A method and device for anti-interference position estimation under multi-source information constraints

Abstract

The present invention provides an anti-interference position estimation method and device under multi-source information constraints, including obtaining initial position coordinates according to a base station; using an inertial navigation integration algorithm and a pedestrian dead reckoning algorithm to continuously recurse position coordinates; using a bidirectional long short-term memory network to perform error assessment processing on the position information provided by an outdoor satellite positioning source and an indoor WiFi positioning source, and eliminating abnormal observations and false position values; obtaining base station positioning results, and using an anti-error Kalman filter to perform position fusion processing on three-dimensional information, base station positioning results, satellite positioning results, and indoor WiFi positioning results to obtain pedestrian trajectory information. The present invention comprehensively considers the important factors that affect the accuracy of positioning trajectory in a large range of space, models and evaluates the error sources, improves positioning stability from the positioning source level; and organically integrates different positioning sources to achieve the anti-interference requirements of pedestrian positions in a large range of intervals.

Description

Technical Field

[0001] The present invention relates to the field of positioning technology, and in particular to an anti-interference position estimation method and device under multi-source information constraints. Background Art

[0002] Location-based services play an indispensable role in the application of smart terminals, and how to accurately obtain continuous and reliable location information of users in a wide range of scenarios has become the key. However, in large urban spaces, affected by the environment such as high-rise buildings, the signals using satellite positioning are seriously interfered with, resulting in a significant reduction in positioning accuracy. Secondly, due to the emergence of virtual positioning software and network agents, smart terminals receive false positioning results, thus affecting the normal operation of location-based services.

[0003] Therefore, the prior art has defects and needs to be improved and developed. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide an interference-resistant position estimation method and device under multi-source information constraints in response to the above-mentioned defects of the prior art, aiming to solve the problem of low accuracy in obtaining continuous and reliable position information of users in a large-scale scenario in the prior art.

[0005] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0006] An anti-interference position estimation method under multi-source information constraints, comprising:

[0007] The initial position coordinates of the terminal are calculated based on the base station connected to the terminal;

[0008] Using the initial position coordinates as the initial position of the inertial navigation integration algorithm, and using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recursively calculate the position coordinates to obtain three-dimensional information corresponding to the pedestrian trajectory;

[0009] The bidirectional long short-term memory network is used to perform error evaluation on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, and abnormal observations and false location values ​​are eliminated to obtain the satellite positioning results and indoor WiFi positioning results.

[0010] The base station positioning result is obtained, and the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result are fused by using an anti-error Kalman filter to obtain pedestrian trajectory information.

[0011] In one implementation, calculating the initial position coordinates of the terminal according to a base station connected to the terminal includes:

[0012] The terminal sequentially connects to the surrounding adjacent base stations and simultaneously obtains the location information and area identification code information of the surrounding adjacent base stations;

[0013] When the number of neighboring base stations detected by the terminal is less than or equal to three, if it is confirmed according to the area identification code information that the connected neighboring base stations are in the same interval identification code, the initial position coordinates of the terminal are calculated using the three-side positioning algorithm.

[0014] In one implementation, after the terminal sequentially connects to the surrounding neighboring base stations and simultaneously obtains the location information and area identification code information of the surrounding neighboring base stations, the method further includes:

[0015] When the number of neighboring base stations detected by the terminal is greater than three, the signal arrival time is calculated to obtain the distance information between the terminal and the neighboring base stations; or, the signal strength of the neighboring base stations is collected and converted into distance information;

[0016] According to the distance information and the location information of the surrounding adjacent base stations, a circular intersection area is constructed. The formula of the circular intersection area is:

[0017] Among them, (x i ,y i ) is the location information of the detected surrounding neighboring base stations, D toa (i) is the distance information from the terminal to different base stations, and (x, y) is the initial location information of the terminal;

[0018] The initial position coordinates of the terminal are calculated using the least squares method, and the calculation formula of the least squares method is:

[0019] Among them, (x, y) is the initial position coordinate of the terminal,

[0020] In one implementation, the initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm are used to continuously recursively calculate the position coordinates to obtain three-dimensional information corresponding to the pedestrian trajectory, including:

[0021] The acceleration value collected by the accelerometer and the angular velocity value collected by the gyroscope in the inertial odometer are obtained, and an error model is established according to the acceleration value and the angular velocity value. The error model is expressed as:

[0022] Among them, the Represents the acceleration value after modeling, Represents the angular velocity value after modeling, the a kRepresents the raw output of the acceleration value, the w k Represents the raw output of the angular velocity value, is the acceleration zero bias, is the angular velocity zero bias, is the gravity related term, the n a is the Gaussian white noise measured by the accelerometer, the n w is the Gaussian white noise measured by the gyroscope;

[0023] The initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the modeled acceleration value and the modeled angular velocity value are fused by the inertial navigation integration algorithm to solve the real-time three-dimensional information, wherein the three-dimensional information includes: three-dimensional position information, three-dimensional velocity information and three-dimensional attitude information;

[0024] The solution formula for the three-dimensional information is:

[0025]

[0026] Said represents three-dimensional position information, the represents three-dimensional velocity information, the Represents 3D posture information, is the posture matrix; k is the corresponding timestamp information, dk represents the integral, represents the 3D posture information of the previous moment, and Ω is an antisymmetric matrix.

[0027] In one implementation, the initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm are used to perform continuous recursion of the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory, and further include:

[0028] The quasi-static magnetic field information is extracted as the observation quantity of the inertial navigation algorithm to correct the heading error accumulated over time; the heading error is expressed as:

[0029] Among them, the is the posture matrix, is the magnetometer output, the is the quasi-static magnetic field information,

[0030] The external acceleration information is extracted as the observation quantity of the inertial navigation algorithm to correct the roll angle and pitch angle solution errors accumulated over time; the calculation formula for the roll angle and pitch angle solution errors accumulated over time is:

[0031] Among them, the f nis the reference acceleration vector, is the external accelerometer output, n a is the error of external acceleration measurement, represents the current moment attitude matrix, I represents the unit matrix, [ψ×] represents the attitude angle, represents an antisymmetric matrix, the f b represents the carrier acceleration vector;

[0032] The velocity and position increments calculated by the pedestrian tracking algorithm are extracted as observations to correct the position and velocity divergence errors accumulated over time; the position and velocity divergence errors accumulated over time are expressed as:

[0033] Among them, the Represents the position observation obtained based on the step length solution, is the velocity observation obtained based on the step length solution, Indicates the position of the inertial odometer output, Indicates the speed value output by the inertial odometer.

[0034] In one implementation, the bidirectional long short-term memory network is used to perform error evaluation processing on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, and abnormal observation values ​​and false location values ​​are eliminated to obtain the satellite positioning result and the indoor WiFi positioning result, including:

[0035] The step-length heading fusion result provided by the inertial navigation algorithm and the pedestrian track algorithm between the position information provided by the two outdoor satellite positioning sources is extracted. The step-length heading fusion result is expressed as: Among them, the represents the acquired step length, Indicates the collected heading value;

[0036] Extract the position difference of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the position difference is expressed as:

[0037]

[0038] Among them, the Indicates the location coordinates provided by the indoor WiFi positioning source, Indicates the location coordinates provided by the outdoor satellite positioning source;

[0039] Extract the reference speed of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the reference speed is expressed as: in, and Indicates the speed update interval;

[0040] Extract the virtual heading of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the virtual heading is expressed as:

[0041]

[0042] Extract the fingerprint strength difference updated by the indoor WiFi positioning source, and the fingerprint strength difference is expressed as: in, and They represent the real-time RSSI vector and the reference RSSI vector respectively, and β is the number of scanned WiFi base stations;

[0043] Extract the signal-to-noise ratio mean of the outdoor satellite positioning source Among them, SNR η (k) represents the signal-to-noise ratio of each received satellite, and α is the number of satellites scanned;

[0044] After extracting the step-length heading indicator, the position difference updated by the outdoor satellite positioning source and the indoor WiFi positioning source, the reference speed, the virtual heading, the fingerprint strength difference and the signal-to-noise ratio mean, the satellite positioning result and the indoor WiFi positioning result are obtained.

[0045] In one implementation, the base station positioning result is obtained, and the three-dimensional information, the base station positioning result, the satellite positioning result, and the indoor WiFi positioning result are subjected to position fusion processing by using an anti-differential Kalman filter to obtain pedestrian trajectory information, including:

[0046] According to the distance information between the terminal and the surrounding adjacent base stations, the flight time information from the base station to the terminal is calculated, and the constraint relationship and observation equation of position and distance are established as follows:

[0047]

[0048] Among them, the d MEMS,m Represents the distance measurement result provided by the inertial odometer positioning, the d station,m Indicates the ranging result provided by base station positioning;

[0049] The location information provided by the surrounding neighboring base stations, the location information provided by the outdoor satellite positioning source, and the location information provided by the indoor WiFi positioning source are obtained as hybrid observations and fused with the location information provided by the inertial odometer. The expression for the fusion processing is:

[0050] Among them, the represents the position observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the speed observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the position observation value output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm, represents the velocity observation output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm;

[0051] The error compensation algorithm is used to compensate the three-dimensional information output by the inertial odometer in real time to constrain the divergence of the three-dimensional information. The expression is:

[0052]

[0053] Among them, the Represents the three-dimensional position information provided by the inertial odometer, Represents the three-dimensional speed information provided by the inertial odometer, Represents the three-dimensional attitude information provided by the inertial odometer; Indicates the accelerometer zero bias, represents the gyroscope zero bias, and the δ represents the derivative; and represents the corresponding Jacobian matrix;

[0054] The forward trajectory is reversely smoothed using a smoothing algorithm to obtain pedestrian trajectory information; the smoothing algorithm is:

[0055]

[0056] Among them, the represents the optimal smoothing result derived from the forward positioning result, P k-1|k represents the current moment robust Kalman filter covariance matrix derived from the optimal covariance matrix of the previous moment, represents the optimal smoothing result at the previous moment, the P k-1 represents the optimal covariance matrix at the previous moment, the φ k T represents the state matrix, Represents the covariance matrix predicted values.

[0057] The present invention also provides an anti-interference position estimation device under multi-source information constraints, comprising:

[0058] A calculation module, used to calculate the initial position coordinates of the terminal according to the base station connected to the terminal;

[0059] A recursion module, used to use the initial position coordinates as the initial position of the inertial navigation integration algorithm, and to use the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recurse the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory;

[0060] The error processing module is used to use a bidirectional long short-term memory network to perform error evaluation on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, eliminate abnormal observation values ​​and false location values, and obtain satellite positioning results and indoor WiFi positioning results;

[0061] The position fusion module is used to obtain the base station positioning result, and use the anti-error Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result to obtain pedestrian trajectory information.

[0062] The present invention also provides a terminal, comprising: a memory, a processor, and an anti-interference position estimation program under multi-source information constraints stored in the memory and executable on the processor, wherein the anti-interference position estimation program under multi-source information constraints, when executed by the processor, implements the steps of the anti-interference position estimation method under multi-source information constraints as described above.

[0063] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program can be executed to implement the steps of the anti-interference position estimation method under multi-source information constraints as described above.

[0064] The present invention provides an anti-interference position estimation method and device under multi-source information constraints, the anti-interference position estimation method under multi-source information constraints includes: calculating the initial position coordinates of the terminal according to the base station connected to the terminal; using the initial position coordinates as the initial position of the inertial navigation integration algorithm, using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recurse the position coordinates, and obtaining the three-dimensional information corresponding to the pedestrian trajectory; using a bidirectional long short-term memory network to perform error evaluation processing on the position information provided by the outdoor satellite positioning source and the position information provided by the indoor WiFi positioning source, eliminating abnormal observations and false position values, and obtaining satellite positioning results and indoor WiFi positioning results; obtaining the base station positioning result, using the anti-error Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result, and obtaining pedestrian trajectory information. The present invention comprehensively considers the important factors affecting the positioning trajectory accuracy in a large range of space, models and evaluates the error source, and improves the positioning stability from the positioning source level; at the same time, a position fusion algorithm model based on multi-source information constraints is designed to organically fuse different positioning sources, and finally realize the anti-interference requirements of pedestrian positions in a large range. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a flow chart of a preferred embodiment of the anti-interference position estimation method under multi-source information constraints in the present invention.

[0066] Figure 2 This is a schematic diagram of least squares positioning based on 4G / 5G base station ranging.

[0067] Figure 3 It is the 4G / 5G base station interval identification code detected and output by the smart terminal (IOS side).

[0068] Figure 4 It is the 4G / 5G base station interval identification code detected and output by the smart terminal (Android side).

[0069] Figure 5 It is a structural diagram of a bidirectional long short-term memory network.

[0070] Figure 6 It is a comparison chart of sensor positioning trajectories.

[0071] Figure 7 This is a comparison chart of sensor positioning errors.

[0072] Figure 8 It is a comparison chart of the combined positioning trajectory results.

[0073] Fig. 9 This is a comparison chart of the combined positioning accuracy results.

[0074] Fig.10 It is a functional principle block diagram of a preferred embodiment of the anti-interference position estimation device under multi-source information constraints in the present invention.

[0075] Fig.11 It is a functional principle block diagram of the terminal in the present invention. DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0077] At present, due to the complexity of urban positioning scenarios and the emergence of virtual positioning software and network agent technologies, the existing positioning technology based on smart terminals in large-scale urban scenarios has the following shortcomings: (1) Lack of effective multi-source fusion methods to comprehensively utilize existing positioning sources; (2) Lack of effective error estimation and detection algorithms to deal with the interference of abnormal data on positioning results; (3) Lack of effective error elimination and large-scale trajectory reconstruction algorithms.

[0078] The present invention proposes a multi-source information constrained intelligent terminal anti-interference position estimation method to improve the positioning and trajectory reconstruction accuracy of intelligent terminals in a large-scale urban space. At the same time, it can effectively detect and eliminate interfered satellite / WiFi observation values ​​and virtual position coordinates. It can be applied to computers, smart phones, Internet of Things terminals and other devices. By collecting, processing, and fusing multi-source sensor and wireless signal data, it can detect and eliminate interfered position information or virtual position information in real time. Finally, the positioning error is further removed through a reverse smoothing algorithm, and finally high-precision trajectory measurement in a large-scale urban space is achieved.

[0079] The embodiment of the present invention first obtains the distance information of the mobile terminal from the surrounding 4G / 5G base stations, and then uses the weighted least squares algorithm combined with the satellite navigation positioning algorithm to realize the high-precision estimation of the initial position of the terminal. After obtaining the initial position, the inertial navigation algorithm is combined with the pedestrian dead reckoning algorithm to realize the continuous recursion of the position coordinates, including the basic positioning model based on the inertial navigation mechanical arrangement and the position and speed constraint model based on the pedestrian track algorithm. Then, the bidirectional long short-term memory network is used to evaluate the error of the position information provided by the outdoor satellite positioning source and the indoor WiFi positioning source, and the abnormal observation values ​​and false position values ​​are eliminated to eliminate the sudden jump error and false observation, and improve the signal robustness at the signal source level. Finally, the anti-error Kalman filter is used to fuse the inertial navigation algorithm, the pedestrian dead reckoning algorithm, the 4G / 5G communication base station positioning results, the satellite positioning results, and the WiFi positioning results to realize the continuous estimation of the pedestrian position in a large range of scenes. In addition, the trajectory in a large range of space is reproduced by the trajectory reconstruction algorithm based on reverse smoothing, and finally the high-precision trajectory measurement in a large range of urban space is realized.

[0080] See also Figure 1 , Figure 1 : is a flow chart of the anti-interference position estimation method under multi-source information constraints in the present invention. Figure 1 As shown, the anti-interference position estimation method under multi-source information constraints described in an embodiment of the present invention includes the following steps:

[0081] Step S100: Calculate the initial location coordinates of the terminal according to the base station connected to the terminal.

[0082] Specifically, 4G / 5G communication base stations are used to achieve initial position estimation, including base station interval identification and initial position estimation based on a three-sided positioning algorithm.

[0083] In one implementation, step S100 specifically includes: the terminal connects to the surrounding neighboring base stations in sequence, and simultaneously obtains the location information and area identification code information of the surrounding neighboring base stations; when the number of surrounding neighboring base stations detected by the terminal is less than or equal to three, if it is confirmed according to the area identification code information that the connected surrounding neighboring base stations are in the same interval identification code, the initial position coordinates of the terminal are calculated using the three-side positioning algorithm.

[0084] Specifically, the terminal first obtains the location information and area identification code information of the currently connected 4G / 5G base station through the terminal, and divides them. After confirming that the current communication base station is in the same interval identification code, the three-side positioning algorithm is used to calculate the initial location information of the terminal. The schematic diagram of the three-side positioning algorithm is shown in the figure. Figure 2 As shown, the 4G / 5G base station identification code recognized by the mobile phone is as follows Figure 3 and Figure 4 shown.

[0085] In one embodiment, after the terminal sequentially connects to the surrounding adjacent base stations and simultaneously obtains the location information and area identification code information of the surrounding adjacent base stations, it also includes: when the number of the surrounding adjacent base stations detected by the terminal is greater than three, the signal arrival time is calculated to obtain the distance information between the terminal and the surrounding adjacent base stations; or, the signal strength of the surrounding adjacent base stations is collected and the signal strength is converted into distance information. That is, after obtaining the location information of more than three adjacent base stations, the distance information D of the intelligent terminal from the current adjacent communication base station is obtained by calculating the signal arrival time. toa =c·T toa , or by collecting the signal strength of the base station and converting it into distance information.

[0086] According to the distance information and the location information of the surrounding adjacent base stations, a circular intersection area is constructed. The formula of the circular intersection area is: Among them, (x i ,y i ) is the location information of the detected surrounding neighboring base stations, D toa (i) is the distance information from the terminal to different base stations, and (x, y) is the initial location information of the terminal.

[0087] The initial position coordinates of the terminal are calculated using the least squares method, and the initial position coordinates are the initial optimal position coordinates. The calculation formula of the least squares method is: Among them, (x, y) is the initial position coordinate of the terminal, b is the construction matrix,

[0088] It can be seen that the mobile terminal position coordinates (x, y) to be solved are:

[0089] When the number of surrounding 4G / 5G base stations is greater than 3, the number of equations obtained is greater than the number of unknowns, which will cause the matrix A to be irreversible. In order to solve this problem, when the number of surrounding 4G / 5G base stations is greater than 3, a pseudo-operator is needed to solve the position estimation problem. The formula at this time can be converted to: The formula can be used to obtain relatively accurate initial position coordinates of the terminal.

[0090] The step S100 is followed by: step S200, using the initial position coordinates as the initial position of the inertial navigation integration algorithm, and using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to perform continuous recursion of the position coordinates to obtain three-dimensional information corresponding to the pedestrian trajectory.

[0091] Specifically, an inertial navigation algorithm is used in combination with a pedestrian dead reckoning algorithm to realize continuous recursion of position coordinates, including a basic positioning model based on inertial navigation mechanical arrangement and a position and speed constraint model based on a pedestrian dead reckoning algorithm.

[0092] In one implementation, step S200 specifically includes:

[0093] The acceleration value collected by the accelerometer and the angular velocity value collected by the gyroscope in the inertial odometer are obtained, and an error model is established according to the acceleration value and the angular velocity value. The error model is expressed as: Among them, the Represents the acceleration value after modeling, Represents the angular velocity value after modeling, the a k Represents the raw output of the acceleration value, the w k Represents the raw output of the angular velocity value, is the acceleration zero bias, is the angular velocity zero bias, is the gravity related term, the n a is the Gaussian white noise measured by the accelerometer, the n w is the Gaussian white noise measured by the gyroscope;

[0094] The initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the modeled acceleration value and the modeled angular velocity value are fused by the inertial navigation integration algorithm to solve the real-time three-dimensional information, wherein the three-dimensional information includes: three-dimensional position information, three-dimensional velocity information and three-dimensional attitude information;

[0095] The solution formula for the three-dimensional information is:

[0096]

[0097] Said represents three-dimensional position information, the represents three-dimensional velocity information, the Represents 3D posture information, is the posture matrix; k is the corresponding timestamp information, dk represents the integral, represents the 3D posture information of the previous moment, and Ω is an antisymmetric matrix.

[0098] In one implementation, in order to further correct the error of the inertial odometer, a multivariate observation is used, and the step S200 further includes:

[0099] The quasi-static magnetic field information is extracted as the observation quantity of the inertial navigation algorithm to correct the heading error accumulated over time; the heading error is expressed as:

[0100] Among them, the is the posture matrix, is the magnetometer output, the is the quasi-static magnetic field information,

[0101] The external acceleration information is extracted as the observation quantity of the inertial navigation algorithm to correct the roll angle and pitch angle solution errors accumulated over time; the calculation formula for the roll angle and pitch angle solution errors accumulated over time is:

[0102] Among them, the f n is the reference acceleration vector, is the external accelerometer output, n a is the error of external acceleration measurement, represents the current moment attitude matrix, I represents the unit matrix, [ψ×] represents the attitude angle, represents an antisymmetric matrix, the f b represents the carrier acceleration vector;

[0103] The velocity and position increments calculated by the pedestrian tracking algorithm are extracted as observations to correct the position and velocity divergence errors accumulated over time; the position and velocity divergence errors accumulated over time are expressed as:

[0104] Among them, the Represents the position observation obtained based on the step length solution, is the velocity observation obtained based on the step length solution, Indicates the position of the inertial odometer output, Indicates the speed value output by the inertial odometer.

[0105] It is also possible to extract the altitude observation error based on the barometer, the calculation formula of the altitude observation error is: Among them, the is the altitude output value based on the barometer, It is the altitude output value of the inertial odometer.

[0106] The step S200 is followed by: step S300, using a bidirectional long short-term memory network to perform error assessment processing on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, eliminating abnormal observation values ​​and false location values, and obtaining satellite positioning results and indoor WiFi positioning results.

[0107] Specifically, a bidirectional long short-term memory network is used to evaluate the error of the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi, and to remove abnormal observations and false location values ​​to eliminate sudden jump errors and false measurement errors. Figure 5 shown.

[0108] In one implementation, step S300 specifically includes:

[0109] The step-length heading fusion result provided by the inertial navigation algorithm and the pedestrian track algorithm between the position information provided by the two outdoor satellite positioning sources is extracted. The step-length heading fusion result is expressed as: Among them, the represents the acquired step length, Indicates the collected heading value;

[0110] Extract the position difference of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the position difference is expressed as:

[0111]

[0112] Among them, the Indicates the location coordinates provided by the indoor WiFi positioning source, Indicates the location coordinates provided by the outdoor satellite positioning source;

[0113] Extract the reference speed of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the reference speed is expressed as: in, and Indicates the speed update interval;

[0114] Extract the virtual heading of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the virtual heading is expressed as:

[0115]

[0116] Extract the fingerprint strength difference updated by the indoor WiFi positioning source, and the fingerprint strength difference is expressed as: in, and They represent the real-time RSSI vector and the reference RSSI vector respectively, and β is the number of scanned WiFi base stations;

[0117] Extract the signal-to-noise ratio mean of the outdoor satellite positioning source Among them, SNR η (k) represents the signal-to-noise ratio of each received satellite, and α is the number of satellites scanned;

[0118] After extracting the step-length heading indicator, the position difference updated by the outdoor satellite positioning source and the indoor WiFi positioning source, the reference speed, the virtual heading, the fingerprint strength difference and the signal-to-noise ratio mean, the satellite positioning result and the indoor WiFi positioning result are obtained.

[0119] The step S300 is followed by: step S400, obtaining the base station positioning result, and using the robust Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result to obtain pedestrian trajectory information.

[0120] Specifically, a unified multi-source fusion positioning model is designed, using the robust Kalman filter to fuse the inertial navigation algorithm, the pedestrian track estimation algorithm, the 4G / 5G communication base station positioning results, the satellite positioning results, and the indoor WiFi positioning results to achieve continuous estimation of pedestrian positions in large-scale urban scenarios.

[0121] In one implementation, step S400 specifically includes:

[0122] The basic algorithm uses the distance information of surrounding base stations obtained by the smart terminal, and establishes the constraint relationship and observation equation of position and distance through the calculated flight time information from the base station to the smart terminal:

[0123]

[0124] Among them, the d MEMS,m Represents the distance measurement result provided by the inertial odometer positioning, the d station,m Indicates the ranging result provided by base station positioning;

[0125] The location information provided by the surrounding neighboring base stations (4G / 5G communication base stations), the location information provided by the outdoor satellite positioning source, and the location information provided by the indoor WiFi positioning source are obtained as hybrid observations and fused with the location information provided by the inertial odometer. The expression for the fusion processing is:

[0126] Among them, the represents the position observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the speed observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the position observation value output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm, represents the velocity observation output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm;

[0127] The error compensation algorithm is used to compensate the three-dimensional information output by the inertial odometer in real time to constrain the divergence of the three-dimensional information. The expression is:

[0128]

[0129] Among them, the Represents the three-dimensional position information provided by the inertial odometer, Represents the three-dimensional speed information provided by the inertial odometer, Represents the three-dimensional attitude information provided by the inertial odometer; Indicates the accelerometer zero bias, represents the gyroscope zero bias, and the δ represents the derivative; and represents the corresponding Jacobian matrix;

[0130] The forward trajectory is reversely smoothed using a smoothing algorithm to further eliminate its accumulated error and effectively restore the trajectory of the person; the smoothing algorithm is:

[0131]

[0132] Among them, the represents the optimal smoothing result derived from the forward positioning result, P k-1|k represents the current moment robust Kalman filter covariance matrix derived from the optimal covariance matrix of the previous moment, represents the optimal smoothing result at the previous moment, the P k-1 represents the optimal covariance matrix at the previous moment, the φ k T represents the state matrix, Represents the covariance matrix predicted values.

[0133] Compared with the existing large-scale urban spatial positioning technology, the embodiments of the present invention comprehensively consider the processing and fusion of multiple positioning source data, the identification and elimination of abnormal data and false position data, and the problems of continuous position estimation and high-precision trajectory reproduction:

[0134] (1) After obtaining the real-time distance information between the surrounding base stations and the mobile terminal, the weighted least squares algorithm is combined with the satellite navigation positioning algorithm to achieve high-precision estimation of the terminal's initial position.

[0135] (2) After obtaining the initial position, the inertial navigation algorithm is used in combination with the pedestrian dead reckoning algorithm to realize the continuous recursion of the position coordinates, including a basic positioning model based on inertial navigation mechanical arrangement and a position and speed constraint model based on the pedestrian dead reckoning algorithm.

[0136] (3) A bidirectional long short-term memory network is used to perform error evaluation on the location information provided by the outdoor satellite positioning source and the indoor WiFi positioning source, and to remove abnormal observations and false position values ​​to eliminate sudden jump errors and false observations, thereby improving signal robustness at the signal source level.

[0137] (4) The robust Kalman filter is used to integrate the inertial navigation algorithm, the pedestrian dead reckoning algorithm, the positioning results of 4G / 5G communication base stations, the satellite positioning results, and the location information provided by indoor WiFi positioning to ultimately achieve continuous estimation of pedestrian positions in a large range of scenarios. In addition, the trajectory reconstruction algorithm based on reverse smoothing is used to reproduce the trajectory in a large space.

[0138] Through the above four points, the embodiment of the present invention can obtain a better anti-interference position estimation result of the intelligent terminal constrained by multi-source information.

[0139] By comparing the positioning accuracy of the combined inertial navigation positioning and pedestrian positioning method (inertial navigation odometer) proposed in the present invention and the traditional pedestrian dead reckoning algorithm, it can be found that the embodiment of the present invention obtains a better positioning result. Figure 6 and 7 shown.

[0140] Figure 8 Further, the combined positioning results of the indoor and outdoor seamless positioning algorithm using the indoor WiFi positioning, inertial navigation odometer, and outdoor GNSS signal involved in the present invention, and the positioning results of the inertial navigation odometer without signal quality evaluation are given. By comparing the positioning effects of several different positioning source combinations and the positioning results of the multi-source fusion indoor and outdoor seamless positioning fusion framework, it can be found that the fusion framework under the multi-source information constraint provided by the embodiment of the present invention can achieve a positioning effect that is better than the conventional positioning source combination, and the result is also closer to the real trajectory. Fig. 9By comparing the positioning accuracies of the corresponding combinations, it can be found that the anti-interference positioning method based on multi-source information constraints provided by the embodiment of the present invention can achieve a positioning accuracy better than 2.16 meters in 75% of cases. Compared with the positioning source using the sensor and WiFi fingerprint, the positioning effect achieved by the GNSS positioning source has been significantly improved, which can effectively meet the anti-interference positioning needs of ordinary people using smart terminals in a large range of urban spaces.

[0141] Furthermore, if Fig.10 As shown, based on the above-mentioned anti-interference position estimation method under multi-source information constraints, the present invention also provides an anti-interference position estimation device under multi-source information constraints, including:

[0142] The calculation module 100 is used to calculate the initial position coordinates of the terminal according to the base station connected to the terminal;

[0143] A recursion module 200 is used to use the initial position coordinates as the initial position of the inertial navigation integration algorithm, and to use the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recurse the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory;

[0144] The error processing module 300 is used to use a bidirectional long short-term memory network to perform error evaluation processing on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, eliminate abnormal observation values ​​and false location values, and obtain satellite positioning results and indoor WiFi positioning results;

[0145] The position fusion module 400 is used to obtain the base station positioning result, and use the robust Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result to obtain pedestrian trajectory information.

[0146] like Fig.11 As shown, the present invention also provides a terminal, comprising: a memory 20, a processor 10, and an anti-interference position estimation program 30 under multi-source information constraints stored on the memory 20 and executable on the processor 10. When the anti-interference position estimation program 30 under multi-source information constraints is executed by the processor 10, the steps of the anti-interference position estimation method under multi-source information constraints as described above are implemented.

[0147] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program can be executed to implement the steps of the anti-interference position estimation method under multi-source information constraints as described above.

[0148] In summary, the present invention discloses an anti-interference position estimation method and device under multi-source information constraints. The anti-interference position estimation method under multi-source information constraints includes: obtaining the initial position coordinates of the terminal according to the base station connected to the terminal; using the initial position coordinates as the initial position of the inertial navigation integration algorithm, and using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recurse the position coordinates to obtain three-dimensional information corresponding to the pedestrian trajectory; using a bidirectional long short-term memory network to perform error evaluation processing on the position information provided by the outdoor satellite positioning source and the position information provided by the indoor WiFi positioning source, eliminating abnormal observations and false position values, and obtaining satellite positioning results and indoor WiFi positioning results; obtaining the base station positioning result, and using an anti-error Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result, and the indoor WiFi positioning result to obtain pedestrian trajectory information. The present invention comprehensively considers the important factors that affect the positioning trajectory accuracy in a large range of space, models and evaluates the error sources, and improves the positioning stability from the positioning source level; at the same time, it designs a position fusion algorithm model based on multi-source information constraints to organically integrate different positioning sources, and finally realizes the anti-interference requirements of pedestrian positions in a large range.

[0149] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. An anti-interference position estimation method under multi-source information constraints, characterized in that: include: The initial position coordinates of the terminal are calculated based on the base station connected to the terminal; Using the initial position coordinates as the initial position of the inertial navigation integration algorithm, and using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recursively calculate the position coordinates to obtain three-dimensional information corresponding to the pedestrian trajectory; The bidirectional long short-term memory network is used to perform error evaluation on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, and abnormal observations and false location values ​​are eliminated to obtain the satellite positioning results and indoor WiFi positioning results. Obtaining the base station positioning result, and using the robust Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result, and the indoor WiFi positioning result to obtain pedestrian trajectory information; The base station is a currently connected 4G or 5G communication base station; The terminal sequentially connects to the surrounding adjacent base stations and simultaneously obtains the location information and area identification code information of the surrounding adjacent base stations; When the number of neighboring base stations detected by the terminal is equal to three, if it is confirmed according to the area identification code information that the connected neighboring base stations are in the same interval identification code, the initial position coordinates of the terminal are calculated using the three-side positioning algorithm; the neighboring base stations are 4G or 5G communication base stations; The obtaining of the base station positioning result, and performing position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result, and the indoor WiFi positioning result by using an anti-differential Kalman filter to obtain pedestrian trajectory information include: According to the distance information between the terminal and the surrounding adjacent base stations, the flight time information from the base station to the terminal is calculated, and the constraint relationship and observation equation of position and distance are established as follows: Among them, the d MEMS,m Represents the distance measurement result provided by the inertial odometer positioning, the d station,m Indicates the ranging result provided by base station positioning; The location information provided by the surrounding neighboring base stations, the location information provided by the outdoor satellite positioning source, and the location information provided by the indoor WiFi positioning source are obtained as hybrid observations and fused with the location information provided by the inertial odometer. The expression for the fusion processing is: Among them, the represents the position observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the speed observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the position observation value output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm, represents the velocity observation output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm; The error compensation algorithm is used to compensate the three-dimensional information output by the inertial odometer in real time to constrain the divergence of the three-dimensional information. The expression is: Among them, the Represents the three-dimensional position information provided by the inertial odometer, Represents the three-dimensional speed information provided by the inertial odometer, Represents the three-dimensional attitude information provided by the inertial odometer; Indicates the accelerometer zero bias, represents the gyroscope zero bias, and the δ represents the derivative; and represents the corresponding Jacobian matrix; The forward trajectory is reversely smoothed using a smoothing algorithm to obtain pedestrian trajectory information; the smoothing algorithm is: Among them, the represents the optimal smoothing result derived from the forward positioning result, P k-1|k represents the current moment robust Kalman filter covariance matrix derived from the optimal covariance matrix of the previous moment, represents the optimal smoothing result at the previous moment, the P k-1 represents the optimal covariance matrix at the previous moment, the φ k T represents the state matrix, Represents the covariance matrix predicted values.

2. The anti-interference position estimation method under multi-source information constraints according to claim 1 is characterized in that: After the terminal sequentially connects to the surrounding adjacent base stations and simultaneously obtains the location information and area identification code information of the surrounding adjacent base stations, the method further includes: When the number of neighboring base stations detected by the terminal is greater than three, the signal arrival time is calculated to obtain the distance information between the terminal and the neighboring base stations; or, the signal strength of the neighboring base stations is collected and converted into distance information; According to the distance information and the location information of the surrounding adjacent base stations, a circular intersection area is constructed. The formula of the circular intersection area is: Among them, (x i ,y i ) is the location information of the detected surrounding neighboring base stations, D toa (i) is the distance information from the terminal to different base stations, and (x, y) is the initial location information of the terminal; The initial position coordinates of the terminal are calculated using the least squares method, and the calculation formula of the least squares method is: Among them, (x, y) is the initial position coordinate of the terminal, 3. The anti-interference position estimation method under multi-source information constraints according to claim 1 is characterized in that: The initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm are used to continuously recursively calculate the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory, including: The acceleration value collected by the accelerometer and the angular velocity value collected by the gyroscope in the inertial odometer are obtained, and an error model is established according to the acceleration value and the angular velocity value. The error model is expressed as: Among them, the Represents the acceleration value after modeling, Represents the angular velocity value after modeling, the a k Represents the raw output of the acceleration value, the w k Represents the raw output of the angular velocity value, is the acceleration zero bias, is the angular velocity zero bias, is the gravity related term, the n a is the Gaussian white noise measured by the accelerometer, the n w is the Gaussian white noise measured by the gyroscope; The initial position coordinates are used as the initial position of the inertial navigation integration algorithm, and the modeled acceleration value and the modeled angular velocity value are fused by the inertial navigation integration algorithm to solve the real-time three-dimensional information, wherein the three-dimensional information includes: three-dimensional position information, three-dimensional velocity information and three-dimensional attitude information; The solution formula for the three-dimensional information is: Said represents three-dimensional position information, the represents three-dimensional velocity information, the Represents 3D posture information, is the posture matrix; k is the corresponding timestamp information, dk represents the integral, represents the 3D posture information of the previous moment, and Ω is an antisymmetric matrix.

4. The anti-interference position estimation method under multi-source information constraints according to claim 3 is characterized in that: The method further comprises: using the initial position coordinates as the initial position of the inertial navigation integration algorithm, and using the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recursively calculate the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory. The quasi-static magnetic field information is extracted as the observation quantity of the inertial navigation algorithm to correct the heading error accumulated over time; the heading error is expressed as: Among them, the is the posture matrix, is the magnetometer output, the is the quasi-static magnetic field information, The external acceleration information is extracted as the observation quantity of the inertial navigation algorithm to correct the roll angle and pitch angle solution errors accumulated over time; the calculation formula for the roll angle and pitch angle solution errors accumulated over time is: Among them, the f n is the reference acceleration vector, is the external accelerometer output, n a is the error of external acceleration measurement, represents the current moment attitude matrix, I represents the unit matrix, [ψ×] represents the attitude angle, represents an antisymmetric matrix, the f b represents the carrier acceleration vector; The velocity and position increments calculated by the pedestrian tracking algorithm are extracted as observations to correct the position and velocity divergence errors accumulated over time; the position and velocity divergence errors accumulated over time are expressed as: Among them, the Represents the position observation obtained based on the step length solution, is the velocity observation obtained based on the step length solution, Indicates the position of the inertial odometer output, Indicates the speed value output by the inertial odometer.

5. The anti-interference position estimation method under multi-source information constraints according to claim 1 is characterized in that: The method uses a bidirectional long short-term memory network to perform error assessment processing on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, removes abnormal observation values ​​and false location values, and obtains satellite positioning results and indoor WiFi positioning results, including: The step-length heading fusion result provided by the inertial navigation algorithm and the pedestrian track algorithm between the position information provided by the two outdoor satellite positioning sources is extracted. The step-length heading fusion result is expressed as: Among them, the represents the acquired step length, Indicates the collected heading value; Extract the position difference of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the position difference is expressed as: Among them, the Indicates the location coordinates provided by the indoor WiFi positioning source, Indicates the location coordinates provided by the outdoor satellite positioning source; Extract the reference speed of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the reference speed is expressed as: in, and Indicates the speed update interval; Extract the virtual heading of different positioning source output results based on the inertial navigation algorithm update and the outdoor satellite positioning source update, and the virtual heading is expressed as: Extract the fingerprint strength difference updated by the indoor WiFi positioning source, and the fingerprint strength difference is expressed as: in, and They represent the real-time RSSI vector and the reference RSSI vector respectively, and β is the number of scanned WiFi base stations; Extract the signal-to-noise ratio mean of the outdoor satellite positioning source Among them, SNR η (k) represents the signal-to-noise ratio of each received satellite, and α is the number of satellites scanned; After extracting the step-length heading indicator, the position difference updated by the outdoor satellite positioning source and the indoor WiFi positioning source, the reference speed, the virtual heading, the fingerprint strength difference and the signal-to-noise ratio mean, the satellite positioning result and the indoor WiFi positioning result are obtained.

6. An anti-interference position estimation device under multi-source information constraints, characterized in that: include: A calculation module, used to calculate the initial position coordinates of the terminal according to the base station connected to the terminal; A recursion module, used to use the initial position coordinates as the initial position of the inertial navigation integration algorithm, and to use the inertial navigation integration algorithm and the pedestrian dead reckoning algorithm to continuously recurse the position coordinates to obtain the three-dimensional information corresponding to the pedestrian trajectory; The error processing module is used to use a bidirectional long short-term memory network to perform error evaluation on the location information provided by the outdoor satellite positioning source and the location information provided by the indoor WiFi positioning source, eliminate abnormal observation values ​​and false location values, and obtain satellite positioning results and indoor WiFi positioning results; A position fusion module is used to obtain the base station positioning result, and use the robust Kalman filter to perform position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result and the indoor WiFi positioning result to obtain pedestrian trajectory information; The base station is a currently connected 4G or 5G communication base station; The terminal sequentially connects to the surrounding adjacent base stations and simultaneously obtains the location information and area identification code information of the surrounding adjacent base stations; When the number of neighboring base stations detected by the terminal is equal to three, if it is confirmed according to the area identification code information that the connected neighboring base stations are in the same interval identification code, the initial position coordinates of the terminal are calculated using the three-side positioning algorithm; the neighboring base stations are 4G or 5G communication base stations; The obtaining of the base station positioning result, and performing position fusion processing on the three-dimensional information, the base station positioning result, the satellite positioning result, and the indoor WiFi positioning result by using an anti-differential Kalman filter to obtain pedestrian trajectory information include: According to the distance information between the terminal and the surrounding adjacent base stations, the flight time information from the base station to the terminal is calculated, and the constraint relationship and observation equation of position and distance are established as follows: Among them, the d MEMS,m Represents the distance measurement result provided by the inertial odometer positioning, the d station,m Indicates the ranging result provided by base station positioning; The location information provided by the surrounding neighboring base stations, the location information provided by the outdoor satellite positioning source, and the location information provided by the indoor WiFi positioning source are obtained as hybrid observations and fused with the location information provided by the inertial odometer. The expression for the fusion processing is: Among them, the represents the position observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the speed observation value output by the indoor WiFi positioning source and the outdoor satellite positioning source, represents the position observation value output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm, represents the velocity observation output by the inertial navigation algorithm and the pedestrian dead reckoning algorithm; The error compensation algorithm is used to compensate the three-dimensional information output by the inertial odometer in real time to constrain the divergence of the three-dimensional information. The expression is: Among them, the Represents the three-dimensional position information provided by the inertial odometer, Represents the three-dimensional speed information provided by the inertial odometer, Represents the three-dimensional attitude information provided by the inertial odometer; Indicates the accelerometer zero bias, represents the gyroscope zero bias, and the δ represents the derivative; and represents the corresponding Jacobian matrix; The forward trajectory is reversely smoothed using a smoothing algorithm to obtain pedestrian trajectory information; the smoothing algorithm is: Among them, the represents the optimal smoothing result derived from the forward positioning result, P k-1|k represents the current moment robust Kalman filter covariance matrix derived from the optimal covariance matrix of the previous moment, represents the optimal smoothing result at the previous moment, the P k-1 represents the optimal covariance matrix at the previous moment, the φ k T represents the state matrix, Represents the covariance matrix predicted values.

7. A terminal, characterized in that: include: A memory, a processor, and an anti-interference position estimation program under multi-source information constraints stored in the memory and executable on the processor, wherein when the anti-interference position estimation program under multi-source information constraints is executed by the processor, the steps of the anti-interference position estimation method under multi-source information constraints as described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program can be executed to implement the steps of the anti-interference position estimation method under multi-source information constraints as claimed in any one of claims 1 to 5.

Images