase integer ambiguity Reset, the reset including resetting the satellite The old ambiguity variables are removed from the set of variables to be estimated in the sliding time window joint estimation and new ambiguity variables are created. This is used for subsequent update estimations, while retaining the ambiguity variables corresponding to satellites that have not undergone cycle slip markings to maintain the continuity of the fixed solution; Then, according to step S3, in the cycle jump epoch... The initial fusion solution result determines the initial range of the carrier phase integer ambiguity after reset, specifically in epochs. Read the time-aligned carrier phase observations and combined The carrier phase prediction value, which does not contain ambiguity terms, is calculated from the ephemeris information. The difference between the two is then used to obtain the initial floating-point value for resetting the ambiguity. And set the initial value for resolving ambiguity to be... At the same time, the initial value range for resetting ambiguity is set to... And the unit is weeks, of which Indicates alignment time satellite The carrier phase observations, in units of cycles. This indicates the satellites obtained by back-calculation from the initial fusion solution results. The initial floating-point value for ambiguity is in weeks. This represents the initial integer value of the ambiguity used for iterative solving after resetting, with the unit being "round". This indicates the rounding operation; Subsequently, re-initialization constraints are constructed based on the inertial prediction state and its uncertainty to suppress the solution jump caused by cycle slip reset. Specifically, at trajectory time... Introducing inertial prediction state The pseudo-observation constraint centered on the trajectory state variable is applied. ,in Indicates the trajectory time obtained from step S2 propagation. The mean value of the inertial prediction state. Representing trajectory time The state variables to be estimated include position, velocity, attitude, and inertial measurement unit error parameters. The uncertainty is calculated using the covariance obtained from propagation in step S2. And through scaling factor Adjust the constraint strength; The scaling factor The constraint weight on the inertial prediction state is determined jointly by cycle slip events and the number of available satellites, and is satisfied by increasing the constraint weight when a cycle slip is detected or the number of available satellites decreases. Specifically, the constraint weight is set such that when the cycle slip processing information is in the epoch... When the cycle slip marker contains at least one satellite, take And the number of satellites available at that epoch. If less than 6, further take ,in Represents the epoch Available satellite set The number of elements; To ensure the differentiability of attitude constraints and consistency with sliding window iterative solutions, attitude deviations are represented by a three-dimensional vector of small-angle errors, which is derived from... attitude quaternions and The attitude quaternion is calculated and, together with the position, velocity, and zero bias, forms the reinitialized constraint residual vector; Finally, the reinitialization constraints are written into the objective function of the subsequent update estimate in a weighted quadratic form, and the reinitialization constraint information is generated. ),in This indicates that the constraint covariance matrix is ​​reinitialized and is determined by... Construct and use to Increasing the constraint weights is converted into increasing the initial ambiguity values, while simultaneously resetting the initial ambiguity values. and its initial value range and trajectory state initial value The re-initialization constraints are collectively encapsulated as initial values ​​for re-initialization calculation, which can be directly used as the initial state for iteration in step S6. The cost term of the re-initialization constraints in the objective function is expressed as follows: ; in Representing trajectory time The cost term corresponding to the re-initialization constraint. Denotes the reinitialized constraint residual vector and is defined as follows: Representing trajectory time The state variable to be estimated, Representing trajectory time The mean value of the inertial prediction state. This indicates that the constraint covariance matrix has been reinitialized. This indicates the transpose operation. This represents the matrix inversion operation. This represents the trajectory time index associated with the cycle jump epoch.

[0031] In this specific embodiment, S6 includes: Trajectory prior information within the sliding time window Synchronous satellite navigation observation data The initial values ​​and constraints of the re-initialization solution are used together to update the estimated continuous-time state trajectory and carrier phase integer ambiguity and output the enhanced navigation solution results. First, using the sliding time window setting from step S3, set the alignment time corresponding to the current output epoch to be... And the length of the sliding time window is Construct an epoch index set and the set of trajectory time indices ,in This indicates the alignment time in step S1, in seconds. Indicates the alignment time index. This indicates the alignment time of the current output epoch, in seconds. Indicates the current output epoch index. The time interval of the trajectory is represented in seconds. Indicates the time index of the trajectory. This indicates the duration of the sliding time window, in seconds. and These represent the epoch index set and the trajectory time index set within the sliding time window, respectively. Then, a set of variables to be estimated is constructed. ,in For trajectory moments The state variable at the location is consistent with step S2 and includes the position. ,speed ,attitude gyroscope zero bias Zero bias of accelerometer Satellite identifier The carrier phase integer ambiguity, in units of cycles. Indicates satellite identifier; Then, the re-initialized initial values ​​are used as the initial state for iteration, where satellites that have experienced cycle slips and have been reset are included. Will Initialize to the value given in step S5 For satellites that did not have cycle slip markers Will Initialize the values ​​to the estimated values ​​within the current sliding time window retained from the previous solution in step S3, for all trajectory time points. Will Initialize to inertial prediction state ,in This represents the initial ambiguity value after reset, in weeks. Representing trajectory time In the state of inertial prediction; Next, re-initialization constraint information is added to the joint estimation problem. This re-initialization constraint information includes the trajectory times at which the re-initialization constraints are applied. The inertial prediction state at this trajectory moment and reinitializing the constraint covariance matrix ,in This represents the trajectory time associated with the cycle jump epoch, in seconds. This indicates the time index of the trajectory. This indicates that the constraint covariance matrix is ​​reinitialized and determined by step S5 based on the inertial prediction uncertainty; Subsequently, Gaussian process prior constraints are applied to the continuous-time state trajectory within the sliding time window, and carrier phase observation equation constraints and Doppler observation equation constraints are applied to the observations. Specifically, for each pair of adjacent trajectory times... Constructing the prior residual of the trajectory: ; in This represents the trajectory prior residual vector. and Representing the trajectory time respectively and The state variable at that point, This represents the inertial deterministic propagation function recursively obtained from the data of the synchronous inertial measurement unit in step S2. and These represent angular velocity aligned data and acceleration aligned data, respectively. For each epoch With the set of satellites available at this epoch Each satellite in Calculate carrier phase observation residuals Residuals from Doppler observations ,in Represents the epoch The available satellite set, This represents the carrier phase observation residual, expressed in cycles. The unknowns in the Doppler observation residuals, expressed in Hz, while maintaining the carrier phase, include at least the following: and Furthermore, the unknowns in the Doppler observation equations include at least... ,in This indicates that the state trajectory in continuous time is at time 1000. The state that can be achieved by seeking value; Based on this, an update objective function is constructed and solved, which includes trajectory prior terms, carrier phase observation terms, Doppler observation terms, and re-initialization constraint terms. The update objective function is written as: ; in This represents the minimization operation. This represents the objective function for updating the estimate. This indicates that the process constructed in step S2 is from... arrive The process noise covariance matrix, This indicates the transpose operation. This represents the matrix inversion operation. This represents the standard deviation of carrier phase observation noise, taken as 0.01 cycles. Indicates the standard deviation of Doppler observation noise and takes Denotes the reinitialized constraint residual vector and is defined as follows: Representing trajectory time The state variable to be estimated, Representing trajectory time The mean value of the inertial prediction state. This indicates the re-initialization of the constraint covariance matrix; The objective function is solved using Gauss-Newton iteration, with a fixed number of iterations of 5. In each iteration, all residual terms are linearized to the first order and a normal equation is constructed. The state increment is solved by Cholesky decomposition and then updated synchronously. and And after each iteration, the value is recalculated based on the updated trajectory state. To update and Until the iteration ends; After iterative convergence, the updated continuous-time state trajectory will be displayed in the output epoch. The value is obtained by performing the calculation at the location. and from Read location information Speed ​​information With attitude information As the output of the enhanced navigation solution, This indicates that the output epoch position vector is in meters. This indicates the output epoch velocity vector with units of 1. This indicates the output epoch attitude quaternion.

[0032] In this specific embodiment, The adaptive threshold update mechanism is that the preset judgment condition in step S4 no longer uses a fixed threshold. Instead, in each epoch For each satellite Online calculation of adaptive threshold Based on this, cycle slips are determined, which improves the robustness of cycle slip determination and reduces the probability of false detection when the uncertainty of the inertial prediction state increases. Specifically, regarding the epoch First, determine the trajectory time closest to its time. And read the covariance of inertial prediction uncertainty ,from Extract the position corresponding to Sub-block covariance matrix And calculate the scalar uncertainty of the inertial prediction state. for The sum of the diagonal elements, the The unit is It also characterizes the magnitude of the uncertainty in the inertial prediction position near that epoch; Meanwhile targeting satellites Read the value used in step S4 to calculate the cycle slip determination quantity. Historical residual standard deviation ,in From length of Historical carrier phase residual buffer The result is calculated and the unit is weeks; Based on this, the adaptive threshold is calculated using the following formula: ; in Represents the epoch satellite The corresponding cycle slip determination adaptive threshold is a dimensionless quantity. Indicates the basic threshold and takes Represents the adjustment coefficient for inertial uncertainty and takes Represents the epoch The corresponding inertial prediction position uncertainty scalar and its unit is , The uncertainty is normalized reference quantity and is taken as Represents the residual variance adjustment coefficient and takes Represents the epoch Previous satellite The historical carrier phase residual standard deviation in weeks. This represents the standard deviation of carrier phase observation noise, consistent with step S3, and taken as 0.01 cycles. Indicates the alignment time index. Indicates satellite identifier; To avoid false negatives due to abnormally amplified thresholds, after calculation... Then limit its amplitude to a range And after the limit As the preset threshold for the judgment condition in step S4, the cycle slip judgment condition is thus determined as follows. and Simultaneously, the satellite was determined. In the calendar A cycle slip occurs, and causes the current Increase or When it increases The synchronous increase is used to suppress false triggering caused by deterioration in observation quality.

[0033] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0034] This invention organically combines a continuous-time trajectory Gaussian process prior, a tightly coupled sliding window joint estimation based on carrier phase and Doppler, and cycle slip detection and constraint re-initialization based on residual sequence changes into an integrated closed-loop process: First, a priori constraints that can be continuously evaluated over time are established using the continuous-time state trajectory driven by inertial measurement unit data, so that the state still has propagable predictive ability and uncertainty characterization when the number of available satellites decreases, the observation geometry deteriorates, or high-dynamic maneuvers occur; then, within the sliding time window, the trajectory prior, carrier phase observation equation, and Doppler observation equation are jointly estimated in a tightly coupled manner to obtain the fused solution result and form a residual sequence; then, change point detection is performed on the carrier phase residual to quickly determine cycle slips and trigger the corresponding satellite ambiguity reset and re-initialization, so that the state degradation caused by cycle slips can be isolated and repaired in a timely manner, thereby improving convergence and fixation capabilities and reducing output discontinuities or jumps caused by switching between floating-point solutions and fixed solutions.

[0035] To address the aforementioned technical issues, this invention further introduces an improved mechanism for "smoothness recovery of occlusion and cycle slip" in its algorithm structure: First, cycle slip determination is based on the residual sequence generated by tightly coupled joint estimation. Change point detection characterizes abrupt changes in the statistical properties of the residuals, improving the robustness of cycle slip detection in complex environments. Second, instead of simply restarting the solution after ambiguity reset, re-initialization constraints are constructed using the inertial prediction state and its uncertainty. The determination threshold and constraint weights can be adaptively adjusted according to the uncertainty, enabling the re-initialization process to both quickly recover the solution and apply controllable constraints to the state changes of the re-initialization epoch. This more effectively suppresses solution jumps and improves the continuity and stability of navigation output.

Claims

1. A satellite navigation enhancement method based on inertial navigation data, characterized in that, include: S1. Collect inertial measurement unit (IMU) data and satellite navigation observation data, and perform time alignment processing. The satellite navigation observation data includes carrier phase observations. S2. Construct a continuous-time state trajectory based on the IMU data, establish a Gaussian process prior model for the continuous-time state trajectory, generate the inertial prediction state and its uncertainty, and obtain trajectory prior information. S3. Construct a tightly coupled observation model based on the trajectory prior information and satellite navigation observation data, construct and solve the joint estimation problem for the continuous-time state trajectory and carrier phase integer ambiguity within a sliding time window, obtain the initial fusion solution result, and calculate the observation residual sequence. S4. Based on the observation residual sequence, perform change point detection on the residual corresponding to the carrier phase observation value to determine cycle slip, output the carrier phase cycle slip detection result, and generate cycle slip processing information, including cycle slip mark and corresponding ambiguity reset instruction. S5. Based on the cycle slip processing information, reset the carrier phase integer ambiguity corresponding to the satellite with the cycle slip mark according to the ambiguity reset command, and construct re-initialization constraints based on the initial fusion solution results, inertial prediction state and its uncertainty, and generate re-initialization initial values ​​and re-initialization constraint information; S6. Within the sliding time window, update and estimate the continuous time state trajectory and carrier phase integer ambiguity based on trajectory prior information, satellite navigation observation data, re-initialization initial values ​​and re-initialization constraint information, and output the enhanced navigation solution results.

2. The satellite navigation enhancement method based on inertial navigation data according to claim 1, characterized in that, S1 includes: Record corresponding time labels for angular velocity data, acceleration data, and satellite navigation observation data respectively; A unified alignment time series is determined, and the angular velocity data and acceleration data are resampled according to the alignment time series to obtain angular velocity aligned data and acceleration aligned data; Time alignment processing is performed on satellite navigation observation data to ensure that the time labels of carrier phase observations and Doppler observations are consistent with the aligned time series. The angular velocity aligned data and the acceleration aligned data are combined to form synchronous inertial measurement unit data, and the time-aligned satellite navigation observation data is determined as synchronous satellite navigation observation data.

3. The satellite navigation enhancement method based on inertial navigation data according to claim 2, characterized in that, S2 include: Multiple trajectory moments are set within the time range corresponding to the synchronous inertial measurement unit data, and state variables of the continuous time state trajectory are defined at each trajectory moment. The state variables include position, velocity, attitude, and inertial measurement unit error parameters. Based on the data from the synchronous inertial measurement unit, inertial recursion is performed between adjacent trajectory times to obtain the state correlation between adjacent trajectory times; Based on the state correlation, a Gaussian process prior model of the continuous-time state trajectory is constructed, and the inertial prediction state and its uncertainty corresponding to each trajectory time are calculated by the Gaussian process prior model. The inertial prediction state and its uncertainty are determined as trajectory prior information.

4. The satellite navigation enhancement method based on inertial navigation data according to claim 2, characterized in that, S3 includes: Based on the prior trajectory information, the predicted value of the continuous time state trajectory at each moment within the sliding time window is determined, and a tightly coupled observation model is constructed by combining the ephemeris information of the corresponding satellite in the synchronous satellite navigation observation data. The tightly coupled observation model includes a carrier phase observation equation corresponding to the carrier phase observation value and a Doppler observation equation corresponding to the Doppler observation value. The unknowns of the carrier phase observation equation include at least the state variables of the continuous-time state trajectory and the carrier phase integer ambiguity. The unknowns of the Doppler observation equation include at least the state variables of the continuous-time state trajectory. Within the sliding time window, the trajectory prior information, the carrier phase observation equation, and the Doppler observation equation are combined to form a joint estimation problem, and the initial fusion solution is obtained by iterative solution. Based on the initial fusion solution results, the observation residuals corresponding to the carrier phase observations and the Doppler observations are calculated respectively, and the observation residual sequence is formed in chronological order.

5. The satellite navigation enhancement method based on inertial navigation data according to claim 1, characterized in that, S4 includes: Extract the carrier phase residual sequence corresponding to the carrier phase observation value from the observation residual sequence; The carrier phase residual sequence is subjected to change point determination. The change point determination adopts the Bayesian online change point detection algorithm or the generalized likelihood ratio test algorithm, and the cycle slip determination quantity is calculated at each epoch. When the cycle slip determination quantity meets the preset determination conditions, the corresponding epoch and the corresponding satellite are marked as cycle slip marks and the carrier phase cycle slip detection result is output. Cycle slip processing information is generated based on the cycle slip marker, and the cycle slip processing information includes the cycle slip marker and the ambiguity reset instruction corresponding to the cycle slip marker.

6. The satellite navigation enhancement method based on inertial navigation data according to claim 1, characterized in that, S5 include: The satellite corresponding to the cycle slip mark is read according to the cycle slip processing information, and the carrier phase integer ambiguity of the satellite corresponding to the cycle slip mark is reset according to the ambiguity reset instruction; The initial range of the reset carrier phase integer ambiguity is determined based on the initial fusion solution results; Based on the inertial prediction state and its uncertainty, a reinitialization constraint is constructed. The reinitialization constraint is used to constrain the deviation between the state variables of the continuous-time state trajectory and the inertial prediction state at the reinitialization epoch. The initial values ​​for reinitialization and the reinitialization constraint information are generated based on the initial value range and the reinitialization constraint conditions.

7. A satellite navigation enhancement method based on inertial navigation data according to claim 2, characterized in that, S6 include: Within the sliding time window, the re-initialized initial value is used as the initial state for iterative solution, and the re-initialization constraint information is added to the joint estimation problem; Gaussian process prior constraints are applied to the continuous time state trajectory based on the trajectory prior information, and carrier phase observation equation constraints and Doppler observation equation constraints are applied to the continuous time state trajectory and carrier phase integer ambiguity based on the synchronous satellite navigation observation data. The updated continuous-time state trajectory and the updated carrier phase integer ambiguity are obtained by iterative solution. The updated continuous-time state trajectory is then evaluated at the output epoch to obtain the enhanced navigation solution result. The enhanced navigation solution results include position information, velocity information, and attitude information.

8. A satellite navigation enhancement method based on inertial navigation data according to claim 5, characterized in that, The preset determination condition for the change point determination is an adaptive threshold, which is updated online based on at least one of the following: the uncertainty of the inertial prediction state and the variance of the carrier phase residual sequence, so that the robustness of the cycle slip determination is improved when the uncertainty of the inertial prediction state increases.

9. A satellite navigation enhancement method based on inertial navigation data according to claim 6, characterized in that, The weights or covariance of the reinitialization constraints are determined by the inertial prediction state and its uncertainty. When a cycle slip or a decrease in the number of available satellites is detected, the constraint weights on the inertial prediction state are increased to suppress the solution jumps caused by reinitialization.

10. A satellite navigation enhancement system based on inertial navigation data, used to execute a satellite navigation enhancement method based on inertial navigation data as described in any one of claims 1 to 9, comprising: The data alignment module is used to collect inertial measurement unit data and satellite navigation observation data, and perform time alignment. The trajectory prior module is used to construct a continuous-time state trajectory based on the data from the inertial measurement unit and establish a Gaussian process prior model, outputting the inertial prediction state and its uncertainty as trajectory prior information. The tightly coupled solution module is used to construct a joint estimation problem based on the trajectory prior information and the satellite navigation observation data within a sliding time window, solve the ambiguity of the continuous time state trajectory and carrier phase integer, obtain the initial fusion solution result and generate the observation residual sequence; The cycle slip detection module is used to detect change points in the carrier phase residual based on the observation residual sequence to determine cycle slips and generate cycle slip processing information including cycle slip markers and ambiguity reset instructions. The reinitialization module is used to reset the carrier phase integer ambiguity corresponding to the satellite that has a cycle slip mark according to the cycle slip processing information, and to generate reinitialization initial values ​​and reinitialization constraint information based on the initial fusion solution results and the inertial prediction state and its uncertainty. The update output module is used to update the estimated continuous-time state trajectory and carrier phase integer ambiguity within the sliding time window based on the trajectory prior information, the satellite navigation observation data, the re-initialization solution initial value, and the re-initialization constraint information, and output the enhanced navigation solution result.

Images