GNSS tight combination positioning method and system with bidirectional constraint and fast re-convergence

Abstract

The present application belongs to the technical field of multi-source fusion positioning, and discloses a GNSS tight combination positioning method and system with bidirectional constraint and rapid re-convergence. In the stage of normal availability of GNSS signals, the present application uses the short-time continuous high-precision motion information provided by the IMU to constrain and test the rationality of the 5G ranging change, so as to reduce the adverse effects of unreliable 5G observation on positioning calculation. In addition, in the stage of GNSS signal rejection or serious degradation, the present application uses the 5G ranging information to continuously constrain and correct the IMU propagation state, thereby improving the constraint ability and positioning stability of the system in complex environments. Furthermore, in the stage of GNSS signal recovery, the present application uses the continuous state maintained by the IMU and 5G in the previous stage as a transition prior, to guide the GNSS calculation process to quickly recover to a stable state, thereby improving the continuity and real-time performance of the system in the intermittent availability scenario of GNSS.

Description

Technical Field

[0001] This invention belongs to the field of multi-source fusion positioning technology, specifically relating to a GNSS compact combination positioning method and system with bidirectional constraints and fast reconvergence. Background Technology

[0002] Global Navigation Satellite System (GNSS), with its advantages of wide coverage and high positioning accuracy, has become a fundamental technology for outdoor navigation and positioning services, and is widely used in fields such as intelligent transportation, mobile terminal positioning, and location services. However, in complex environments such as urban canyons, under overpasses, and tunnel entrances and exits, GNSS signals are susceptible to obstruction, multipath effects, and non-line-of-sight propagation, leading to a reduction in the number of available satellites, a significant decrease in observation quality, and consequently, deterioration in positioning accuracy or even interruption of calculation.

[0003] To compensate for the shortcomings of GNSS, inertial measurement units (IMUs) are often introduced to assist in positioning. IMUs have the advantages of high short-term accuracy and strong autonomy, but the zero bias and scaling factor errors of their sensors accumulate over time. In the event of a long-term interruption of GNSS signals, it is difficult to maintain high-precision positioning by relying solely on IMUs; and when GNSS signals are restored, the excessively large prior position error of the system often leads to an excessively long GNSS reconvergence time, making it difficult to achieve a smooth transition in positioning accuracy.

[0004] In recent years, with the large-scale deployment of 5G communication networks, their high bandwidth, low latency, and high-density base station layout have given them excellent potential for local ranging. By introducing 5G ranging information, it is expected to constrain IMU drift during GNSS degradation and provide additional auxiliary information support for the system after GNSS signal recovery.

[0005] However, existing GNSS, IMU, and 5G fusion positioning methods still have significant shortcomings in practical applications: First, when 5G ranging is affected by non-line-of-sight propagation, multipath effects, or abrupt errors, existing methods lack a mechanism to effectively constrain and verify abnormal 5G ranging using short-term high-precision motion information from the IMU, easily introducing unreliable 5G observations into the solution process. Second, during the GNSS rejection phase, existing methods fail to fully utilize the suppression effect of 5G ranging on IMU cumulative drift, leading to a decline in the quality of system continuity. Third, during the GNSS signal recovery phase, due to the lack of effective utilization of the IMU and 5G continuity from the previous phase, the system reconvergence process is slow, making it difficult to meet the requirements of high-continuity positioning applications.

[0006] Therefore, there is an urgent need to propose a GNSS compact combination positioning method that can achieve IMU constraint on 5G abnormal ranging, 5G suppression of IMU drift, and assist in rapid reconvergence during the GNSS signal recovery phase. Summary of the Invention

[0007] The purpose of this invention is to propose a GNSS compact combination positioning method with bidirectional constraints and fast reconvergence. This method achieves continuous and stable positioning in complex environments by uniformly modeling GNSS, IMU and 5G multi-source observations and combining the bidirectional constraint mechanism of IMU and 5G.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A GNSS compact combination positioning method with bidirectional constraints and fast reconvergence includes the following steps: Step 1. Using a positioning terminal that integrates a GNSS receiver, an IMU sensor, and a 5G communication module, collect multi-source observation data in real time and construct a continuous multi-epoch observation sequence; Step 2. Perform validity screening and gross error removal on the multi-source observation data in the observation sequence; Step 3. Perform error modeling and correction on the multi-source observation data in the observation sequence;

[0009] Step 4. Construct a tightly combined state vector of GNSS, IMU and 5G within a unified state space; Step 5. Establish the state propagation model and the observation model; Step 6. Based on the state propagation model and observation model established in Step 5, perform joint positioning calculation to obtain the values ​​of the tightly combined state vector of GNSS, IMU and 5G, which include three-dimensional position, velocity and attitude and positioning quality index information; During the joint positioning solution process, when the GNSS signal is available, the IMU is used to constrain the 5G abnormal ranging; when the GNSS signal is rejected or degraded, the 5G is used to constrain the IMU drift; and when the GNSS signal is restored, a fast reconvergence process is performed.

[0010] Furthermore, based on the GNSS compact combination positioning method with bidirectional constraints and fast reconvergence, this invention also proposes a corresponding GNSS compact combination positioning system with bidirectional constraints and fast reconvergence, the technical solution of which is as follows: A tightly coupled GNSS positioning system with bidirectional constraints and fast reconvergence includes: The observation sequence construction module is used to collect multi-source observation data in real time through a positioning terminal that integrates a GNSS receiver, an IMU sensor, and a 5G communication module, and construct a continuous multi-epoch observation sequence. The data preprocessing module is used to screen the validity of multi-source observation data in the observation sequence and remove gross errors; The error correction module is used to model and correct errors in multi-source observation data in the observation sequence. The state vector construction module is used to construct tightly combined state vectors of GNSS, IMU and 5G in a unified state space; The State Propagation and Observation Model Building Module is used to build state propagation models and observation models; And a joint positioning solution module, which is used to perform joint positioning solution based on the state propagation model and the observation model to obtain the values ​​of the tightly combined state vector of GNSS, IMU and 5G, including three-dimensional position, velocity and attitude and positioning quality index information; During the joint positioning solution process, when the GNSS signal is available, the IMU is used to constrain the 5G abnormal ranging; when the GNSS signal is rejected or degraded, the 5G is used to constrain the IMU drift; and when the GNSS signal is restored, a fast reconvergence process is performed.

[0011] The present invention has the following advantages: As described above, this invention discloses a GNSS compact combination positioning method with bidirectional constraints and fast reconvergence. This method proposes a 5G anomaly ranging constraint and verification scheme based on IMU short-time high-precision estimation results. Specifically, during the phase when GNSS signals are normally available, IMU-based constraints on 5G anomaly ranging are implemented. Utilizing the short-time continuous high-precision motion information provided by the IMU, the rationality of 5G ranging changes is constrained and verified. This enables the identification, downweighting, or elimination of abnormal 5G observations such as non-line-of-sight propagation, multipath effects, and ranging abrupt changes, thereby reducing the adverse effects of unreliable 5G observations on positioning solutions. Furthermore, this invention also constructs an auxiliary constraint mechanism for continuously suppressing 5G-based IMU cumulative drift. Specifically, during the phase of GNSS signal rejection or severe degradation, 5G-based IMU drift constraints are implemented. 5G ranging information is used to continuously constrain and correct the IMU propagation state, which can slow down the trend of IMU error accumulation over time, maintain a high-quality continuous state prior, and thus improve the system's constraint capability and positioning stability in complex environments. In addition, the method of the present invention also designs a GNSS fast reconvergence mechanism based on IMU and 5G continuous state assistance. That is, during the GNSS signal recovery phase, a fast reconvergence process is executed, using the continuous state maintained by IMU and 5G in the previous phase as a transitional prior to guide the GNSS solution process to quickly recover to a stable state. This can shorten the reconvergence time, reduce the fluctuation of positioning results, and thus improve the continuity and real-time performance of the system in GNSS intermittent availability scenarios. Attached Figure Description

[0012] Figure 1 This is a flowchart of the GNSS compact combination positioning method with bidirectional constraints and fast reconvergence in an embodiment of the present invention. Detailed Implementation

[0013] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments: Example 1 The existing GNSS, IMU and 5G converged positioning technology still faces the following key problems in practical applications: First, there is a lack of an effective mechanism to constrain and verify 5G anomaly ranging by utilizing the short-time, high-precision calculation results of IMU.

[0014] In the existing fusion framework, when there are non-line-of-sight propagation, multipath effects or abrupt errors, 5G ranging observations usually rely mainly on statistical residuals or empirical thresholds for anomaly detection. This fails to fully utilize the short-term continuous high-precision calculation results provided by the IMU to constrain and verify the rationality of 5G ranging changes, and easily introduces abnormal 5G observations into the state estimation process, affecting the stability and accuracy of the positioning results.

[0015] Secondly, during the GNSS denial phase, the inhibitory effect of 5G on IMU cumulative drift was not fully realized.

[0016] While IMUs offer the advantage of short-term autonomous continuity, their position, velocity, and other state errors accumulate over time. Existing methods, when GNSS is unavailable, do not fully utilize 5G ranging information to continuously constrain and correct the IMU's propagation state, leading to a continuous increase in IMU drift errors. This makes it difficult to maintain high-quality continuous state priors, thus limiting the system's overall positioning performance in complex environments.

[0017] Third, the GNSS signal recovery phase lacks a fast reconvergence mechanism based on the continuous state of IMU and 5G.

[0018] In scenarios where GNSS is intermittently available, such as tunnel entrances and exits, existing methods typically rely directly on GNSS observations from the initial recovery period to re-establish the solution state after GNSS recovery. This lacks effective utilization of the continuous state maintained by the IMU and 5G in the previous stage, resulting in a long re-convergence time and even state oscillations or sudden changes in positioning results, making it difficult to meet the requirements for high-continuity positioning.

[0019] To address the aforementioned issues, this embodiment proposes a GNSS tightly coupled positioning method based on IMU and 5G bidirectional constraints and fast reconvergence. This method constructs a bidirectional constraint mechanism between IMU and 5G by jointly modeling GNSS, IMU, and 5G observations in a unified state space: on the one hand, it uses short-time high-precision calculations from the IMU to constrain and verify abnormal 5G ranging; on the other hand, it uses 5G ranging to suppress the cumulative drift of the IMU under GNSS rejection conditions, and in the GNSS signal recovery phase, it uses the continuous state maintained by the IMU and 5G to assist in fast reconvergence.

[0020] The GNSS tight combination positioning method with bidirectional constraints and fast reconvergence proposed in this invention is particularly suitable for high-continuity positioning requirements in environments where GNSS signals are intermittently available, such as urban canyons, long tunnels and their entrances and exits.

[0021] like Figure 1 As shown, the GNSS compact combination positioning method with bidirectional constraints and fast reconvergence specifically includes the following steps: Step 1. Data Acquisition: A positioning terminal integrating a GNSS receiver, IMU sensor, and 5G communication module is used to acquire multi-source observation data in real time, constructing a continuous multi-epoch observation sequence. The 5G communication module refers to the 5G receiver.

[0022] In this embodiment, step 1 specifically includes: The positioning terminal, which integrates a GNSS receiver, an IMU sensor, and a 5G communication module, collects multi-source observation data in real time, including: raw observation values ​​such as GNSS pseudorange, carrier phase, Doppler frequency shift, and carrier-to-noise ratio; IMU triaxial acceleration and triaxial angular velocity measurements; and TOA (Time of Arrival) ranging values ​​from the 5G base station to the 5G receiver.

[0023] Simultaneously, the multi-source observation data is uniformly time-stamped, recording the receiver's local time and the timestamp information of each sensor. Time alignment of observation data at different frequencies is achieved through interpolation or synchronization mechanisms, constructing a continuous multi-epoch observation sequence to provide basic data support for subsequent fusion and solution.

[0024] Step 2. Data preprocessing: Perform validity screening and gross error removal on the multi-source observation data in the observation sequence to improve the quality of the input data.

[0025] In this embodiment, step 2 specifically includes: GNSS observation quality control: Based on the carrier-to-noise ratio (CNR) threshold, the raw GNSS observations collected by the GNSS receiver are screened for validity, and observations with a CNR lower than the CNR threshold are removed. The consistency of GNSS pseudorange, carrier phase, and Doppler shift observations is checked using an inter-epoch differential check method. If the differential result exceeds a preset threshold, it is determined to be an abnormal observation and is removed.

[0026] Cycle slip detection and marking: By combining the relationship between the carrier phase observations and the Doppler frequency shift observations, cycle slips are detected in the carrier phase observations, and those with cycle slips are marked or re-initialized.

[0027] Preliminary screening of 5G ranging observations: The validity of ranging observations from 5G base stations to 5G receivers is screened based on the preset physical feasible range and the rate of change between epochs. If the ranging observations exceed the preset physical feasible range or the changes between epochs are abnormal, they are judged as gross errors and removed.

[0028] IMU zero-bias initialization: The original IMU measurements, namely the IMU triaxial acceleration and triaxial angular velocity measurements, are corrected for zero bias and the sensor coordinate system is transformed into the navigation coordinate system to ensure the consistency of subsequent state propagation.

[0029] Step 3. Error Modeling and Correction: Perform error modeling and correction on the multi-source observation data in the observation sequence.

[0030] In this embodiment, step 3 specifically includes: Systematic errors in GNSS observations are modeled and corrected to improve the accuracy of the observation model. This includes corrections for tropospheric dry delay, ionospheric delay, satellite clock bias, and hardware delay. These error terms can be calculated and corrected using existing mature models; the specific implementation details will not be elaborated further.

[0031] It should be noted that the tropospheric delay includes the dry tropospheric delay and the wet tropospheric delay. The dry tropospheric delay is corrected here, while the wet tropospheric delay is solved as an unknown state variable in the subsequent formula (1).

[0032] Meanwhile, the IMU measurement noise is modeled based on the power spectral density parameters provided by the sensor and discretized into process noise.

[0033] The 5G ranging error is modeled using a random noise model, and further suppressed through a constraint mechanism during the subsequent fusion process.

[0034] The above data acquisition, preprocessing, and error correction processes are routine basic processing steps in multi-source fusion localization, and their purpose is to provide unified and stable input data for subsequent bidirectional constraints and joint estimation.

[0035] After completing the preprocessing and error correction of multi-source observation data, in order to achieve joint estimation of GNSS, IMU and 5G multi-source information under a unified framework, a unified state space vector and the corresponding state propagation model and observation model are further constructed.

[0036] Step 4. Construction of unified state space vectors: Construct tightly combined state vectors of GNSS, IMU and 5G in a unified state space.

[0037] In this embodiment, step 4 specifically includes: Constructing tightly coupled state vectors of GNSS, IMU, and 5G within a unified state space The state vector is shown in formula (1): (1) in, Indicates location; Indicates speed; Represents attitude quaternions; and These represent the IMU accelerometer and gyroscope bias, respectively. and These represent the clock biases of GNSS satellites and 5G receivers, respectively. Indicates tropospheric wet delay; This indicates the carrier integer ambiguity.

[0038] Constructed state vector This is to clarify the parameters to be solved. As the core state parameter in the unified state-space model, the state vector is used for state propagation and state correction in the subsequent state update process and fast reconvergence process.

[0039] This step introduces state parameters such as position, velocity, attitude, IMU bias, receiver clock error, tropospheric wet delay, and carrier ambiguity to achieve a unified expression of multi-source observation information, providing a foundation for subsequent constraint construction and joint estimation.

[0040] Step 4 of the method of the present invention constructs a tightly combined state vector of GNSS, IMU and 5G in a unified state space, and establishes the corresponding state propagation model and observation model in the subsequent step 5, so that multi-source observations can participate in the solution under the same framework. The unified state space model corresponds to formulas (1) to (5), which includes the unified state space vector shown in formula (1) and the state propagation model and observation model shown in formulas (2) to (5).

[0041] Step 5. Establish the state propagation model and the observation model.

[0042] In this embodiment, step 5 specifically includes: The equation for GNSS undifferentiated observations is: (2) in, Numbering of GNSS satellite receivers Numbering GNSS satellites; These are pseudorange observations; Wavelength; These are carrier phase observations; This refers to the geometric distance between the GNSS satellite receiver and the GNSS satellite. The speed of light in a vacuum; Clock bias for GNSS satellite receivers; For GNSS satellite clock bias; and These are the ionospheric delay and the tropospheric delay, respectively. For carrier phase ambiguity; and These are the pseudorange hardware delays at the GNSS satellite receiver end and the GNSS satellite end, respectively. and These are the carrier phase hardware delays at the GNSS satellite receiver end and the GNSS satellite end, respectively. and Other errors in pseudorange and carrier phase observations.

[0043] Considering that relativistic effects, Earth's rotation effects, tidal effects, antenna errors, and other error terms are not the focus of this invention and can be corrected by existing models, they are not listed here.

[0044] The IMU measurement model is: (3) in, The values ​​are accelerometer measurements, i.e., triaxial acceleration measurements. The direction cosine matrix represents the attitude. Represents actual acceleration. Represents the gravity vector; and Bias for the IMU accelerometer and gyroscope; and This indicates the noise measured by the accelerometer and gyroscope; This indicates the gyroscope measurement value, which is the three-axis angular velocity measurement value; This represents the true angular velocity.

[0045] In time interval Above, among which and Representing the start and end times respectively, pre-integrating the IMU measurements yields: (4) in, , , These represent the pre-integral quantities for position, velocity, and attitude, respectively. express The direction cosine matrix from the carrier coordinate system to the navigation coordinate system at any given time. express The direction cosine matrix from the carrier coordinate system to the navigation coordinate system at any given time. express Location at any given moment Right now Location at any given moment Right now The speed of time, express The speed of time, Indicates from arrive The time interval or integration time step.

[0046] The 5G distance-based observation model is as follows: (5) in, This is the serial number of the 5G receiver. Number the 5G base station; These are 5G observations, specifically the ranging observations from the 5G base station to the 5G receiver. This refers to the geometric distance between the 5G base station and the 5G receiver. For 5G receiver clock bias; This refers to the clock bias of 5G base stations; clock bias is the clock error. Other measurement errors.

[0047] The established state propagation model includes formulas (3) and (4), and the established observation model includes formulas (2) and (5).

[0048] After completing the unified observation modeling, the joint positioning and solution phase begins. In step 6, different processing procedures will be executed for different environmental conditions and observation states, including IMU-to-5G anomaly ranging constraints, 5G-to-IMU drift constraints, and rapid reconvergence during the GNSS signal recovery phase. These are different constraints and update mechanisms dynamically triggered based on the GNSS and 5G observation states.

[0049] Step 6. Based on the state propagation model and observation model established in Step 5, perform joint positioning calculation, and finally output information including high-precision three-dimensional position, velocity, attitude, positioning quality index, etc., that is, obtain the corresponding values ​​in the tightly combined state vector of GNSS, IMU and 5G.

[0050] During the joint positioning solution process, when the GNSS signal is normally available, the IMU is used to constrain the 5G abnormal ranging; when the GNSS signal is rejected or severely degraded, the 5G is used to constrain the IMU drift; and when the GNSS signal is restored, a fast reconvergence process is performed.

[0051] In step 6 of this embodiment, the IMU's constraints on 5G anomaly ranging include motion prediction, consistency verification, and anomaly removal. The specific process is as follows: The location calculated based on the IMU is shown in formula (6): (6) in, Indicates the result calculated based on the IMU. The position at that moment; express Estimated location at any given time; Indicates the expected component of the location.

[0052] The IMU predicted location is obtained based on formula (6). Further calculations were performed on the 5G receiver and 5G base station. The predicted 5G ranging values ​​between them are shown in formula (7): (7) in, This represents the predicted 5G ranging value calculated based on the IMU's predicted location. 5G base station The three-dimensional known position coordinates.

[0053] Considering the clock difference between the 5G receiver and the base station, to eliminate the impact of the 5G receiver clock difference, the calculation of the clock difference between the 5G receiver and the 5G base station is performed. and The distance difference between them is: (8) Among them, 5G base stations That is, the first One 5G base station, 5G base station That is, the first One 5G base station; for The data from the 5G receiver to the 5G base station is calculated using 5G observations at all times. and The difference in distance between them; express 5G receiver to 5G base station The true geometric distance, express 5G receiver to 5G base station The true geometric distance, express 5G base station Ranging observations to the 5G receiver express 5G base station Ranging observations to the 5G receiver Indicates 5G base station Clock error, Indicates 5G base station Clock error.

[0054] Assuming high-precision time synchronization has been achieved between 5G base stations, that is At that time, the 5G receiver connects to the 5G base station. and The distance difference between them is: (9) Because base stations equipped with 5G high-precision positioning capabilities have high clock synchronization accuracy, the clock error between 5G base stations is temporarily ignored in the subsequent processing of the method of this invention.

[0055] Construct the difference between the actual 5G base station ranging difference and the 5G base station ranging difference predicted by the IMU. And treat it as a consistent residual: (10) in, Used to express the degree of inconsistency between the IMU estimation shown in formula (7) and 5G ranging; This is the distance difference between 5G base stations obtained by subtracting the 5G observation values. That is, the distance difference between 5G base stations predicted by the IMU; express The 5G receiver and 5G base station are calculated based on the IMU's predicted location at all times. Predicted 5G ranging values ​​between express The 5G receiver and 5G base station are calculated based on the IMU's predicted location at all times. Predicted 5G ranging values ​​between them.

[0056] when At that time, it is determined that the 5G base station and Ranging observations to the 5G receiver were considered outliers and were either removed or downweighted. To determine the threshold, the threshold is determined. The range can be set according to the 5G ranging accuracy and IMU calculation error level, preferably between 1m and 3m; the downweighting process increases the 5G base station's range. and The observation noise covariance of ranging observations to the 5G receiver is used to reduce the influence weight of outlier observations during the state update process.

[0057] During the joint positioning calculation process, when the number of available GNSS satellites is detected to be lower than a set threshold, or when the GNSS positioning results are lost or the observation quality is severely degraded, the system is determined to enter the GNSS rejection phase.

[0058] During the GNSS rejection phase, raw 5G ranging observations are directly incorporated into the constructed unified state space model, participating in joint state estimation together with the IMU state propagation results. 5G observations do not form independent positioning results independently; instead, they directly participate in the state update process as raw ranging observations. Multi-base station ranging information continuously constrains and corrects the IMU propagation state, thereby achieving tight-binding positioning of 5G and IMU.

[0059] In step 6 of this embodiment, the 5G IMU drift constraint includes ranging constraint, joint correction, and drift suppression, and the specific execution process is as follows: During the GNSS signal rejection or degradation phase, i.e., when GNSS is unavailable, state propagation is first performed based on the IMU pre-integration result shown in formula (4) to obtain the predicted state of the 5G receiver in the current epoch, i.e., the state vector shown in formula (1). The values ​​of each state parameter in the equation; where, as shown in formula (6), the receiver predicted position calculated based on the IMU is expressed as... .

[0060] The predicted locations are used to construct the 5G ranging residuals between the receiver and each 5G base station. The 5G ranging residuals are defined as follows: (11) in, express 5G base station 5G ranging residuals for IMU predicted location express 5G base station Ranging observations to the 5G receiver.

[0061] In a multi-base station environment, the 5G ranging residuals corresponding to all available 5G base stations are jointly constructed into a residual vector. : (12) in, This indicates the number of available 5G base stations in the current epoch; to They represent Time of the first To the 5G ranging residuals of IMU predicted location for each 5G base station.

[0062] Linearize the 5G distance measurement-based observation model shown in formula (5) to construct the corresponding observation equation: (13) in, This represents the observation design matrix corresponding to 5G ranging observations, whose elements are derived from the observation equations and the unified state vector. The first-order partial derivatives constitute; Represents the unified state vector The corresponding error state correction amount, Operators that indicate error or correction amounts; This represents the noise vector of 5G ranging observation.

[0063] The 5G ranging residual is introduced as an observation update term into the state update process, which is a tightly combined state vector of GNSS, IMU, and 5G. The state parameters in the data are jointly corrected; the error state correction amount is... Represented as: (14) in, The state update gain matrix is ​​represented by the gain matrix. It can be calculated based on the Kalman filter update formula.

[0064] The prior state obtained from IMU propagation is corrected using the error state correction to obtain the updated posterior state: (15) in, The prior state obtained through IMU state propagation; This represents the posterior state after the tight combination update of 5G and IMU, i.e., the posterior state after introducing 5G ranging constraints.

[0065] Through the above joint update process, the IMU propagation state can be continuously constrained and corrected using the raw ranging observations of multiple base stations 5G, thereby suppressing the cumulative divergence of state errors such as position, velocity and attitude during the GNSS rejection phase, maintaining a high-quality continuous state prior, and improving the continuity and stability of system positioning results in complex environments.

[0066] During the joint positioning solution process, when the system detects that the GNSS signal has been restored to usability and the restored GNSS observations meet the preset quality conditions, the system enters the GNSS signal recovery phase.

[0067] During GNSS signal interruption or degradation, the system continuously maintains a unified state vector through tight combination positioning of 5G and IMU. Therefore, when GNSS is restored, the system has a high-quality continuous state prior, including state parameters such as position, velocity, attitude and IMU bias.

[0068] Unlike traditional GNSS reinitialization methods that rely directly on GNSS observations from the initial recovery phase to re-establish the solution state after signal recovery, this invention utilizes the continuous state maintained by the tight combination of 5G and IMU during the GNSS rejection phase as the prior state for the GNSS signal recovery phase. The original GNSS observations after recovery are reintroduced into the unified state space model and participate in joint state estimation together with IMU and 5G observations.

[0069] In step 6 of this embodiment, fast reconvergence includes prior guidance, progressive fusion, and fast recovery, and the process is as follows: During the GNSS signal recovery phase, the posterior state updated by the tight combination of 5G and IMU in the previous epoch is used as the prior state of the current epoch. The recovered original GNSS observations and 5G ranging observations are jointly introduced into the unified state space model constructed by formulas (1) to (5), and a joint observation update equation is constructed based on the prior state obtained from IMU state propagation: (16) in, This represents the residual from joint GNSS and 5G observations. Represents the joint observation design matrix; This represents the joint observation noise vector.

[0070] Considering that there may still be problems such as observation instability, multipath effect or cycle slip residue in the early stage of GNSS signal recovery, in order to avoid abnormal GNSS observations in the early stage of recovery from causing drastic disturbances to joint state estimation, this invention further adopts a progressive fusion strategy of GNSS observations to dynamically adjust the weight of the recovered GNSS observations.

[0071] Specifically, in the early stages of GNSS recovery, GNSS observations are assigned a low weight; as the number of consecutive effective GNSS observation epochs increases, the weight of GNSS observations in joint state estimation is gradually increased, so that GNSS observations gradually transition from a weakly constrained state to a normally constrained state.

[0072] The GNSS observation weighting adjustment coefficient is expressed as follows: (17) in, For the GNSS signal recovery phase GNSS observation weight adjustment coefficients corresponding to each epoch; This represents the number of consecutive valid epochs since the GNSS signal was recovered. The time constant represents the rate of gradual change of the GNSS observation weights. Its value can be set according to the convergence speed requirements of the GNSS signal recovery stage, and is preferably 5 to 30 epochs.

[0073] By dynamically adjusting the GNSS observation noise covariance matrix, the gradual recovery of GNSS observation weights can be achieved. (18) in, Indicates the GNSS signal recovery phase. The GNSS observation noise covariance matrix for each epoch; This represents the standard observation noise covariance matrix corresponding to the normal and stable phase of GNSS.

[0074] In the early stages of GNSS recovery, due to The value is relatively small, so the corresponding GNSS observation weight is low. At this time, the joint state estimation mainly inherits the continuous state maintained by the tight combination of 5G and IMU in the previous stage.

[0075] As GNSS observations gradually stabilize As the number of GNSS observations gradually increases, their corrective role in joint state estimation is enhanced, and the system eventually recovers to a normal tight combination positioning state of GNSS, IMU, and 5G.

[0076] Through the aforementioned continuous state inheritance and progressive fusion mechanism of GNSS observation, the state oscillation and error mutation during the GNSS signal recovery stage can be effectively reduced, the speed of ambiguity re-fixation and state re-convergence can be improved, and the rapid and smooth re-convergence of the joint positioning results of GNSS, IMU and 5G can be achieved.

[0077] This invention aims to provide a GNSS tightly coupled positioning method based on IMU and 5G bidirectional constraints and fast reconvergence. By jointly modeling GNSS, IMU, and 5G multi-source observations within a unified state space, a 5G anomaly ranging constraint and verification mechanism based on short-time high-precision IMU calculation results is constructed. This enables the identification and suppression of anomalies in 5G observations such as non-line-of-sight propagation, multipath effects, and ranging abrupt changes. Simultaneously, a continuous constraint mechanism for the IMU state by 5G is established. In GNSS rejection environments, 5G ranging information is used to suppress IMU cumulative drift, maintaining a high-quality continuous state prior, thereby improving the overall constraint capability and solution stability of the system in complex environments. Furthermore, a fast reconvergence mechanism is constructed for GNSS intermittent availability scenarios. During the GNSS signal recovery phase, the continuous state maintained by the IMU and 5G is used as a transitional prior to assist GNSS in quickly restoring a stable solution state, thus shortening the reconvergence time and improving the continuity and stability of the positioning results.

[0078] In addition, the present invention also includes the following alternative embodiments: When 5G ranging observations are unavailable or of poor quality, Time Difference of Arrival (TDOA), Round Trip Range (RTT), Angle of Arrival (AOA), or combinations thereof can be used as alternative observation information. In addition, near-range communication observation data such as Ultra Wideband (UWB), WiFi RTT, or Bluetooth ranging can be introduced to provide auxiliary constraints on the IMU state in order to achieve continuous correction of the positioning state.

[0079] In the process of multi-source fusion localization solution, in addition to the error state Kalman filter method under the compact combination modeling framework, other methods such as extended Kalman filter (EKF), unscented Kalman filter (UKF), capacitive Kalman filter (CKF), particle filter (PF), or factor graph optimization (FGO) can be selected to solve the unified state space model according to specific application requirements and computing resource conditions.

[0080] In the specific implementation of IMU for constraining and verifying 5G anomaly ranging, in addition to consistency judgment based on the displacement change relationship between adjacent epochs, it can also be achieved based on velocity change, acceleration continuity, motion trajectory smoothness, or motion model constraints; correspondingly, in addition to rejection, the processing method for abnormal 5G observations can also be adjusted by using robust processing strategies.

[0081] During the GNSS rejection phase, in addition to being based on the ranging residual of a single base station, 5G can also constrain the IMU state through joint geometric constraints of multiple base stations, combined ranging observations, or equivalent position observations, thereby continuously suppressing and correcting the cumulative drift of the IMU.

[0082] In the process of achieving rapid reconvergence during the GNSS signal recovery phase, different transition strategies can be designed according to different application scenarios. For example, there are gradual update methods based on the continuation of prior states, phased convergence methods based on observation weight adjustment, or continuous reinitialization methods based on sliding window state inheritance, so as to achieve a smooth transition from weakly constrained states to strongly constrained states.

[0083] Compared with existing technologies, the present invention has the following advantages: First, the method of this invention can enhance the constraint and discrimination capabilities of 5G anomaly ranging and improve the reliability of fusion calculation.

[0084] This invention utilizes short-time, high-precision IMU calculation results to constrain and verify the consistency of 5G ranging changes. Compared with methods that rely solely on statistical residuals or empirical thresholds, it can identify and suppress abnormal observations such as non-line-of-sight propagation, multipath effects, and ranging abrupt changes from the perspective of motion continuity. This reduces the interference of unreliable 5G data on positioning calculations and improves the stability and reliability of the fusion process.

[0085] Secondly, the method of the present invention can suppress the accumulation of IMU errors during the GNSS rejection phase and improve the continuous state quality.

[0086] During the GNSS unavailable phase, this invention utilizes 5G ranging information to continuously constrain the IMU propagation state, effectively mitigating the accumulation and divergence of state errors such as position and velocity, enabling the system to maintain a high-quality continuous state prior, and providing a stable foundation for subsequent positioning calculations.

[0087] Third, the method of the present invention can improve the convergence conditions in the GNSS signal recovery stage and achieve faster and more stable reconvergence.

[0088] This invention utilizes the continuous state maintained by the IMU and 5G in the previous stage as a transitional prior during the GNSS signal recovery phase, avoiding direct reliance on the initial state with large errors for calculation. This reduces the solution space range, decreases calculation oscillations, accelerates convergence speed, and improves the smoothness and continuity of positioning results.

[0089] Example 2 This embodiment 2 describes a GNSS compact combination positioning system with bidirectional constraints and fast reconvergence, which is based on the same inventive concept as the GNSS compact combination positioning method with bidirectional constraints and fast reconvergence in embodiment 1.

[0090] Specifically, this bidirectional constrained and fast reconvergence GNSS compact combination positioning system includes the following modules: The observation sequence construction module is used to collect multi-source observation data in real time through a positioning terminal that integrates a GNSS receiver, an IMU sensor, and a 5G communication module, and construct a continuous multi-epoch observation sequence.

[0091] The data preprocessing module is used to screen the validity of multi-source observation data in the observation sequence and remove gross errors.

[0092] The error correction module is used to model and correct errors in multi-source observation data in the observation sequence.

[0093] The state vector construction module is used to construct tightly combined state vectors of GNSS, IMU and 5G within a unified state space.

[0094] The State Propagation and Observation Model Building Module is used to build state propagation models and observation models.

[0095] It also includes a joint positioning solution module, which performs joint positioning solution based on the state propagation model and the observation model to obtain the values ​​of the tightly combined state vector of GNSS, IMU and 5G, including three-dimensional position, velocity and attitude, and positioning quality index information.

[0096] During the joint positioning solution process, when the GNSS signal is available, the IMU is used to constrain the 5G abnormal ranging; when the GNSS signal is rejected or degraded, the 5G is used to constrain the IMU drift; and when the GNSS signal is restored, a fast reconvergence process is performed.

[0097] It should be noted that in the GNSS tightly coupled positioning system with bidirectional constraints and fast reconvergence, the implementation process of the functions and roles of each functional module is detailed in the implementation process of the corresponding steps in the method of Example 1, and will not be repeated here.

[0098] Of course, the above description is only a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. It should be noted that any equivalent substitutions or obvious modifications made by those skilled in the art under the guidance of this specification fall within the scope of this specification and should be protected by the present invention.

Claims

1. A GNSS compact combination positioning method with bidirectional constraints and fast reconvergence, characterized in that, Includes the following steps: Step 1. Using a positioning terminal that integrates a GNSS receiver, an IMU sensor, and a 5G communication module, collect multi-source observation data in real time and construct a continuous multi-epoch observation sequence; Step 2. Perform validity screening and gross error removal on the multi-source observation data in the observation sequence; Step 3. Perform error modeling and correction on the multi-source observation data in the observation sequence; Step 4. Construct a tightly combined state vector of GNSS, IMU and 5G within a unified state space; Step 5. Establish the state propagation model and the observation model; Step 6. Based on the state propagation model and observation model established in Step 5, perform joint positioning calculation to obtain the values ​​of the tightly combined state vector of GNSS, IMU and 5G, which include three-dimensional position, velocity and attitude and positioning quality index information; During the joint positioning calculation process, when GNSS signals are normally available, IMU is used to constrain 5G abnormal ranging. During the GNSS signal rejection or degradation phase, 5G-based IMU drift constraints are implemented; during the GNSS signal recovery phase, a fast reconvergence process is implemented. In step 6, the process of executing the IMU for 5G anomaly ranging constraints is specifically as follows: The location calculated based on the IMU is shown in formula (6): (6) in, Indicates the result calculated based on the IMU. The position at that moment; express Estimated location at any given time; Indicates the expected location component; The IMU predicted location is obtained based on formula (6). Further calculations were performed on the 5G receiver and 5G base station. The predicted 5G ranging values ​​between them are shown in formula (7): (7) in, This represents the predicted 5G ranging value calculated based on the IMU's predicted location. 5G base station The three-dimensional known position coordinates; To eliminate the impact of 5G receiver clock bias, the calculation of the distance from the 5G receiver to the 5G base station is performed. and The distance difference between them is: (8) Among them, 5G base stations That is, the first One 5G base station, 5G base station That is, the first One 5G base station; for The data from the 5G receiver to the 5G base station is calculated using 5G observations at all times. and The difference in distance between them; express 5G receiver to 5G base station The true geometric distance, express 5G receiver to 5G base station The true geometric distance, express 5G base station Ranging observations to the 5G receiver express 5G base station Ranging observations to the 5G receiver The speed of light in a vacuum. Indicates 5G base station Clock error, Indicates 5G base station Clock error; At that time, the 5G receiver connects to the 5G base station. and The distance difference between them is: (9) Construct the difference between the actual 5G base station ranging difference and the 5G base station ranging difference predicted by the IMU. And treat it as a consistent residual: (10) in, This is the distance difference between 5G base stations obtained by subtracting the 5G observation values. That is, the distance difference between 5G base stations predicted by the IMU; express The 5G receiver and 5G base station are calculated based on the IMU's predicted location at all times. Predicted 5G ranging values ​​between express The 5G receiver and 5G base station are calculated based on the IMU's predicted location at all times. Predicted 5G ranging values ​​between them; when At that time, it is determined that the 5G base station and Ranging observations to the 5G receiver were considered outliers and were either removed or downweighted. To determine the threshold, the weighting process involves increasing the weight of 5G base stations. and The observation noise covariance of ranging observations to the 5G receiver is used to reduce the influence weight of outlier observations during the state update process.

2. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 1, characterized in that, Step 1 specifically involves: The positioning terminal, which integrates a GNSS receiver, an IMU sensor, and a 5G communication module, collects multi-source observation data in real time, including: raw observation values ​​of GNSS pseudorange, carrier phase, Doppler frequency shift, and carrier-to-noise ratio; IMU triaxial acceleration and triaxial angular velocity measurements; and ranging observation values ​​from the 5G base station to the 5G receiver. Simultaneously, a unified time stamp is applied to the multi-source observation data, recording the receiver's local time and the timestamp information of each sensor. Time alignment of observation data at different frequencies is achieved through interpolation or synchronization mechanisms to construct a continuous multi-epoch observation sequence.

3. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 2, characterized in that, Step 2 specifically involves: The validity of the raw observations collected by the GNSS receiver, i.e., the GNSS observations, is screened based on the carrier-to-noise ratio threshold, and observations with a carrier-to-noise ratio lower than the carrier-to-noise ratio threshold are removed. The consistency of GNSS pseudorange, carrier phase and Doppler shift observations is checked by the interepoch difference check method. If the difference result exceeds the preset threshold, it is judged as an abnormal observation and is removed. By combining the relationship between the carrier phase observations and the Doppler frequency shift observations, cycle slip detection is performed on the carrier phase observations, and carrier phase observations with cycle slips are marked or reinitialized. The validity of the ranging observations from the 5G base station to the 5G receiver is screened based on the preset physical feasible range and the rate of change between epochs. If the ranging observations exceed the preset physical feasible range or the change between epochs is abnormal, they are judged as gross errors and removed. The original IMU measurements, namely the IMU triaxial acceleration and triaxial angular velocity measurements, are corrected for zero bias initial values, and the sensor coordinate system is transformed into the navigation coordinate system.

4. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 3, characterized in that, Step 3 specifically involves: Modeling and correction of systematic errors in GNSS observations, including tropospheric dry delay correction, ionospheric delay correction, satellite clock error correction and hardware delay correction; The IMU measurement noise is modeled based on the power spectral density parameters provided by the sensor and discretized into process noise; The 5G ranging error is modeled using a random noise model.

5. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 4, characterized in that, Step 4 specifically involves: Constructing tightly coupled state vectors of GNSS, IMU, and 5G within a unified state space The state vector is shown in formula (1): (1) in, Indicates location; Indicates speed; Represents attitude quaternions; and These represent the IMU accelerometer and gyroscope bias, respectively. and These represent the clock biases of GNSS satellites and 5G receivers, respectively. Indicates tropospheric wet delay; This indicates the carrier integer ambiguity.

6. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 5, characterized in that, Step 5 specifically involves: The equation for GNSS undifferentiated observations is: (2) in, The GNSS satellite receiver number, Number the GNSS satellites; These are pseudorange observations; Wavelength; These are carrier phase observations; This refers to the geometric distance between the GNSS satellite receiver and the GNSS satellite. Clock bias for GNSS satellite receivers; For GNSS satellite clock bias; and These are the ionospheric delay and the tropospheric delay, respectively. For carrier phase ambiguity; and These are the pseudorange hardware delays at the GNSS satellite receiver end and the GNSS satellite end, respectively. and These are the carrier phase hardware delays at the GNSS satellite receiver end and the GNSS satellite end, respectively. and Other errors in pseudorange and carrier phase observations; The IMU measurement model is: (3) in, The values ​​are accelerometer measurements, i.e., triaxial acceleration measurements. The direction cosine matrix represents the attitude. Represents actual acceleration. Represents the gravity vector; and Bias for the IMU accelerometer and gyroscope; and This indicates the noise measured by the accelerometer and gyroscope; This indicates the gyroscope measurement value, which is the three-axis angular velocity measurement value; Represents the true angular velocity; In the time interval Above, among which and Representing the start and end times respectively, pre-integrating the IMU measurements yields: (4) in, , , These represent the pre-integral quantities for position, velocity, and attitude, respectively. express The direction cosine matrix from the carrier coordinate system to the navigation coordinate system at any given time. express The direction cosine matrix from the carrier coordinate system to the navigation coordinate system at any given time. express Location at any given moment Right now Location at any given moment Right now The speed of time, express The speed of time, Indicates from arrive The time interval or integration time step; The 5G distance-based observation model is as follows: (5) in, This is the serial number of the 5G receiver. Number the 5G base station; These are 5G observations, specifically the ranging observations from the 5G base station to the 5G receiver. This refers to the geometric distance between the 5G base station and the 5G receiver. For 5G receiver clock bias; This refers to the clock bias of 5G base stations; clock bias is the clock error. Other measurement errors; The established state propagation model includes formulas (3) and (4), and the established observation model includes formulas (2) and (5).

7. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 6, characterized in that, In step 6, the process of performing 5G-based IMU drift constraint is specifically as follows: During the GNSS signal rejection or degradation phase, i.e., when GNSS is unavailable, state propagation is first performed based on the IMU pre-integration result shown in formula (4) to obtain the predicted state of the 5G receiver in the current epoch, i.e., the state vector shown in formula (1). The values ​​of each state parameter in the equation; where, as shown in formula (6), the receiver predicted position calculated based on the IMU is expressed as... ; The predicted locations are used to construct the 5G ranging residuals between the receiver and each 5G base station. The 5G ranging residuals are defined as follows: (11) in, express 5G base station 5G ranging residuals for IMU predicted location express 5G base station Ranging observations to the 5G receiver; In a multi-base station environment, the 5G ranging residuals corresponding to all available 5G base stations are jointly constructed into a residual vector. : (12) in, This indicates the number of available 5G base stations in the current epoch; to They represent Time of the first To the 5G ranging residuals of IMU predicted location from 5G base stations; Linearize the 5G distance measurement-based observation model shown in formula (5) to construct the corresponding observation equation: (13) in, This represents the observation design matrix corresponding to 5G ranging observations, whose elements are derived from the observation equations and the unified state vector. The first-order partial derivatives constitute; Represents the unified state vector The corresponding error state correction amount; This represents the 5G ranging observation noise vector; The 5G ranging residual is introduced as an observation update term into the state update process, which is a tightly combined state vector of GNSS, IMU, and 5G. The state parameters in the data are jointly corrected; the error state correction amount is... Represented as: (14) in, This represents the state update gain matrix; The prior state obtained from IMU propagation is corrected using the error state correction to obtain the updated posterior state: (15) in, The prior state obtained through IMU state propagation; This represents the posterior state after the tight combination update of 5G and IMU, i.e., the posterior state after introducing 5G ranging constraints.

8. The GNSS compact combination positioning method with bidirectional constraints and fast reconvergence according to claim 7, characterized in that, In step 6, the process of performing fast reconvergence is specifically as follows: During the GNSS signal recovery phase, the posterior state updated by the tight combination of 5G and IMU in the previous epoch is used as the prior state of the current epoch. The recovered original GNSS observations and 5G ranging observations are jointly introduced into the unified state space model constructed by formulas (1) to (5), and a joint observation update equation is constructed based on the prior state obtained from IMU state propagation: (16) in, This represents the residual from joint GNSS and 5G observations. Represents the joint observation design matrix; Represents the joint observation noise vector; A progressive fusion strategy for GNSS observations is adopted, and the weights of the recovered GNSS observations are dynamically adjusted. The GNSS observation weight adjustment coefficient is expressed as follows: (17) in, For the GNSS signal recovery phase GNSS observation weight adjustment coefficients corresponding to each epoch; This represents the number of consecutive valid epochs since the GNSS signal was recovered. This represents the time constant that controls the rate of gradual change of GNSS observation weights; By dynamically adjusting the GNSS observation noise covariance matrix, the gradual recovery of GNSS observation weights can be achieved. (18) in, Indicates the GNSS signal recovery phase. The GNSS observation noise covariance matrix for each epoch; This represents the standard observation noise covariance matrix corresponding to the normal and stable phase of GNSS.

9. A GNSS compact positioning system for implementing the bidirectional constraint and fast reconvergence GNSS compact positioning method as described in claim 1, characterized in that, The bidirectional constrained and fast reconvergence GNSS compact combination positioning system includes: The observation sequence construction module is used to collect multi-source observation data in real time through a positioning terminal that integrates a GNSS receiver, an IMU sensor, and a 5G communication module, and construct a continuous multi-epoch observation sequence. The data preprocessing module is used to screen the validity of multi-source observation data in the observation sequence and remove gross errors; The error correction module is used to model and correct errors in multi-source observation data in the observation sequence. The state vector construction module is used to construct tightly combined state vectors of GNSS, IMU and 5G in a unified state space; The State Propagation and Observation Model Building Module is used to build state propagation models and observation models; And a joint positioning solution module, which is used to perform joint positioning solution based on the state propagation model and the observation model to obtain the values ​​of the tightly combined state vector of GNSS, IMU and 5G, including three-dimensional position, velocity and attitude and positioning quality index information; During the joint positioning solution process, when the GNSS signal is available, the IMU is used to constrain the 5G abnormal ranging; when the GNSS signal is rejected or degraded, the 5G is used to constrain the IMU drift; and when the GNSS signal is restored, a fast reconvergence process is performed.

Images