A Pseudorange Gross Error Detection Method and System Based on Rigid Trajectory Constraints

Abstract

This invention discloses a pseudorange gross error detection method based on rigid trajectory constraints. The method includes: collecting pseudorange observations within a preset distance window and obtaining a rigid trajectory within the corresponding time period using dead reckoning; modeling the sources of rigid trajectory error as initial position error, initial heading error, and scale error; based on the pseudorange observations within the preset distance window, using the initial position, initial heading, and scale of the rigid trajectory as parameters to be estimated, solving for these parameters using an optimization method, and saving the pseudorange residuals after optimization convergence; based on the optimized pseudorange residuals, extracting residual consistency and continuity features for each satellite, and marking faulty satellites as normal satellites according to a preset threshold. This invention utilizes dead reckoning to obtain a rigid trajectory and uses the rigid trajectory to measure pseudorange gross error changes, thereby effectively identifying abrupt and gradually varying gross errors and significantly improving the detection success rate of pseudorange gross errors.

Description

Technical Field

[0001] This invention relates to the field of GNSS positioning, and specifically to a pseudorange gross error detection method and system based on rigid trajectory constraints. Background Technology

[0002] Global Navigation Satellite System (GNSS) signals are radio signals. Due to the characteristics of radio signals, they are susceptible to ionospheric delay, multipath effects, and non-line-of-sight (NLOS) propagation. In densely built-up urban areas, GNSS signals are frequently blocked or reflected, resulting in multipath effects and NLOS signals. Currently, in most GNSS applications, gross errors in GNSS observations caused by multipath effects and NLOS signals are considered the main factors affecting GNSS positioning performance. Existing typical robust GNSS algorithms are mainly classified as follows:

[0003] (1) Residual consistency test method

[0004] The fundamental assumption behind consistency checks is that observations containing multipath or NLOS signals are inconsistent with other normal observations, which can be reflected in the residuals after localization. Residual-based consistency checks utilize redundant observation information, eliminating one or more observations based on the observation residuals to find the most reliable localization solution. Typical consistency check schemes include Receiver Autonomous Integrity Monitoring (RAIM) and Random Sample Consensus (RANSAC) algorithms. However, in urban scenarios with a high proportion of gross errors, these methods struggle to guarantee the reliability of the localization solution and cannot effectively detect gradually varying gross errors.

[0005] (2) Chi-square test of filtered innovation

[0006] The chi-square test based on filtered information typically performs fault detection directly based on the information from the Kalman filter in GNSS / INS integrated navigation or multi-source fusion navigation. It has advantages such as simplicity, low computational cost, and the ability to quickly identify abrupt errors. However, compared to abrupt GNSS gross errors, gradually varying gross errors are more challenging. Such methods usually require sufficient time to accumulate before reaching the judgment threshold, making them unsuitable for detecting gradually varying gross errors.

[0007] (3) Factor plot optimization robustness method

[0008] Unlike Kalman filtering, which recursively estimates the current state, Factor Graph Optimization (FGO) solves the problem by minimizing a cost function of the observation residuals within a window. In recent years, FGO-based GNSS positioning and fusion positioning have received widespread attention from academia and industry. FGO, through multiple iterations and the use of historical data within a window during state estimation, better uncovers the temporal correlation between measurements and the state; simultaneously, it employs a robust cost function to reduce the weight of gross observations, typically exhibiting better robustness compared to filtering frameworks. However, such methods still suffer from the following problems: first, they do not fully utilize the rigid constraints of trajectory extrapolation, making convergence difficult in the presence of gross errors; second, the computational cost increases threefold with the window length, making it difficult to adapt to the real-time processing requirements of low-cost hardware.

[0009] In summary, GNSS robustness (especially gradually varying gross errors) remains a serious challenge in complex electromagnetic propagation scenarios such as urban canyons. Summary of the Invention

[0010] To address the robustness issue of GNSS in complex scenarios, this invention provides a pseudorange gross error detection method and system based on rigid trajectory constraints. By utilizing rigid trajectories to measure changes in pseudorange gross errors, the success rate of pseudorange gross error detection is effectively improved.

[0011] According to one aspect of the present invention, the present invention provides a pseudorange gross error detection method based on rigid trajectory constraints. The pseudorange gross error detection method includes: collecting pseudorange observations within a preset distance window and obtaining a rigid trajectory within the corresponding time period using dead reckoning; modeling the sources of rigid trajectory error as initial position error, initial heading error, and scale error; based on the pseudorange observations within the preset distance window, using the initial position, initial heading, and scale of the rigid trajectory as parameters to be estimated, solving for the parameters to be estimated using an optimization method, and saving the pseudorange residuals after optimization convergence; based on the optimized pseudorange residuals, extracting residual consistency features and continuity features for each satellite, and marking faulty satellites and normal satellites according to a preset threshold.

[0012] Furthermore, the dead reckoning methods include inertial navigation mechanical arrangement, vehicle dead reckoning, pedestrian dead reckoning, visual odometry, laser odometry, visual inertial odometry, or laser inertial odometry.

[0013] Furthermore, based on the pseudorange observations within a preset distance window, the initial position, initial heading, and scale of the rigid trajectory are used as parameters to be estimated. An optimization method is used to solve for these parameters, and the pseudorange residuals after optimization convergence are saved. This includes: constructing pseudorange residuals based on the difference between the pseudorange observations within the preset distance window and the pseudorange predictions calculated from the rigid trajectory; minimizing the objective function by continuously adjusting the initial position, initial heading, and scale of the rigid trajectory; reducing the weight of gross observations in the objective function by combining a robust kernel function; obtaining a set of parameters that minimizes the objective function; and saving the pseudorange residuals after optimization convergence.

[0014] Furthermore, the optimization method includes the Gauss-Newton method, the Newton method, or the Levenberg-Marquardt method; the robust kernel function includes the Huber kernel function, the Cauchy kernel function, the Turkey kernel function, or the Geman-McClure kernel function.

[0015] Furthermore, based on the optimized pseudorange residuals, residual consistency and continuity features are extracted for each satellite. Faulty satellites are marked as normal satellites according to preset thresholds. This includes: calculating the residual fluctuation amplitude of each satellite to verify whether the satellite meets the residual consistency requirement based on the residual fluctuation amplitude and a first preset threshold; calculating the effective observation epoch ratio of each satellite to verify whether the satellite meets the residual continuity requirement based on the effective observation epoch ratio and a second preset threshold; and marking satellites that simultaneously meet both residual consistency and residual continuity as normal satellites, otherwise marking them as faulty satellites.

[0016] Further, the residual fluctuation amplitude of each satellite is calculated to verify whether the satellite meets the residual consistency criteria based on the residual fluctuation amplitude and a first preset threshold. This includes: obtaining the maximum and minimum pseudorange residuals of each satellite in all epochs within a preset distance window; calculating the difference between the maximum and minimum pseudorange residuals of each satellite; wherein the difference between the maximum and minimum pseudorange residuals is the residual fluctuation amplitude; comparing the residual fluctuation amplitude with the first preset threshold; if the residual fluctuation amplitude is less than or equal to the first preset threshold, the satellite meets the residual consistency criteria; if the residual fluctuation amplitude is greater than the first preset threshold, the satellite does not meet the residual consistency criteria.

[0017] Further, the effective observation epoch ratio of each satellite is calculated to verify whether the satellite meets the residual continuity requirement based on the effective observation epoch ratio and a second preset threshold. This includes: obtaining the number of consecutive effective pseudorange observation epochs and the total number of epochs for each satellite within the corresponding time period of a preset distance window; calculating the ratio of the number of consecutive effective pseudorange observation epochs to the total number of epochs; wherein the ratio of the number of consecutive effective pseudorange observation epochs to the total number of epochs is the effective observation epoch ratio; comparing the effective observation epoch ratio with the second preset threshold; if the effective observation epoch ratio is greater than or equal to the second preset threshold, the satellite meets the residual continuity requirement; if the effective observation epoch ratio is less than the second preset threshold, the satellite does not meet the residual continuity requirement.

[0018] Furthermore, the pseudorange gross error detection method also includes: dynamically adjusting the weight of pseudorange observations of each satellite participating in the fusion positioning based on the marking results of each satellite, and downweighting or removing faulty satellites.

[0019] According to one aspect of the present invention, the present invention provides a pseudorange gross error detection system based on rigid trajectory constraints, comprising: a rigid trajectory estimation module, which collects pseudorange observations within a preset distance window and obtains the rigid trajectory within the corresponding time period using dead reckoning; the sources of rigid trajectory error are modeled as initial position error, initial heading error, and scale error; a pseudorange residual optimization module, which, based on the pseudorange observations within the preset distance window, uses the initial position, initial heading, and scale of the rigid trajectory as parameters to be estimated, solves for the parameters to be estimated using an optimization method, and saves the pseudorange residuals after optimization convergence; and a satellite verification and marking module, which, based on the optimized pseudorange residuals, extracts residual consistency features and continuity features for each satellite, and verifies and marks faulty satellites and normal satellites according to a preset threshold.

[0020] According to one aspect of the present invention, the present invention provides a non-transitory computer-readable storage medium storing computer instructions that cause the computer to execute the pseudorange gross error detection method based on rigid trajectory constraints.

[0021] The above technical solution uses dead reckoning to obtain a rigid trajectory and uses the rigid trajectory to measure the changes in pseudorange gross errors, thereby achieving effective identification of abrupt and gradual gross errors and effectively improving the detection success rate of pseudorange gross errors.

[0022] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0023] (1) This invention fully exploits the characteristic that the gross error of the slowly varying pseudorange of the same satellite is difficult to keep consistent under the distance window perspective. By using dead reckoning to obtain the relative change of the gross error of the rigid trajectory measurement, the detection success rate of the gross error of pseudorange is effectively improved.

[0024] (2) This invention optimizes the rigid trajectory while extracting the pseudorange residual, giving full play to the high relative accuracy of the rigid trajectory, while reducing the dependence on the trajectory extrapolation performance and having good generalization ability.

[0025] (3) By modeling the sources of rigid trajectory error as initial position, initial heading and scale error, this invention effectively reduces the solution scale of the optimization problem and is computationally efficient. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart of a pseudorange gross error detection method based on rigid trajectory constraints, provided for an embodiment of the present invention.

[0028] Figure 2 The first sub-flowchart of a pseudorange gross error detection method based on rigid trajectory constraints provided in an embodiment of the present invention.

[0029] Figure 3 The second sub-flowchart of a pseudorange gross error detection method based on rigid trajectory constraints provided in an embodiment of the present invention.

[0030] Figure 4 The third sub-flowchart of a pseudorange gross error detection method based on rigid trajectory constraints provided in an embodiment of the present invention.

[0031] Figure 5 This is a schematic diagram illustrating the principle of a pseudorange gross error detection method based on rigid trajectory constraints, provided in an embodiment of the present invention.

[0032] Figure 6 The above is a flowchart of a pseudorange gross error detection method based on rigid trajectory constraints provided in an embodiment of the present invention.

[0033] Figure 7 This is a schematic diagram of a pseudorange gross error detection system based on rigid trajectory constraints, provided as an embodiment of the present invention. Detailed Implementation

[0034] The terms “comprising” and “having”, and any variations thereof, in the specification, claims, and accompanying drawings of this invention are intended to cover a non-exclusive inclusion, such as a process, method, system, product, or apparatus that includes a series of steps or units, not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0035] Please refer to the appendix for details. Figure 1 , 5 6. The present invention provides a pseudorange gross error detection method based on rigid trajectory constraints, and the carrier velocity estimation method includes steps S101-S107.

[0036] Step S101: Collect pseudorange observations within a preset distance window and use dead reckoning to obtain the rigid trajectory within the corresponding time period; the sources of rigid trajectory error are modeled as initial position error, initial heading error, and scale error.

[0037] In step S101, the pseudorange observation is the raw data directly measured by the GNSS receiver, reflecting the "apparent distance" between the satellite and the GNSS receiver, including gross errors caused by multipath / non-line-of-sight. The theoretically calculated value based on the rigid trajectory is the pseudorange prediction, which ideally does not contain gross errors. The difference between the pseudorange observation and the pseudorange prediction is the pseudorange residual. The preset distance window refers to the travel distance, which can be set according to actual conditions (e.g., 100m, 50m), and is not limited here. Dead reckoning methods include, but are not limited to, one of the following: inertial navigation mechanical orchestration, vehicle dead reckoning, pedestrian dead reckoning, visual odometry, laser odometry, visual inertial odometry, and laser inertial odometry. In this embodiment, the rigid trajectory calculated by the dead reckoning method does not require reliable absolute position, heading, and scale, because the pseudorange residual will be optimized by adjusting the initial position, initial heading, and scale of the rigid trajectory. Please refer to the following text for details; further elaboration is not provided here.

[0038] In this embodiment, the pseudorange observations (i.e., the pseudorange observation model) are represented as follows: In the formula, the subscript k represents the epoch number; the superscript k represents the epoch number. Indicates the j-th satellite; Indicates the geometric distance between the GNSS receiver and the satellite; The speed of light; This indicates satellite clock bias, which can be obtained and corrected through ephemeris. Indicates receiver clock bias; and These represent the ionospheric and tropospheric delays, respectively. The tropospheric delay can be corrected using the Hopfield model, and the ionospheric delay can be corrected using the Klobuchar model. This indicates gross errors caused by multipath / non-line-of-sight. This represents the observation noise, assumed to follow a zero-mean Gaussian distribution, i.e. Standard deviation of observed noise Represented using an elevation angle model: In the formula, The standard deviation of pseudorange noise in the a priori zenith direction, This indicates the elevation angle of the satellite observed by the receiver at the epoch.

[0039] Please refer to the appendix for details. Figure 2 The following section will further introduce the steps for obtaining the rigid trajectory within the corresponding time period using dead reckoning, and then calculating the pseudorange prediction value based on the rigid trajectory (including steps S1011-S1015).

[0040] Step S1011: Obtain the initial value of the rigid trajectory position sequence through dead reckoning. In this embodiment, the initial value of the rigid trajectory position sequence obtained through dead reckoning can be expressed as: In the formula, This represents the position coordinate vector calculated in the navigation coordinate system (n-frame); the subscript k indicates the epoch number. Simply put, This refers to the uncorrected position information directly output by dead reckoning. However, due to errors in dead reckoning (mainly from initial position error, initial heading error, and scale error), it is necessary to establish a rigid trajectory position error model to describe the impact of these errors on the true rigid trajectory position, thereby allowing for correction and optimization of the rigid trajectory position.

[0041] Optionally, depending on the characteristics of dead reckoning methods, the rigid trajectory position error model can be constructed based on one or more of the initial position error, initial heading error, and scale error. In this embodiment, initial position error, initial heading error, and scale error are all modeled. Understandably, for ordinary optimization algorithms (such as factor graph optimization), assuming the parameter to be estimated in a single epoch is n and the distance window length is m, the corresponding dimension of the state to be estimated is n*m. This invention completely models the sources of rigid trajectory error as initial position, velocity, heading, and sensor errors; that is, the parameter to be estimated in a single epoch is n, and the dimension of the state to be estimated is also n. Therefore, the computational complexity of this application is very small, significantly reducing the problem-solving scale, improving computational efficiency, and making subsequent pseudorange residual optimization easier to converge.

[0042] Step S1013: Construct a rigid trajectory position error model based on the initial position error, initial heading error, and scale error. In this embodiment, the construction of the rigid trajectory position error model based on the initial position error, initial heading error, and scale error is described as follows: In the formula, the symbol This represents the variable to be estimated or calculated. This represents the calculated position vector at time k (i.e., the position information after error correction). Represents an unknown initial position vector; Indicates an unknown scale parameter; Indicates an unknown course The corresponding rotation matrix can be written as: , The relative displacement vector representing a rigid trajectory: Understandably, the optimization variables to be determined for a rigid trajectory can be expressed as: In the formula, The optimization variables representing the rigid trajectory to be determined; Indicates the initial position for optimization; Indicates the scale parameter for optimization; This indicates the optimized initial heading.

[0043] Step S1015: The pseudorange prediction value is calculated based on the rigid trajectory position error model. In this embodiment, the pseudorange prediction value is expressed as: In the formula, This represents the position vector calculated by the rigid trajectory position error model. Represents the satellite's position vector; This represents the pseudorange prediction value, which is the distance calculated based on the dead reckoning trajectory position and the satellite position (i.e., the Euclidean distance between the two points). The speed of light; This indicates the receiver clock bias.

[0044] In one implementation, the receiver clock bias at time k It is modeled as a clock drift model, and the clock drift is modeled as a constant value in the short term: In the formula, and These represent the clock drift and receiver clock bias vector at the first epoch within the time period corresponding to the preset distance window, respectively (the receiver clock bias vector is determined based on the number of satellites involved in the solution). , Let represent the clock drift and receiver clock difference vectors at time k, respectively; This represents the time difference between the k-th epoch and the first epoch.

[0045] Step S103: Based on the pseudorange observations within the preset distance window, the initial position, initial heading, and scale of the rigid trajectory are used as parameters to be estimated. The parameters to be estimated are solved using an optimization method, and the pseudorange residuals after optimization convergence are saved.

[0046] In step S103, to optimize the pseudorange residuals described above, the optimization algorithms that can be used include, but are not limited to, the Gauss-Newton method, Newton's method, or the Levenberg-Marquardt method. Please refer to the appendix for further details. Figure 3 The following will further describe step S103 (including steps S1031-S1033).

[0047] Step S1031: Construct pseudorange residuals based on the difference between pseudorange observations within a preset distance window and pseudorange predictions calculated from the rigid trajectory.

[0048] In step S1031, the pseudorange residual represents the difference between the pseudorange observed value and the corresponding pseudorange predicted value. The standardized squared pseudorange residuals of all epochs within the time period corresponding to the preset distance window are accumulated to obtain the objective function of the optimization problem (i.e., the weighted sum of squares of the pseudorange residuals): ,in, Represents the squared pseudo-range residual after standardization; Represents pseudorange observations. This represents the pseudorange prediction value. The objective function of the optimization problem is expressed as: N represents the total number of epochs within the time period corresponding to the preset distance window; S(k) represents the number of pseudorange observations at time k (i.e., the number of satellites that perform pseudorange observations at time k).

[0049] Step S1033: By continuously adjusting the initial position, initial heading, and scale of the rigid trajectory, the objective function is minimized.

[0050] In step S1033, the optimization problem in this embodiment is to optimize the pseudorange residual by minimizing the objective function. Therefore, the optimization problem can be described as follows: In the formula, This represents the optimal parameters that minimize the objective function.

[0051] Furthermore, due to the pseudorange residual model It is constructed based on the rigid trajectory position error model, pseudorange observation model, and clock error / drift state model. Therefore, the system state to be determined in the optimization problem can be represented as: In the formula, Represents the system optimization variables. The variables to be optimized for the rigid trajectory are: initial position, initial heading, and scale. This represents the variables to be optimized for the GNSS receiver (i.e., the clock drift and receiver clock error vector in the first epoch within the time period corresponding to the preset distance window). In other words, the optimization problem can be simplified to finding the optimal parameters that minimize the objective function. .

[0052] Step S1035: Combine the robust kernel function to reduce the weight of gross observations in the objective function, find the set of parameters to be estimated that minimizes the objective function, and save the pseudo-range residuals after optimization convergence.

[0053] In step S1035, using a robust kernel function, and suppressing gross observation interference, a set of estimated parameters that minimizes the objective function is obtained. The corresponding pseudorange residuals are the pseudorange residuals after optimization convergence. Specifically, using a robust kernel function, the optimization problem can be further expressed as: Robust kernel functions include, but are not limited to, robust kernel functions such as the Huber kernel, Cauchy kernel, Turkey kernel, or Geman-McClure kernel.

[0054] The Huber kernel function will be used as an example below. The Huber kernel function is defined as follows: s represents the squared pseudo-range residual after standardization. In other words, when the squared standardized pseudorange residual of any epoch within the time period corresponding to the preset distance window is less than or equal to 1, it directly participates in the cumulative calculation of the pseudorange residual. When the weighted squared pseudorange residual of any epoch within the time period corresponding to the preset distance window is greater than 1, the weight of the pseudorange residual participating in the cumulative calculation of the weighted sum of squares is reduced.

[0055] Step S105: Based on the optimized pseudorange residuals, extract residual consistency features and continuity features for each satellite, and verify and mark faulty satellites and normal satellites according to preset thresholds.

[0056] In step S105, from the perspective of a distance window (e.g., 100 m), it is difficult to maintain relative consistency in the pseudorange gross error observations of the same satellite. This is because in complex urban environments, there are almost no completely unchanging reflection paths within a 100 m distance window. Rigid trajectories possess the characteristic of high relative accuracy and can be used to measure changes in pseudorange gross errors. The pseudorange residuals after the optimization problem converges in step S103 incorporate the relatively accurate rigid trajectories. Therefore, the changing trend of the pseudorange residuals can roughly reflect the actual changes in pseudorange gross errors. The characteristics of the pseudorange residual sequence changes can be extracted and verified, thereby achieving effective identification of pseudorange gross errors (including abrupt and gradually varying gross errors).

[0057] In this embodiment, residual consistency detection is used to verify both slowly varying gross errors and abruptly varying gross errors, and observation continuity detection is used to further verify slowly varying gross errors. Optionally, features reflecting residual consistency include, but are not limited to, residual fluctuation amplitude and residual standard deviation, while features reflecting observation continuity include, but are not limited to, the proportion of effective observation epochs. In this embodiment, residual fluctuation amplitude is used to reflect the residual consistency of the pseudorange residual sequence, and the proportion of effective observation epochs is used to reflect the observation continuity of the pseudorange residual sequence.

[0058] Please refer to the appendix for details. Figure 4 The following will further describe step S105 (including steps S1051-S1057).

[0059] Step S1051: Extract the residual fluctuation amplitude of each satellite to verify whether the satellite meets the residual consistency requirement. Specifically, obtain the maximum and minimum pseudorange residuals of each satellite in all epochs within the time period corresponding to the preset distance window. Calculate the difference between the maximum and minimum pseudorange residuals of each satellite; where the difference between the maximum and minimum pseudorange residuals is the residual fluctuation amplitude. Compare the residual fluctuation amplitude with a first preset threshold; if the residual fluctuation amplitude is less than or equal to the first preset threshold, the satellite meets the residual consistency requirement; if the residual fluctuation amplitude is greater than the first preset threshold, the satellite does not meet the residual consistency requirement.

[0060] In this embodiment, verifying whether a satellite meets the residual consistency standard based on the residual fluctuation amplitude is expressed as follows: In the formula, Indicates the residual fluctuation amplitude; This indicates the time k corresponding to the optimization convergence in step S103. The pseudorange residuals of the satellites; This indicates the first preset threshold.

[0061] Step S1053: Calculate the effective observation epoch ratio for each satellite to verify whether the satellite meets the residual continuity requirement. Specifically, obtain the number of consecutive effective pseudorange observation epochs and the total number of epochs for each satellite within the corresponding time period of a preset distance window. Calculate the ratio of the number of consecutive effective pseudorange observation epochs to the total number of epochs; wherein, the ratio of the number of consecutive effective pseudorange observation epochs to the total number of epochs is the effective observation epoch ratio. Compare the effective observation epoch ratio with a second preset threshold; if the effective observation epoch ratio is greater than or equal to the second preset threshold, the satellite meets the residual continuity requirement; if the effective observation epoch ratio is less than the second preset threshold, the satellite does not meet the residual continuity requirement.

[0062] In this embodiment, the verification of whether the satellite meets the residual continuity requirement based on the proportion of effective observation epochs is expressed as follows: In the formula, This indicates the percentage of valid observation epochs; and These represent the number of consecutive valid pseudorange observation epochs and the total number of epochs for a single satellite within the time period corresponding to the preset distance window, respectively. This represents the second preset threshold. Understandably, intermittent pseudorange observations (i.e., the alternation of valid and invalid epochs, resulting in a low proportion of valid observation epochs) are likely gross errors.

[0063] In step S1055, satellites that simultaneously meet the requirements of residual consistency and residual continuity are marked as normal satellites; otherwise, they are marked as faulty satellites. Furthermore, based on the marking results of each satellite, the weights of the pseudorange observations corresponding to each satellite participating in the fusion positioning can be adjusted, and faulty satellites can be downweighted or removed. This can significantly suppress the destructive effects of gradually varying and abruptly changing gross errors on GNSS positioning and its fusion positioning performance.

[0064] Optionally, in the fusion positioning process after gross error detection, the sensors used may include, in addition to GNSS receivers, one or more of the following: gyroscopes, accelerometers, magnetometers, barometers, cameras, and lidar.

[0065] Optionally, the data fusion algorithms used in the fusion localization process include, but are not limited to: Kalman filtering and its improved algorithms, particle filtering and its improved algorithms, and graph optimization algorithms.

[0066] Optionally, the application scenarios of the pseudorange gross error detection method based on rigid trajectory constraints provided by the present invention include, but are not limited to, scenarios where GNSS positioning can be used, such as pedestrian scenarios and vehicle scenarios.

[0067] Please refer to the appendix for details. Figure 7 Based on the same inventive concept as the aforementioned method embodiments, this embodiment of the invention also provides a pseudorange gross error detection system 100 based on rigid trajectory constraints, including a rigid trajectory estimation module 1, a pseudorange residual optimization module 2, and a satellite verification and marking module 3. The rigid trajectory estimation module 1 collects pseudorange observations within a preset distance window and obtains the rigid trajectory within the corresponding time period using dead reckoning; the sources of rigid trajectory error are modeled as initial position error, initial heading error, and scale error. The pseudorange residual optimization module 2, based on the pseudorange observations within the preset distance window, uses the initial position, initial heading, and scale of the rigid trajectory as parameters to be estimated, solves for the parameters using optimization methods, and saves the optimized and converged pseudorange residuals. The satellite verification and marking module 3, based on the optimized pseudorange residuals, extracts residual consistency features and continuity features for each satellite, and verifies and marks faulty satellites and normal satellites according to a preset threshold.

[0068] The pseudorange gross error detection system 100 based on rigid trajectory constraints provided in this invention addresses the GNSS robustness problem in complex electromagnetic propagation scenarios such as urban canyons. Figure 7 Several modules within the system fully exploit the characteristic that gross errors in slowly varying pseudoranges of the same satellite are difficult to maintain consistency under the range window perspective. By utilizing the relative changes in gross errors of rigid trajectory measurements obtained through dead reckoning, the success rate of pseudorange gross error detection is effectively improved. Simultaneously optimizing the rigid trajectory while extracting pseudorange residuals fully leverages the high relative accuracy of the rigid trajectory, while reducing dependence on trajectory reckoning performance and exhibiting good generalization ability. Furthermore, by completely modeling the sources of rigid trajectory errors as initial position, initial heading, and scale errors, the solution scale of the optimization problem is effectively reduced, resulting in computational efficiency.

[0069] It should be noted that the system embodiments provided by the present invention are used not only to implement the methods in the above method embodiments, but also to implement the methods in other method embodiments provided by the present invention. The only difference is that corresponding functional modules are set. The principle is basically the same as that of the above system embodiments provided by the present invention. As long as those skilled in the art can improve the modules in the above system embodiments by referring to the specific technical solutions in other method embodiments and combining technical features to obtain corresponding technical means and technical solutions composed of these technical means, on the basis of the above system embodiments, and on the premise of ensuring the practicality of the technical solutions, they can obtain corresponding system-like embodiments for implementing the methods in other method-like embodiments.

[0070] In summary, this invention first obtains a rigid trajectory through a trajectory extrapolation method; then, based on pseudorange observations, it optimizes the rigid trajectory, and the changes in the observed residuals after optimization reflect the changes in pseudorange gross errors; it extracts residual consistency and continuity features for each satellite, and marks faulty and normal satellites according to preset thresholds; based on the marked results, it can significantly suppress the destructive impact of slowly varying and abruptly changing gross errors on GNSS positioning and its fusion positioning performance. Furthermore, in the optimization process of the rigid trajectory, this invention, on the one hand, links pseudorange observations from multiple epochs, diluting the proportion of gross error observations, thus facilitating the identification of gross errors; on the other hand, it uses a robust kernel function to reduce the weight of observations with large residuals, reducing the impact of gross error observations on parameter estimation, thereby effectively suppressing the interference of gross error observations during the optimization process.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting pseudorange gross errors based on rigid trajectory constraints, characterized in that, The pseudo-range coarse error detection method comprises: Collect pseudo-range observations in a preset distance window, and obtain a rigid trajectory in a corresponding time period by using a dead reckoning means; the error source of the rigid trajectory is modeled as an initial position error, an initial heading error, and a scale error; Based on the pseudo-range observations in the preset distance window, the initial position, the initial heading, and the scale of the rigid trajectory are taken as estimated parameters, and an optimization method is used to solve the estimated parameters, and the pseudo-range residuals after optimization convergence are saved, including: constructing pseudo-range residuals based on the difference between the pseudo-range observations in the preset distance window and the pseudo-range prediction values calculated according to the rigid trajectory; by continuously adjusting the initial position, the initial heading, and the scale of the rigid trajectory, the objective function is minimized; the weight of the gross error observation in the objective function is reduced by combining a robust kernel function, and a set of estimated parameters that minimizes the objective function is solved, and the pseudo-range residuals after optimization convergence are saved; Based on the optimized pseudo-range residuals, the residual consistency features and the continuity features of each satellite are extracted, and the faulty satellites and the normal satellites are checked and marked according to a preset threshold.

2. The method of pseudo-range gross error detection based on rigid trajectory constraint according to claim 1, wherein, The dead reckoning means includes inertial navigation mechanical arrangement, vehicle dead reckoning, pedestrian dead reckoning, visual odometry, laser odometry, visual inertial odometry, or laser inertial odometry. 3.The method of claim 1, wherein, The optimization method includes Gauss-Newton method, Newton method, or Levenberg-Marquardt method; the robust kernel function includes Huber kernel function, Cauchy kernel function, Turkey kernel function, or Geman-McClure kernel function.

4. The method of claim 1, wherein the method comprises: Based on the optimized pseudo-range residuals, the residual consistency features and the continuity features of each satellite are extracted, and the faulty satellites and the normal satellites are checked and marked according to a preset threshold, including: Calculate the residual fluctuation amplitude of each satellite to check whether the satellite meets the residual consistency according to the residual fluctuation amplitude and a first preset threshold; Calculate the effective observation epoch proportion of each satellite to check whether the satellite meets the residual continuity according to the effective observation epoch proportion and a second preset threshold; Satellites that meet both the residual consistency and the residual continuity are marked as normal satellites, otherwise as faulty satellites.

5. The method of claim 4, wherein the method comprises: Calculate the residual fluctuation amplitude of each satellite to check whether the satellite meets the residual consistency according to the residual fluctuation amplitude and a first preset threshold, including: Obtain the maximum pseudo-range residual and the minimum pseudo-range residual of each satellite in all epochs in the time period corresponding to the preset distance window; Calculate the difference between the maximum pseudo-range residual and the minimum pseudo-range residual of each satellite; wherein the difference between the maximum pseudo-range residual and the minimum pseudo-range residual is the residual fluctuation amplitude; Compare the residual fluctuation amplitude with the first preset threshold; if the residual fluctuation amplitude is less than or equal to the first preset threshold, the satellite meets the residual consistency; if the residual fluctuation amplitude is greater than the first preset threshold, the satellite does not meet the residual consistency.

6. The method of claim 4, wherein the method comprises: Calculate the effective observation epoch proportion of each satellite to check whether the satellite meets the residual continuity according to the effective observation epoch proportion and a second preset threshold, including: Obtaining the number of continuous effective pseudorange observation epochs and the total number of epochs of each satellite in the preset distance window corresponding period; Calculating the ratio of the number of continuous effective pseudorange observation epochs and the total number of epochs; wherein the ratio of the number of continuous effective pseudorange observation epochs and the total number of epochs is the effective observation epoch ratio; Comparing the effective observation epoch ratio with the second preset threshold value; if the effective observation epoch ratio is greater than or equal to the second preset threshold value, the satellite meets the residual continuity; if the effective observation epoch ratio is less than the second preset threshold value, the satellite does not meet the residual continuity.

7. The method of claim 1, wherein the method comprises: The pseudorange rough error detection method further comprises: dynamically adjusting the weight of the pseudorange observation value of each satellite participating in the fusion positioning according to the marking result of each satellite, and reducing the weight or eliminating the faulty satellite. 8.A pseudo-range gross error detection system based on rigid trajectory constraint, characterized in that, Comprise: The rigid trajectory calculation module collects the pseudorange observation value of the preset distance window, and obtains the rigid trajectory in the corresponding period by using the dead reckoning method; The error source modeling of the rigid trajectory is the initial position error, the initial heading error and the scale error; The pseudorange residual error optimization module, based on the pseudorange observation value in the preset distance window, takes the initial position, the initial heading and the scale of the rigid trajectory as the estimated parameters, uses the optimization method to solve the estimated parameters, and saves the optimized converged pseudorange residual error, comprising: constructing the pseudorange residual error based on the difference between the pseudorange observation value in the preset distance window and the pseudorange prediction value calculated according to the rigid trajectory; by continuously adjusting the initial position, the initial heading and the scale of the rigid trajectory, the objective function is minimized; combined with the robust kernel function to reduce the weight of the rough error observation in the objective function, a set of estimated parameters that minimize the objective function is solved, and the optimized converged pseudorange residual error is saved; The satellite verification marking module extracts the residual error consistency feature and the continuity feature based on the optimized pseudorange residual error, and according to the preset threshold value, checks and marks the faulty satellite and the normal satellite.

9. A non-transitory computer-readable storage medium, comprising: The non-transitory computer readable storage medium stores computer instructions, which make the computer execute the rigid trajectory constraint based pseudorange rough error detection method in any one of claims 1 to 7.

Images