An indoor radio frequency positioning carrier signal phase anomaly data detection and repair method

Abstract

The application discloses an indoor radio frequency positioning carrier signal phase abnormal data detection and repair method, and belongs to the technical field of indoor high-precision positioning. The method realizes the detection and repair of indoor radio frequency positioning carrier signal phase abnormal data by using UWB positioning signal assistance. In the abnormal carrier data detection process, the Kalman filter algorithm is combined with the traditional least square method to establish a polynomial fitting method based on Kalman filter preprocessing, which can reduce the detection of false abnormal data and realize more accurate detection and repair. In the carrier data repair process, the UWB observation data is used to assist the fixing of the whole week part, thereby improving the reliability of the repair. The application solves the problem that the indoor radio frequency positioning carrier signal data is frequently abnormal due to the complex building structure and the changeable actual environment, and makes the positioning result more accurate and reliable.

Description

Technical Field

[0001] This invention relates to a method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals, belonging to the field of indoor high-precision positioning technology. Background Technology

[0002] With continuous urban development, the economic value of underground space is becoming increasingly prominent. The emergence of underground shopping malls, subway stations, and integrated utility tunnels has made location-based services (LBS) supported by high-precision positioning a research hotspot. Compared with open above-ground environments, underground positioning is more complex in terms of signal propagation models, spatial layout, and motion constraints, posing significant challenges to high-precision underground positioning services. The shortcomings of existing solutions and the surge in industry demand have spurred academia and industry to explore and research new high-precision indoor positioning technologies. High-precision indoor radio frequency (RF) positioning systems were proposed against this backdrop. RF positioning systems can provide carrier phase observations, achieving centimeter-level positioning accuracy through carrier phase differential positioning. Currently, many applications and terminals integrate GNSS (Global Navigation Satellite System) positioning functionality. Indoor RF receivers, due to their structural consistency with GNSS receivers, have significant advantages in hardware availability and seamless positioning both above and below ground. Summary of the Invention

[0003] Purpose of the Invention: To address the issue of inaccurate positioning in complex indoor radio frequency (RF) positioning environments due to frequent carrier phase data anomalies, this invention aims to solve the technical problem of accurately detecting and repairing the location of phase anomalies in indoor RF positioning carrier signals. This invention proposes a method for detecting and repairing phase anomaly data in indoor RF positioning carrier signals.

[0004] The method of the present invention to solve the above problems is achieved through the following technical solution:

[0005] Step 1: Obtain the observed values ​​of the indoor radio frequency positioning carrier signal phase and UWB observation values;

[0006] Step 2: Select m observations of the phase of the radio frequency positioning carrier signal, perform Kalman filtering preprocessing on them, and take the observed values ​​of the phase of the radio frequency positioning carrier signal after processing as the fitting data;

[0007] Step 3: Use the fitting data obtained in Step 2 to establish a fitting polynomial, and use the least squares method to estimate the coefficients of the fitting curve.

[0008] Step 4: Use the fitting polynomial obtained in Step 3 to extrapolate the observed value of the radio frequency positioning carrier signal phase of the next epoch, and compare the extrapolated value with the actual value. By observing whether the difference exceeds the allowable range, it is determined whether the data at this epoch is abnormal.

[0009] Step 5: Determine whether the abnormal data inspection has been completed for the observation values ​​of the radio frequency positioning carrier signal phase of all epochs to be measured.

[0010] In step 1, the UWB observations are obtained by quickly acquiring the distance measurements between the user receiver and each base station using the TOF positioning method;

[0011] The equation for the phase observation of the indoor radio frequency positioning carrier signal is shown in Equation 1:

[0012]

[0013] In the formula, This represents the phase observation value of the indoor radio frequency positioning carrier signal, where N is the undetermined integer ambiguity. It is the integer part of the observation. It refers to the portion of the carrier phase value received by the receiver that is less than one revolution.

[0014] In step 2, the observed values ​​of the first m radio frequency positioning carrier signal phases are selected and subjected to Kalman filtering. The Kalman filtering discrete system model is shown in equation (2):

[0015]

[0016] In the formula, X k Let X be the n-dimensional state vector of the system at time k; k-1 Let A be the n-dimensional state vector of the system at time k-1; k,k-1 W is the dimensional state transition matrix of the system. k-1 The noise sequence generated by the system during this process has a mean of zero and satisfies the condition that the covariance matrix Q... k It conforms to a multivariate normal distribution, i.e., W k ~N(0,Q) k );y k It is an m-dimensional measurement vector; C k V is a dimensional observation matrix; k With a mean of zero, the covariance matrix is ​​R k The observed noise sequence V follows a normal distribution. k ~N(0,R k Meanwhile, assume the initial state and the noise at each time step {X0, W1, W2, ..., W...}. k ,V1,V2,…,V k Each of these entities exists independently.

[0017] State vector X k minimum mean square error estimate The following system of recursive equations is used to obtain:

[0018]

[0019] In the formula, A k,k-1 It is an n×n dimensional non-singular one-step transition matrix; C k It is an m×n dimensional observation matrix. It is the state vector X k-1 The minimum mean square error estimate, K k It is the Kalman filter gain, P k|k-1 It is the predicted value of covariance. It is the transpose of the m×n dimensional observation matrix, P k It is the filter covariance. Q is the transpose of the one-step transition matrix of an n×n dimensional nonsingular state. k-1 It is state noise, and the filter error covariance matrix P0 = pI, where p is a large positive number so as to cover the maximum range of variation of the initial estimation error; I is the identity matrix of the corresponding dimension.

[0020] View the measurement data as a time series, i.e., Y = {y k} k∈Z The measurement data series after Kalman filtering is the new measurement data series. Replace the original measurement data column.

[0021] In step 3, the polynomial fitting model is shown in equation (4):

[0022]

[0023] In the formula, i = 1, 2, ..., p, p ≥ q + 1, a i The coefficients of the polynomial are obtained using the least squares method; t i The observation time for the fitted epoch; t o Let q be an initial time; q be the fitting order; p be the number of epochs participating in one fitting. Since the fourth or fifth derivative generally tends to zero and cannot be fitted by a polynomial, the order q in polynomial fitting is generally taken as 3 to 4.

[0024] In step 4, the mean square error σ is calculated using equation (5). j Solving

[0025]

[0026] In the formula, j represents the j-th fitting; V iThe residuals after each fitting;

[0027] When the comparison result is: Δ j+m+1 <3σ j If Δ is not an outlier, then the observed value is determined to be non-outlier data. j+m+1 σ is the difference between the actual value of the (m+1)th observation in the j-th fitting and the predicted value obtained from the polynomial fitting. j Let be the error of the j-th fitting;

[0028] When the comparison result is: Δ j+m+1 ≥3σ j If the phase observation value of the observation epoch is determined to be abnormal data, then the positioning result obtained by referring to the UWB observation value in step 1 is used for initialization to obtain the initial value of the floating-point solution of the carrier phase integer ambiguity, thereby realizing the auxiliary fixation of the integer part, while the part that is less than one cycle still uses the part in the actual observation value.

[0029] In step 5, the position of the current detection epoch is determined. If the current detection epoch is not the last epoch, it indicates that the signal verification has not been completed. At this time, the value of the first epoch is discarded, the value of the next epoch is added, and the fitting program continues. If the current detection epoch is the last epoch, it indicates that the verification has been completed, and the program ends.

[0030] The technical effects and advantages of the present invention are as follows:

[0031] Indoor radio frequency positioning signals are frequently affected by complex environments and numerous interference factors, resulting in abnormal carrier phase observations and a significant decrease in positioning accuracy. This invention proposes a method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals. It combines the Kalman filtering algorithm with the traditional least squares method, establishing a polynomial fitting method based on Kalman filtering preprocessing. This method reduces the detection of false anomalies, achieving more accurate verification and repair. Furthermore, it uses UWB observation data to assist in fixing the integer portion, improving the reliability of anomaly data repair. Attached Figure Description

[0032] To more clearly illustrate the technical solutions and specific embodiments of the present invention, the accompanying drawings used in the detailed description of the technical solutions and specific embodiments will be briefly introduced below. Obviously, the drawings listed below are only a part of the embodiments of the present invention. Without creative modifications, drawings of other embodiments can be obtained based on these drawings.

[0033] Figure 1 This is an overall flowchart of the indoor radio frequency positioning carrier signal phase abnormality data detection and repair method of the present invention.

[0034] Figure 2These are the steps of the indoor radio frequency positioning carrier signal phase abnormality data detection and repair method of the present invention.

[0035] Figure 3 This is a diagram showing the results of applying the indoor radio frequency positioning carrier signal phase anomaly data detection and repair method. Detailed Implementation

[0036] The following description, in conjunction with the accompanying drawings, further illustrates one embodiment of the present invention. Clearly, the described embodiments are merely a subset of the present invention, provided as examples to illustrate the technical solutions. All other embodiments without inventive modifications are within the scope of protection of the present invention.

[0037] like Figure 2 As shown in the figure, this embodiment provides a method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals. The method includes the following steps:

[0038] Step 1: Obtain the observed values ​​of the indoor radio frequency positioning carrier signal phase and UWB observation values;

[0039] Step 2: Select m observations of the phase of the radio frequency positioning carrier signal, perform Kalman filtering preprocessing on them, and take the observed values ​​of the phase of the radio frequency positioning carrier signal after processing as the fitting data;

[0040] Step 3: Use the fitting data obtained in Step 2 to establish a fitting polynomial, and use the least squares method to estimate the coefficients of the fitting curve.

[0041] Step 4: Use the fitting polynomial obtained in Step 3 to extrapolate the observed value of the radio frequency positioning carrier signal phase of the next epoch, and compare the extrapolated value with the actual value. By observing whether the difference exceeds the allowable range, it is determined whether the data at this epoch is abnormal.

[0042] Step 5: Determine whether the abnormal data inspection has been completed for the observation values ​​of the radio frequency positioning carrier signal phase of all epochs to be measured.

[0043] In step 1, the UWB ranging value is obtained by quickly acquiring the distance measurement values ​​between the user receiver and each base station using the TOF positioning method. Further calculations can yield the UWB positioning result, which is then initialized to obtain the initial value of the carrier phase integer ambiguity floating-point solution.

[0044] The equation for the phase observation of the indoor radio frequency positioning carrier signal is shown in Equation 1:

[0045]

[0046] In the formula, This represents the phase observation value of the indoor radio frequency positioning carrier signal, where N is the undetermined integer ambiguity. It is the integer part of the observation. It refers to the portion of the carrier phase value received by the receiver that is less than one revolution.

[0047] In step 2, the observations of the first m epochs are selected and subjected to Kalman filtering. The Kalman-filtered discrete system model is shown in equation (2):

[0048]

[0049] In the formula, X k Let X be the n-dimensional state vector of the system at time k; k-1 Let A be the n-dimensional state vector of the system at time k-1; k,k-1 W is the dimensional state transition matrix of the system. k-1 The noise sequence generated by the system during this process has a mean of zero and satisfies the condition that the covariance matrix Q... k It conforms to a multivariate normal distribution, i.e., W k ~N(0,Q) k );y k It is an m-dimensional measurement vector; C k V is a dimensional observation matrix; k With a mean of zero, the covariance matrix is ​​R k The observed noise sequence V follows a normal distribution. k ~N(0,R k Meanwhile, assume the initial state and the noise at each time step {X0, W1, W2, ..., W...}. k ,V1,V2,…,V k Each of them exists independently.

[0050] State vector X k minimum mean square error estimate The following system of recursive equations is used to obtain:

[0051]

[0052] In the formula, A k,k-1 It is an n×n dimensional non-singular one-step transition matrix; C k It is an m×n dimensional observation matrix. It is the state vector X k-1 The minimum mean square error estimate, K k It is the Kalman filter gain, P k|k-1 It is the predicted value of covariance. It is the transpose of the m×n dimensional observation matrix, P k It is the filter covariance. Q is the transpose of the one-step transition matrix of an n×n dimensional nonsingular state. k-1It is state noise, and the filter error covariance matrix P0 = pI, where p is a large positive number so as to cover the maximum range of variation of the initial estimation error; I is the identity matrix of the corresponding dimension.

[0053] View the measurement data as a time series, i.e., Y = {y k} k∈Z (Typically non-stationary), the measurement data series after Kalman filtering is used as this new measurement data series. Replace the original measurement data column.

[0054] In step 3, the polynomial fitting model is shown in equation (4):

[0055]

[0056] In the formula, i = 1, 2, ..., p, p ≥ q + 1, a i The coefficients of the polynomial are obtained using the least squares method; t i The observation time for the fitted epoch; t o Let q be an initial time; q be the fitting order; p be the number of epochs participating in one fitting. Since the fourth or fifth derivative generally tends to zero and cannot be fitted by a polynomial, the order q in polynomial fitting is generally taken as 3 to 4.

[0057] In step 4, the mean square error σ is calculated using equation (5). j Solving

[0058]

[0059] In the formula, j represents the j-th fitting; V i The residuals after each fitting;

[0060] When the comparison result is: Δ j+m+1 <3σ j If Δ is not an outlier, then the observed value is determined to be non-outlier data. j+m+1 σ is the difference between the actual value of the (m+1)th observation in the j-th fitting and the predicted value obtained from the polynomial fitting. j Let be the error of the j-th fitting;

[0061] When the comparison result is: Δ j+m+1 ≥3σ j If the phase observation value of the observation epoch is determined to be abnormal data, then the positioning result obtained by referring to the UWB observation value in step 1 is used for initialization to obtain the initial value of the floating-point solution of the carrier phase integer ambiguity, thereby realizing the auxiliary fixation of the integer part, while the part that is less than one cycle still uses the part in the actual observation value.

[0062] Finally, the current epoch of detection is determined. If the current epoch is not the last epoch, it indicates that the signal verification is not yet complete. In this case, the value of the first epoch is discarded, the value of the next epoch is added, and the fitting program continues. If the current epoch is the last epoch, it indicates that the verification is complete, and the program ends.

[0063] The indoor radio frequency positioning carrier signal phase abnormality data detection and repair process in this embodiment corresponds to the above-mentioned indoor radio frequency positioning carrier signal phase abnormality data detection and repair method.

[0064] The carrier phase observation results after applying the indoor radio frequency positioning carrier signal phase anomaly data detection and repair method in this embodiment are shown in the figure below. Figure 3 As shown in the figure. The "-." line represents the predicted carrier phase value at each epoch after polynomial fitting, and the "*" line represents the actual carrier phase value at each epoch. Figure 3 The abnormal data detected at the 62nd, 72nd, and 88th epochs and their corresponding carrier phase deviation values ​​are displayed. Compensating for the corresponding carrier phase deviation values ​​before positioning can yield more accurate positioning results.

[0065] The above embodiments of the present invention are merely examples for clearly illustrating the present invention. Those skilled in the art should understand that the above description is not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is impossible to exhaustively list all embodiments here. All obvious variations or modifications derived from the technical solutions of the present invention are still within the protection scope of the present invention. The scope of protection of the present invention is defined by the claims and their equivalents.

Claims

1. A method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals, characterized in that, The method includes the following steps: Step 1: Obtain the observed values ​​of the indoor radio frequency positioning carrier signal phase and UWB observation values; Step 2: Select m observations of the phase of the radio frequency positioning carrier signal, perform Kalman filtering preprocessing on them, and take the observed values ​​of the phase of the radio frequency positioning carrier signal after processing as the fitting data; Step 3: Use the fitting data obtained in Step 2 to establish a fitting polynomial, and use the least squares method to estimate the coefficients of the fitting curve. Step 4: Use the fitting polynomial obtained in Step 3 to extrapolate the observed value of the radio frequency positioning carrier signal phase of the next epoch, and compare the extrapolated value with the actual value. By observing whether the difference exceeds the allowable range, it is determined whether the data at this epoch is abnormal. Step 5: Determine whether the abnormal data inspection has been completed for the observation values ​​of the radio frequency positioning carrier signal phase of all epochs to be measured.

2. The method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals according to claim 1, characterized in that, The specific method for step 1 is as follows: UWB observations are obtained by rapidly acquiring distance measurements between the user receiver and each base station using the TOF positioning method. The equation for the phase observation of the indoor radio frequency positioning carrier signal is shown in Equation 1: In the formula, This represents the phase observation value of the indoor radio frequency positioning carrier signal, where N is the undetermined integer ambiguity. It is the integer part of the observation. It refers to the portion of the carrier phase value received by the receiver that is less than one revolution.

3. The method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals according to claim 1, characterized in that, The specific method for step 2 is as follows: The observed values ​​of the first m radio frequency positioning carrier signal phases are selected and subjected to Kalman filtering. The discrete system model of the Kalman filter is shown in Equation (2): In the formula, X k Let X be the n-dimensional state vector of the system at time k; k-1 Let A be the n-dimensional state vector of the system at time k-1; k,k-1 W is the dimensional state transition matrix of the system. k-1 The noise sequence generated by the system during this process has a mean of zero and satisfies the condition that the covariance matrix Q... k It conforms to a multivariate normal distribution, i.e., W k ~N(0,Q) k );y k It is an m-dimensional measurement vector; C k V is a dimensional observation matrix; k With a mean of zero, the covariance matrix is ​​R k The observed noise sequence V follows a normal distribution. k ~N(0,R k Meanwhile, assume the initial state and the noise at each time step {X0, W1, W2, ..., W...}. k ,V1,V2,…,V k Each of these entities exists independently. State vector X k minimum mean square error estimate The following system of recursive equations is used to obtain: In the formula, A k,k-1 It is an n×n dimensional non-singular one-step transition matrix; C k It is an m×n dimensional observation matrix. It is the state vector X k-1 The minimum mean square error estimate, K k It is the Kalman filter gain, P k|k-1 It is the predicted value of covariance. It is the transpose of the m×n dimensional observation matrix, P k It is the filter covariance. Q is the transpose of the one-step transition matrix of an n×n dimensional nonsingular state. k-1 It is state noise, and the filter error covariance matrix P0 = pI, where p is a large positive number so as to cover the maximum range of variation of the initial estimation error; I is the identity matrix of the corresponding dimension. View the measurement data as a time series, i.e., Y = {y k } k∈Z The measurement data series after Kalman filtering is the new measurement data series. Replace the original measurement data column.

4. The method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals according to claim 1, characterized in that, The specific method for step 3 is as follows: The polynomial fitting model is shown in equation (4): In the formula, i = 1, 2, ..., p, p ≥ q + 1, a i The coefficients of the polynomial are obtained using the least squares method; t i The observation time for the fitted epoch; t o Let q be an initial time; q be the fitting order; p be the number of epochs participating in one fitting. Since the fourth or fifth derivative generally tends to zero and cannot be fitted by a polynomial, the order q in polynomial fitting is generally taken as 3 to 4.

5. The method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals according to claim 1, characterized in that, The specific method for step 4 is as follows: The mean square error σ is calculated using equation (5). j Solving In the formula, j represents the j-th fitting; V i The residuals after each fitting; When the comparison result is: Δ j+m+1 <3σ j If Δ is not an outlier, then the observed value is determined to be non-outlier data. j+m+1 σ is the difference between the actual value of the (m+1)th observation in the j-th fitting and the predicted value obtained from the polynomial fitting. j Let be the error of the j-th fitting; When the comparison result is: Δ j+m+1 ≥3σ j If the phase observation value of the observation epoch is determined to be abnormal data, then the positioning result obtained by referring to the UWB observation value in step 1 is used for initialization to obtain the initial value of the floating-point solution of the carrier phase integer ambiguity, thereby realizing the auxiliary fixation of the integer part, while the part that is less than one cycle still uses the part in the actual observation value.

6. The method for detecting and repairing abnormal phase data of indoor radio frequency positioning carrier signals according to claim 1, characterized in that, The specific method for step 5 is as follows: The current detection epoch is determined. If the current detection epoch is not the last epoch, it means that the signal verification has not been completed. At this time, the value of the first epoch is discarded and the value of the next epoch is added. The fitting program continues. If the current detection epoch is the last epoch, it means that the verification has been completed and the program ends.

Images