A satellite / inertial deep coupling method with a multipath error estimator

Abstract

The present invention discloses a satellite / inertial deep coupling method with a multipath error estimator, which is applied to the combined navigation problem of a global satellite navigation system and an inertial navigation system in an urban canyon environment and the dynamic short-delay multipath problem. First, initialize various parameters; perform relevant operations and coherent integration; calculate the multipath error and the cumulative mean of the multipath error using the coherent integration result; correct the pseudorange measurement and perform positioning solution; construct a line-of-sight vector using the position and velocity information after satellite / inertial combination and estimate the true pseudorange; estimate the unmodeled error of the pseudorange using Kalman filtering; predict the code phase and construct a closed-loop feedback loop; reset the position component of the filtered state value to zero, and repeat the above steps until all sampled data is processed. This method has strong implementability, can improve the tracking accuracy and robustness of the satellite receiver tracking loop, reduce the risk of continuous transmission of multipath errors, and thus improve the accuracy and availability of positioning information.

Description

Technical Field

[0001] The present invention belongs to the field of satellite and integrated navigation positioning, and particularly relates to a satellite / inertial deep coupling method with a multipath error estimator. Background Art

[0002] In high-precision satellite navigation positioning applications, especially in the increasing number of in-vehicle applications in dense urban canyon environments, dynamic short-delay multipath effects are the main reasons affecting satellite positioning accuracy. The frequently occurring dynamic short multipaths and accompanying signal attenuation in dense urban canyon environments impose a heavy burden on the satellite receiver tracking loop, and its tracking robustness is severely challenged, resulting in serious degradation or even unavailability of satellite measurements. Existing multipath suppression and elimination methods cannot well solve the problem of dynamic short multipaths for two reasons. First, existing methods have poor effects in suppressing short-delay multipaths, but have better effects in processing medium- and long-delay multipaths. Second, the structures of existing methods proposed for short-delay multipaths are relatively complex, and the huge computational amount cannot meet the real-time requirements of dynamic multipaths. Vector tracking has inherent advantages in multipath suppression by means of mutual assistance between different tracking channels and effective utilization of spatial geometric relationships. However, due to the high requirement of the loop feedback of vector tracking for the accuracy of positioning results, a single vector receiver solution is difficult to meet its own working conditions in situations such as limited satellite visibility, weak signals, multipath interference, and signal occlusion in dense urban canyon environments. Therefore, the position information provided by external sensors, such as inertial navigation systems, will bring significant enhancement assistance effects to vector tracking receivers. However, existing satellite / inertial deep coupling systems based on vector tracking are not specifically designed for multipath problems, and most still utilize their inherent advantages. And vector tracking and deep coupling have a natural internal need for accurate modeling and estimation of multipath errors when constructing tracking feedback. The method of usually treating multipath errors as noise and processing them together with other error sources cannot meet this need. In this case, the deep coupling system is again at risk of performance degradation. Summary of the Invention

[0003] Object of the Invention: Aiming at the above defects, the present invention provides a satellite / inertial deep coupling method with a multipath error estimator, which can real-time estimate the residual multipath error in satellite pseudorange measurements with a small adjustment of the receiver structure and computational amount, thereby improving the positioning solution accuracy. At the same time, in a closed-loop manner, the feedback information of the tracking loop is constructed according to the line-of-sight vector, and the estimated multipath error is used to compensate this information, thereby improving the tracking accuracy and robustness and reducing the risk of continuous transmission of multipath errors.

[0004] The technical solution adopted by the present invention to solve the above technical problems is as follows: A satellite / inertial deep coupling method with a multipath error estimator includes the following steps:

[0005] (1) Initialize various parameters, including the cumulative mean of multipath error, the estimated value of the filtering state, and the sliding time window K win ;

[0006] (2) Start a data processing cycle T. During this data processing cycle, use the locally reproduced carrier signal and pseudo-random code signal to perform a correlation operation with the received signal, and perform coherent integration on the result;

[0007] (3) Calculate the multipath error using the coherent integration result and perform low-pass filtering, and then subtract the cumulative mean of the multipath error from the multipath error;

[0008] (4) Use the obtained multipath error and other modeling errors to correct the satellite pseudorange measurement, and use the corrected pseudorange for positioning solution. Other modeling errors specifically include modeling other optional error sources such as satellite clock error, ionospheric error, tropospheric error, and satellite orbit error according to requirements;

[0009] (5) Use the result of satellite positioning solution to perform data fusion with the position and velocity information given by the inertial navigation system, use the combined position and velocity information to construct a line-of-sight vector with the satellite position and velocity broadcast by the ephemeris, and estimate the true pseudorange and pseudorange rate;

[0010] (6) Use the difference between the corrected measured pseudorange and the estimated pseudorange, and the difference between the measured pseudorange rate and the estimated pseudorange rate as the Kalman filter observation quantity to estimate the unmodeled error in the pseudorange;

[0011] (7) Use the estimated values of multipath error, other modeling errors, and unmodeled errors to predict the code phase and construct a closed-loop feedback loop;

[0012] (8) Reset the position component in the filtering state value to zero and enter the next data processing cycle. Repeat steps (1) to (8) until all sampled data is processed.

[0013] In step (3), the specific calculation method for calculating the multipath error using the coherent integration result is:

[0014]

[0015] where M k is the multipath error at the k-th T moment, D is twice the correlator spacing, I0 is the coherent integration result of the prompt branch, I2 is the coherent integration result of the twice-late branch, and the delay of the I2 branch relative to the I0 branch is D. α is a scaling coefficient obtained by fitting the multipath error envelope.

[0016] In step (3), the specific calculation method for the cumulative mean of the multipath error is:

[0017] Perform multipath determination based on the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio changes suddenly or is lower than a certain threshold, it is determined that there is a multipath signal, and the cumulative mean of the multipath error at the current moment is the cumulative mean of the multipath error at the previous moment. Otherwise, it is determined that there is no multipath signal, and the cumulative mean of the multipath error is updated to the cumulative value of the multipath error M within the sliding time window K win within the multipath error M k relative to K win mean value of

[0018] In step (4), the satellite pseudorange measurement is corrected using the obtained multipath error and other modeling errors. The specific method is as follows:

[0019] Perform multipath determination based on the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio changes suddenly or is lower than a certain threshold, it is determined that there is a multipath signal, and the multipath error and other modeling errors are subtracted from the pseudorange measurement. Otherwise, it is determined that there is no multipath signal, and other modeling errors are subtracted from the pseudorange measurement.

[0020] In step (5), the line-of-sight vector is constructed using the combined position and velocity information and the satellite position and velocity broadcast by the ephemeris, and the true pseudorange and pseudorange rate are estimated. The specific method is as follows:

[0021]

[0022]

[0023]

[0024]

[0025] Among them, and are the estimates of the true pseudorange and pseudorange rate of the m-th satellite at the k-th T moment respectively, t ub,k and t ud,k are the clock bias and clock drift of the user at the k-th T moment respectively, and are the magnitude of the line-of-sight vector from the user to the satellite and the magnitude of the change rate of the line-of-sight vector respectively, (x k , y k , z k ) and are the three-dimensional position and three-dimensional velocity of the user respectively, and are the three-dimensional position and three-dimensional velocity of the satellite respectively.

[0026] In step (6), the difference between the corrected measured pseudorange and the estimated pseudorange, and the difference between the measured pseudorange rate and the estimated pseudorange rate are used as the Kalman filter observables to estimate the unmodeled error in the pseudorange. The specific method is as follows:

[0027] (6-1) Use the corrected measured pseudorange Estimated pseudorange Measured pseudorange rate And the estimated pseudorange rate Construct a 2m×1-dimensional Kalman filter measurement Z k :

[0028]

[0029] (6-2) The Kalman filter state quantity X k Consists of three-dimensional position errors (Δx k , Δy k , Δz), three-dimensional velocity errors The error Δt of the user clock offset ub,k And the error Δt of the user clock drift ud,k Composed of:

[0030]

[0031] (6-3) The state transition matrix Φ of the Kalman filter k+1,k And the measurement matrix H k Are respectively

[0032]

[0033]

[0034] Where I and 0 are the identity matrix and the zero matrix respectively. Are the three components of the unit line-of-sight vector Can be specifically expressed as The estimated value of the three-dimensional position error obtained by the Kalman filter After Transferred by Is the estimate of the unmodeled error in the pseudorange.

[0035] In step (7), the predicted value of the code phase is used with the estimated values of the multipath error, other modeled errors, and unmodeled errors to construct a closed-loop feedback loop. The specific steps are as follows:

[0036] (7-1) Calculate the estimated pseudorange at the next moment Perform the calculation

[0037]

[0038] Among them, is the magnitude of the line-of-sight vector from the user at the (k + 1)-th T moment to the m-th satellite, and are the error estimates of the user clock offset and the user clock drift respectively.

[0039] (7-2) predicts the code phase at the (k + 1)-th T moment and

[0040]

[0041] where is the code phase at the k-th T moment. is fed back to the code loop controller to generate a new local replica pseudorandom code, forming a closed-loop control loop.

[0042] Compared with the prior art, the advantage of the present invention is that it can estimate the residual multipath error in the pseudorange measurement with a small structure and computational complexity, and reduce the influence of the multipath error on the positioning result by correcting the pseudorange measurement. At the same time, this multipath error is added to the prediction of the code phase when constructing the vector feedback tracking loop to accurately track the received signal with multipath error, thereby blocking the propagation of the multipath error to the next processing cycle. At the same time, taking advantage of the fact that the multipath error in the positioning result is smaller than that in the pseudorange, a closed-loop control quantity of the code loop is constructed by means of line-of-sight vector feedback to track the direct signal in the multipath reception. At the same time, taking advantage of the fact that the inertial navigation system is not affected by the external environment, the multipath error in the satellite positioning result is further reduced through information fusion. Combining the above features and advantages, compared with traditional methods such as narrow correlation, this method has a better suppression effect on dynamically changing short-delay multipath. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flowchart of the method according to an embodiment of the present invention.

[0044] Figure 2 is a flowchart of the method for calculating the cumulative mean of the multipath error. DETAILED DESCRIPTION OF THE INVENTION

[0045] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. After reading the present invention, those skilled in the art's various equivalent forms of modification of the present invention all fall within the scope defined by the appended claims of this application.

[0046] The main idea of the present invention is as follows: The least squares estimation and minimum variance estimation criteria determine that the multipath error in the position solution result is less than the pseudorange multipath error in a certain channel. Therefore, it is beneficial to use this position information to construct the loop feedback control parameter according to the line-of-sight vector for the multipath channel to track the direct signal. Using the position information provided by the inertial navigation system for data fusion can further reduce the multipath error in the position, and at the same time can overcome the problem that a single satellite receiver solution cannot provide reliable high-precision positioning data for vector feedback in an urban canyon environment. On this basis, using a multipath error estimator to further enhance the performance of this structure, and taking advantage of the inherent requirement of vector tracking deep coupling to accurately model each error, and organically integrating the two can further improve the dynamic short-delay multipath suppression ability of the system in a dense urban canyon environment.

[0047] As Figure 1 shown, a satellite / inertial deep coupling method with a multipath error estimator disclosed in an embodiment of the present invention includes the following steps:

[0048] S1: Initialize various parameters, including the cumulative mean of the multipath error, the estimated value of the filtering state, and the sliding time window K win .

[0049] S2: At the start of a data processing period T, within this data processing period, use the locally reproduced carrier signal and pseudorandom code signal to perform a correlation operation with the received signal, and perform coherent integration on the result.

[0050] S3: Calculate the multipath error using the coherent integration result and perform low-pass filtering, and then subtract the cumulative mean of the multipath error from the multipath error, specifically including

[0051] S3-1: The calculation of the multipath error using the coherent integration result, the specific calculation method is:

[0052]

[0053] where M k is the multipath error at the k-th T moment, D is twice the correlator spacing, I0 is the coherent integration result of the prompt branch, I2 is the coherent integration result of the double-delay branch, and the delay of the I2 branch relative to the I0 branch is D. α is a scaling coefficient obtained by fitting the multipath error envelope. Here, D can be selected as 0.2 chips using the narrow correlation technique. Taking the GPS L1 C / A signal as an example, the value of α is 0.18, and the low-pass filter can be selected as a second-order or fourth-order Butterworth filter with a cut-off frequency of 1 Hz or other low-pass filters.

[0054] S3-2: The cumulative mean of the multipath error, as Figure 2 shown, the specific calculation method is:

[0055] Perform multipath judgment based on the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio changes suddenly or is lower than the threshold, and the specific threshold is 42 dBHz, it is determined that there is a multipath signal, and the cumulative mean of the multipath error at the current moment is the cumulative mean of the multipath error at the previous moment That is Otherwise, it is determined that there is no multipath signal, and the cumulative mean of the multipath error is updated to the average value of the cumulative value of the multipath error M win within the sliding time window K k relative to K win which can be approximately calculated as

[0056] S4: Use the obtained multipath error and other modeling errors to correct the satellite pseudorange measurement, and use the corrected pseudorange for positioning solution. Other modeling errors specifically include modeling of satellite clock error, ionospheric error, tropospheric error, satellite orbit error, and other optional error sources according to requirements.

[0057] S4-1: As Figure 2 shown, the method of using the obtained multipath error and other modeling errors to correct the satellite pseudorange measurement is as follows:

[0058] Perform multipath judgment based on the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio changes suddenly or is lower than the threshold, and the specific threshold is 42 dBHz, it is determined that there is a multipath signal, and the multipath error is calculated as and subtract the multipath error and other modeling errors from the pseudorange measurement. Otherwise, it is determined that there is no multipath signal, the multipath error is regarded as zero, and other modeling errors are subtracted from the pseudorange measurement.

[0059] S5: Use the result of satellite positioning solution and the position and velocity information given by the inertial navigation system for data fusion, use the combined position and velocity information and the satellite position and velocity broadcast by the ephemeris to construct the line-of-sight vector, and estimate the true pseudorange and pseudorange rate.

[0060] S5-1: The method of using the combined position and velocity information and the satellite position and velocity broadcast by the ephemeris to construct the line-of-sight vector and estimate the true pseudorange and pseudorange rate is as follows:

[0061]

[0062]

[0063]

[0064]

[0065] Among them, and are the estimates of the true pseudorange and pseudorange rate of the m-th satellite at the k-th T moment, respectively, t ub,k and t ud,k are the clock bias and clock drift of the user at the k-th T moment, respectively and are the magnitude of the line-of-sight vector from the user to the satellite and the magnitude of the rate of change of the line-of-sight vector, respectively. (x k , y k , z k ) and are the three-dimensional position and three-dimensional velocity of the user, respectively and are the three-dimensional position and three-dimensional velocity of the satellite, respectively

[0066] S5-2: The result of satellite positioning solution is fused with the position and velocity information given by the inertial navigation system, which can be realized by loose integration and tight integration methods. Here, the realization method is given taking loose integration as an example as follows:

[0067] The state variables of this combined filter include the position errors in latitude, longitude and altitude δP = [δL, δλ, δh] T , the velocity errors in the east, north and up directions δV = [δv E , δv N , δv U T , the three-axis attitude error the error of the constant bias of the three-axis accelerometer the error of the constant bias of the three-axis gyroscope ε = [ε bx , ε by , ε bz T .

[0068] The measurement of this combined filter is the difference between the position of the inertial navigation system and the position solved by the satellite, and the difference between the velocity of the inertial navigation system and the velocity solved by the satellite. The measurement matrix H and the velocity solved by the satellite . The measurement matrix H k,loose can be expressed as

[0069]

[0070] S6: Using the difference between the calibrated measured pseudorange and the estimated pseudorange, and the difference between the measured pseudorange rate and the estimated pseudorange rate as the Kalman filter observables to estimate the unmodeled errors in the pseudorange. The specific method is as follows:

[0071] ​​(6-1) Use the calibrated measured pseudorange Estimate the pseudorange Measure the pseudorange rate And estimate the pseudorange rate Construct the 2m×1-dimensional Kalman filter measurement Z k :

[0072]

[0073] (6-2) The Kalman filter state quantity X k Consists of the three-dimensional position error (Δx k , Δy k , Δz), the three-dimensional velocity error The error Δt of the user clock offset ub,k And the error Δt of the user clock drift ud,k As follows:

[0074]

[0075] (6-3) The state transition matrix Φ of the Kalman filter k+1,k And the measurement matrix H k Are respectively

[0076]

[0077]

[0078] Where I and 0 are the identity matrix and the zero matrix respectively. Are respectively the three components of the unit line-of-sight vector And can be specifically expressed as The estimated value of the three-dimensional position error obtained by the Kalman filter After Transferred Is the estimate of the unmodeled error in the pseudorange.

[0079] S7: Use the estimated values of multipath error, other modeled errors, and unmodeled errors to predict the code phase and construct a closed-loop feedback loop. The specific steps are as follows:

[0080] (7-1) Calculate the estimated pseudorange at the next moment Perform the calculation

[0081]

[0082] Where Is the magnitude of the line-of-sight vector from the user to the mth satellite at the (k + 1)th T moment, And Are respectively the error estimate of the user clock offset and the error estimate of the user clock drift.

[0083] (7-2) Predict the code phases at k+1 T moments for prediction

[0084]

[0085] where is the code phase at the k-th T moment. Feed back to the code loop controller to generate a new local replicated pseudo-random code, forming a closed-loop control loop.

[0086] S8: Reset the position component in the filtering state value to zero and enter the next data processing cycle. Repeat steps S1 to S8 until all sampled data is processed.

Claims

1. A satellite / inertial deep coupling method with a multipath error estimator, characterized in that It includes the following steps: (1) Initialize each parameter, including the cumulative mean of multipath error, the estimated value of the filtering state, and the sliding time window K win ; (2) A data processing cycle T starts. During this data processing cycle, the locally reproduced carrier signal and the pseudo-random code signal are used to perform a correlation operation with the received signal, and the result is coherently integrated. (3) The multipath error is calculated using the coherent integration result and low-pass filtered, and then the cumulative mean of the multipath error is subtracted from the multipath error. (4) The satellite pseudorange measurement is corrected using the obtained multipath error and other modeling errors, and positioning solution is performed using the corrected pseudorange; the other modeling errors specifically include modeling of satellite clock error, ionospheric error, tropospheric error, and satellite orbit error. (5) The result of satellite positioning solution is fused with the position and velocity information given by the inertial navigation system. The combined position and velocity information and the satellite position and velocity broadcast by the ephemeris are used to construct the line-of-sight vector, and the true pseudorange and pseudorange rate are estimated. (6) The difference between the corrected measured pseudorange and the estimated pseudorange, and the difference between the measured pseudorange rate and the estimated pseudorange rate are used as the Kalman filter observables to estimate the unmodeled error in the pseudorange. (7) The estimated values of multipath error, other modeling errors, and unmodeled errors are used to predict the code phase and construct a closed-loop feedback loop. (8) The position component in the filter state value is reset to zero and enters the next data processing cycle; steps (1) to (8) are repeated until all sampled data is processed. In step (3), the specific calculation method for calculating the multipath error using the coherent integration result is: where M k is the multipath error at the k-th T moment, D is twice the correlator spacing, I0 is the coherent integration result of the prompt branch, I2 is the coherent integration result of the double-delay branch, and the delay of the I2 branch relative to the I0 branch is D; α is a scaling coefficient obtained by fitting the multipath error envelope; In step (3), the specific calculation method for the cumulative mean of the multipath error is: Perform multipath determination based on the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio changes suddenly or is lower than a certain threshold, it is determined that there is a multipath signal, and the cumulative mean of the multipath error at the current moment is the cumulative mean of the multipath error at the previous moment; otherwise, it is determined that there is no multipath signal, and the cumulative mean of the multipath error is updated to the cumulative value of the multipath error M within the sliding time window K win relative to the mean value of K win . win The multipath error M within win k The cumulative value of k win Relative to the mean value of K win .

2. A satellite / inertial deep coupling method with a multipath error estimator according to claim 1, characterized in that In step (4), the specific method for correcting the satellite pseudorange measurement using the obtained multipath error and other modeling errors is: Multipath is judged according to the carrier-to-noise ratio of the signal. If the carrier-to-noise ratio suddenly changes or is lower than a certain threshold, it is determined that there is a multipath signal, and the multipath error and other modeling errors are subtracted from the pseudorange measurement; otherwise, it is determined that there is no multipath signal, and only other modeling errors are subtracted from the pseudorange measurement.

3. A satellite / inertial deep coupling method with a multipath error estimator according to claim 1, characterized in that In step (5), the specific method for constructing the line-of-sight vector using the combined position and velocity information and the satellite position and velocity broadcast by the ephemeris, and estimating the true pseudorange and pseudorange rate is: wherein, and are respectively the estimated true pseudorange and pseudorange rate of the m-th satellite at the k-th T moment, t ub，k and t ud，k are respectively the clock bias and clock drift of the user at the k-th T moment, and are respectively the magnitude of the line-of-sight vector from the user to the satellite and the magnitude of the line-of-sight vector change rate, (x k , y k , z k ) and are respectively the three-dimensional position and three-dimensional velocity of the user, and are respectively the three-dimensional position and three-dimensional velocity of the satellite.

4. A satellite / inertial deep coupling method with a multipath error estimator according to claim 3, characterized in that In step (6), the specific method for using the difference between the corrected measured pseudorange and the estimated pseudorange, and the difference between the measured pseudorange rate and the estimated pseudorange rate as the Kalman filter observables to estimate the unmodeled error in the pseudorange is: (6-1) Use the calibrated measured pseudorange to estimate the pseudorange the measured pseudorange rate and the estimated pseudorange rate Construct the 2m×1-dimensional Kalman filter measurement Z k : (6-2) Kalman filter state quantity X k Composed of three-dimensional position error (Δx k , Δy k , Δz), three-dimensional velocity error Error Δt of user clock deviation ub，k And error △t of user clock drift ud，k Composition: (6-3) State transition matrix φ of Kalman filter k+1，k and measurement matrix H k respectively Wherein, I and 0 are the identity matrix and the zero matrix, respectively; are the three components of the unit line-of-sight vector respectively, and are specifically expressed as The three-dimensional position error estimated value obtained by Kalman filtering After Transmitted to obtain That is the estimate of the unmodeled error in the pseudorange.

5. A satellite / inertial deep coupling method with a multipath error estimator according to claim 4, characterized in that In step (7), the specific steps for predicting the code phase using the estimated values of multipath error, other modeling errors, and unmodeled errors and constructing a closed-loop feedback loop are: (7-1) Estimated pseudorange for the next moment Perform calculation Among them, is the magnitude of the line-of-sight vector from the user at the (k + 1)-th T moment to the m-th satellite, and are the error estimates of the user clock offset and the error estimates of the user clock drift, respectively; (7-2) Predict the code phase at k+1 T moments Perform prediction wherein is the code phase at the k-th T moment; and is fed back to the code loop controller to generate a new local replicated pseudo-random code, forming a closed-loop control loop.

Images