A terminal positioning method and related apparatus

Abstract

The application discloses a terminal positioning method and related device, the method comprises the following steps: determining a state vector and a process noise covariance matrix according to motion state information of a target terminal, that is, high-order motion state information is introduced in the application to position the target terminal, and the accuracy and stability are improved; further determining a measurement vector according to position information of N base stations and the state vector based on a measurement model, reducing the dependence on absolute high-order motion state information, and enhancing the robustness of terminal positioning; then, using a Kalman filtering algorithm, predicting and updating the target terminal according to the state vector, the process noise covariance matrix and the measurement vector, obtaining a target trajectory of the target terminal, realizing effective fusion of motion state information and high-order differential state information (i.e. the measurement vector) of the terminal, and improving the accuracy of terminal positioning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of terminal positioning, and in particular to a terminal positioning method and related device. BACKGROUND

[0002] At present, in order to enable users to implement location-based services such as navigation, location sharing, etc., the positions of all terminals (such as vehicle-mounted intelligent devices, mobile phones, etc.) of the user are generally positioned and monitored.

[0003] In related technologies, a positioning method of TDOA (Time Difference of Arrival), FDOA (Frequency Difference of Arrival) or AOA (Angle of Arrival) is generally used to position terminal devices. However, the positioning method provided by related technologies still has the problem of low positioning accuracy. SUMMARY

[0004] Based on the above problems, the present application provides a terminal positioning method and related device, aiming to improve the accuracy of terminal positioning.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, the embodiments of the present application provide a terminal positioning method, which comprises:

[0007] Obtaining motion state information of a target terminal; the target terminal is in communication with N base stations, and N is an integer greater than or equal to 2;

[0008] Determining a state vector and a process noise covariance matrix according to the motion state information; the state vector represents state components of the target terminal when moving in a multi-dimensional space, and the process noise covariance matrix represents a change process of the state vector from a previous time to a current time in the case of including driving noise;

[0009] Determining a measurement vector according to the position information of the N base stations and the state vector based on a measurement model; the measurement model represents a mapping relationship between the state vector and the measurement vector, the measurement vector represents a differential motion state between M reference base stations and (N-M) base stations, M is an integer greater than or equal to 1 and less than N, and the measurement model is constructed based on a nonlinear function and measurement noise;

[0010] Using a Kalman filtering algorithm, the target terminal is predicted and updated according to the state vector, the process noise covariance matrix and the measurement vector, to obtain a target trajectory of the target terminal.

[0011] In a possible implementation manner of the first aspect, the target terminal is predicted and updated according to the state vector, the process noise covariance matrix and the measurement vector by using a Kalman filtering algorithm to obtain the target trajectory of the target terminal, including:

[0012] determining the sample point set according to the historical state estimation value and the historical state covariance matrix of the last moment;

[0013] propagating the sample point set through a motion model to determine a predicted state mean value and a predicted state covariance matrix, wherein the predicted state covariance matrix is determined based on the predicted state mean value and the process noise covariance matrix, and the motion model is constructed based on the state vector;

[0014] propagating the sample point set through a measurement model to determine a predicted measurement mean value, an innovation covariance matrix and a state-measurement cross covariance matrix;

[0015] determining a posterior state estimation value and a posterior state covariance matrix according to the predicted state mean value, the predicted state covariance matrix, the predicted measurement mean value, the innovation covariance matrix, the state-measurement cross covariance matrix and the measurement vector;

[0016] obtaining the target trajectory of the target terminal according to the posterior state estimation value and the posterior state covariance matrix.

[0017] In a possible implementation manner of the first aspect, the posterior state estimation value and the posterior state covariance matrix are determined according to the predicted state mean value, the predicted state covariance matrix, the predicted measurement mean value, the innovation covariance matrix, the state-measurement cross covariance matrix and the measurement vector, including:

[0018] determining a Kalman gain according to the innovation covariance matrix and the state-measurement cross covariance matrix;

[0019] determining the posterior state estimation value according to the predicted state mean value, the measurement vector, the predicted measurement mean value and the Kalman gain;

[0020] determining the posterior state covariance matrix according to the predicted state covariance matrix, the innovation covariance matrix and the Kalman gain.

[0021] In a possible implementation manner of the first aspect, the sample point set is determined according to the historical state estimation value and the historical state covariance matrix of the last moment, including:

[0022] decomposing the historical state covariance matrix to obtain a first numerical value;

[0023] determining the sample point set according to the historical state estimation value and the first numerical value.

[0024] In a possible implementation manner of the first aspect, the sample point set is propagated through the motion model to determine the predicted state mean and the predicted state covariance matrix, including:

[0025] The first weight and the second weight are obtained;

[0026] The sample point set is propagated through the motion model to obtain a propagated sample point set;

[0027] The predicted state mean is determined according to the propagated sample point set and the first weight;

[0028] The predicted state covariance matrix is determined according to the second weight, the propagated sample point set, the predicted state mean and the process noise covariance matrix.

[0029] In a possible implementation manner of the first aspect, the first weight and the second weight are obtained, including:

[0030] The scaling parameter of the Kalman filter algorithm and the total number of variables estimated by the Kalman filter algorithm are obtained;

[0031] The compound scaling parameter is calculated according to the scaling parameter and the total number;

[0032] The first weight and the second weight are calculated according to the compound scaling parameter.

[0033] In a possible implementation manner of the first aspect, the Kalman filter algorithm is an unscented Kalman filter algorithm.

[0034] In a second aspect, an embodiment of the present application provides a terminal positioning device, which comprises:

[0035] An obtaining module is configured to obtain motion state information of a target terminal; the target terminal communicates with N base stations, and N is an integer greater than or equal to 2;

[0036] A determining module is configured to determine a state vector and a process noise covariance matrix according to the motion state information; the state vector represents state components of the target terminal when moving in a multi-dimensional space, and the process noise covariance matrix represents a change process of the state vector from a previous time to a current time in the case of including driving noise;

[0037] A measurement vector determining module is configured to determine a measurement vector according to the position information of the N base stations and the state vector based on a measurement model; the measurement model represents a mapping relationship between the state vector and the measurement vector, and the measurement vector represents a differential motion state between M reference base stations and (N-M) base stations, M is an integer greater than or equal to 1 and less than N; the measurement model is constructed based on a nonlinear function and measurement noise;

[0038] The terminal positioning module is configured to utilize a Kalman filtering algorithm to predict and update the target terminal according to the state vector, the process noise covariance matrix, and the measurement vector, so as to obtain a target trajectory of the target terminal.

[0039] In a third aspect, an embodiment of the present application provides a control device, including a processor and a memory, the memory being configured to store programs, instructions or codes, and the processor being configured to execute the programs, instructions or codes in the memory to complete the terminal positioning method in the first aspect.

[0040] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is loaded by a processor to execute the terminal positioning method in the first aspect.

[0041] Beneficial effects:

[0042] The terminal positioning method provided by the embodiment of the present application determines the state vector and the process noise covariance matrix according to the motion state information of the target terminal, that is, high-order motion state information is introduced in the present application to position the target terminal, thereby improving the accuracy and stability; further, the measurement vector is determined according to the position information of the N base stations and the state vector based on the measurement model, thereby reducing the dependence on absolute high-order motion state information and enhancing the robustness of terminal positioning; then, the Kalman filtering algorithm is utilized to predict and update the target terminal according to the state vector, the process noise covariance matrix and the measurement vector, so as to obtain the target trajectory of the target terminal, thereby realizing effective fusion of the motion state information and the high-order differential state information (i.e. the measurement vector) of the terminal and improving the accuracy of terminal positioning.

[0043] The target terminal communicates with N base stations, and N is an integer greater than or equal to 2; the state vector represents the state component of the target terminal when moving in a multi-dimensional space, and the process noise covariance matrix represents the change process of the state vector from the last time to the current time in the case of including driving noise; the measurement model represents the mapping relationship between the state vector and the measurement vector, and the measurement vector represents the differential motion state between the M reference base stations and the (N-M) base stations, and the measurement model is constructed based on a nonlinear function and measurement noise. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0045] Figure 1A flowchart of a terminal positioning method provided by an embodiment of the present application is shown in FIG. 1.

[0046] Figure 2 A flowchart of a target trajectory determination process provided by an embodiment of the present application is shown in FIG. 2.

[0047] Figure 3a A schematic diagram of a target trajectory in a one-dimensional scene provided by an embodiment of the present application is shown in FIG. 3.

[0048] Figure 3b A schematic diagram of a target trajectory in a two-dimensional scene provided by an embodiment of the present application is shown in FIG. 4.

[0049] Figure 3c A schematic diagram of a target trajectory in a three-dimensional scene provided by an embodiment of the present application is shown in FIG. 5.

[0050] Figure 4 A structural schematic diagram of terminal positioning provided by an embodiment of the present application is shown in FIG. 6.

[0051] Figure 5 A structural schematic diagram of a control device provided by an embodiment of the present application is shown in FIG. 7. DETAILED DESCRIPTION

[0052] As an example, terminal positioning takes uplink wireless positioning as an example. A terminal device to be positioned transmits a radio signal. A base station receives the radio signal and measures positioning parameters such as angle of arrival, angle of departure, and time of arrival of the radio signal. Then, the positioning parameters are measured to obtain a positioning result, so as to realize position estimation and monitoring of the terminal device.

[0053] In related technologies, TDOA, FDOA, or AOA can be determined according to the positioning parameters, and then the terminal device is positioned by using the TDOA, FDOA, or AOA. However, when the number of base stations is small or part of the observation errors is large, the accuracy and reliability of positioning by using TDOA are low. When the terminal is positioned by using FDOA or AOA, the adaptation to the non-cellular positioning scene is not considered, and higher-order motion parameters are not considered, which leads to inadaptation of the positioning scene and low accuracy and reliability of terminal positioning.

[0054] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor are within the scope of protection of the present application.

[0055] First, the terms that may appear in each of the following embodiments are explained.

[0056] TOA (Time of Arrival) refers to the time taken by a radio signal to reach a base station from a terminal device.

[0057] TDOA (Time Difference of Arrival) refers to the difference in time taken by a terminal device to reach multiple base stations. Unlike TOA, TDOA does not require strict time synchronization between the terminal device and the base station, only synchronization between the base stations.

[0058] DOA (Direction of Arrival) refers to the direction angle of a radio signal when it reaches a base station.

[0059] DDOA (Doppler Difference of Arrival) refers to the difference in Doppler frequency observed by two base stations. Doppler frequency is the difference between the frequency of a radio signal received by a base station and the frequency at which the radio signal was transmitted, reflecting the relative motion between the transmitting and receiving stations.

[0060] FDOA (Frequency Difference of Arrival) refers to the difference in frequency of a radio signal observed by two base stations. FDOA focuses more on describing the difference in frequency of a radio signal, rather than the frequency change caused by the Doppler effect.

[0061] AOA (Angle of Arrival) refers to the direction angle of a radio signal when it reaches a base station, similar to the concept of DOA. AOA may focus more on describing the situation where a radio signal arrives at a base station from a specific direction.

[0062] A base station, also known as a Transmit Receive Point (TRP) or Access Point (AP), generally includes an antenna and a radio frequency processing unit, distributed in space, for receiving and transmitting wireless signals. The base station is responsible for distributing in space, receiving and transmitting wireless signals, thereby providing wireless access services for users.

[0063] A non-cellular network means a network that differs from a cellular mobile communication network in terms of cell division concept, with all access points densely deployed throughout the region to collectively serve users.

[0064] Higher-order motion state information (or motion state information) refers to the derivative information of the moving distance of a terminal device during motion, with the first derivative being velocity, the second derivative being acceleration, and the third derivative being jerk. When the terminal device is doing complex motion, the higher-order derivative is not zero.

[0065] High-order difference motion state information (or referred to as difference motion state information) refers to distance difference, speed difference, acceleration difference, jerk difference, etc.

[0066] Kalman filtering is a recursive linear optimal filtering algorithm used to estimate the state of a dynamic system from noisy measurement data. It is based on the state-space model of the system and alternates between prediction and update steps to gradually approach the true state of the system.

[0067] Unscented Kalman Filter (UKF) is a nonlinear extension of Kalman filter for state estimation problems of nonlinear systems. Unscented Kalman Filter algorithm approximates the probability distribution of nonlinear functions through Unscented Transform (UT).

[0068] Referring to Figure 1 , the figure is a flowchart of a terminal positioning method provided by an embodiment of the present application.

[0069] In combination Figure 1 with

[0070] S11: Obtain motion state information of a target terminal.

[0071] The target terminal means a terminal device to be positioned, which can include but is not limited to a mobile phone, a smart watch, a tablet, a computer, or other mobile and communication-enabled devices. The target terminal communicates with N base stations, where N is an integer greater than or equal to 2.

[0072] In the embodiment of the present application, the motion state information means high-order motion state information of the target terminal, such as speed, acceleration, jerk, etc., which is not limited here.

[0073] S12: Determine a state vector and a process noise covariance matrix according to the motion state information.

[0074] The state vector represents the state components of the target terminal when moving in a multi-dimensional space. That is, the state vector describes the motion state vector (component) of the target terminal in a multi-dimensional space at a certain time.

[0075] As an example, the motion of the target terminal in one dimension is taken as an example, where the dimension refers to the spatial dimension, which is the physical space in which the terminal device moves. One dimension represents any one of the horizontal direction (x), the vertical direction (y), or the depth direction (y) of the target terminal.

[0076] For single-dimensional motion, the state vector contains the first State vector of the derivative (dimension Here, the dimension m refers to the order of the motion state, which refers to how many levels (i.e., derivatives) of motion state information are contained.

[0077] Equation (1)

[0078] wherein, represents a state vector in one dimension, is a position, is a velocity (first derivative), is an acceleration (second derivative), is the derivative of the order of .

[0079] Further, one dimension is extended to multiple dimensions, that is, it is assumed that the terminal device moves in an N-dimensional space, and the state vector of the target terminal can be obtained by stacking the state components of each dimension, as shown in the following equation (2):

[0080] Equation (2)

[0081] wherein, represents a state vector in an N-dimensional space, is a single-dimensional state vector of the Nth dimension, the total dimension of .

[0082] The process noise covariance matrix represents the change of the state vector from the last time to the current time in the case of including driving noise. That is, the process noise covariance matrix is a matrix that describes the influence of random changes in the state evolution caused by unmodeled parameter uncertainty or external disturbances in the change process of the state vector from the last time to the current time. In one possible implementation, the process noise covariance matrix can be obtained based on the state transition matrix.

[0083] As an example, taking the motion of the target terminal in one dimension as an example, when the state vector in one dimension is determined, the state transition matrix describes how the state vector evolves from to in the case of no process noise. For the time interval , the element of the state transition matrix (the row, the column) represents the change of the state vector in the corresponding dimension from the last time to the current time.The Taylor expansion of the state transition matrix

[0084] Equation (3)

[0085] where is the state transition matrix, is the factorial of , and is the time step.

[0086] Further, the process noise covariance matrix describes the state uncertainty caused by the random variation of the th derivative. Then, define the input noise influence vector whose elements (index starts from 1) represent the influence of the unit driving noise on the state component :

[0087] Equation (4)

[0088] where is the influence vector of the unit driving noise on the state component , is the time step, and m represents the order of the motion state. The process noise covariance matrix is

[0089]

[0090] Equation (5) where

[0091] is the time step, represents the driving noise, represents the input noise influence vector, represents the variance of the driving noise. Further, extend one dimension to multi-dimensions, i.e., assume that the terminal device moves in

[0092] -dimensional space, and assume that the motion and process noise between each dimension are independent. Then, determine the state vector in the N-dimensional space, and the state transition matrix is a block-diagonal matrix:

[0093] Equation (6) Then, the process noise covariance matrix

[0094] is also a block-diagonal matrix:

[0095] Equation (7) ​

[0096] where blkdiag is a function or tool for constructing or operating a block diagonal matrix, and the core function thereof is to arrange an input matrix along the main diagonal to form a square matrix in which the diagonal blocks are the input matrix and the off-diagonal blocks are all zeros.

[0097] It should be understood that in the embodiments of the present application, the high-order motion state information is introduced for terminal positioning, which can effectively utilize the high-order motion state information and is beneficial to improving the accuracy and reliability of terminal positioning.

[0098] S13: determining a measurement vector based on a measurement model and the position information of the N base stations and the state vector.

[0099] The measurement vector represents the differential motion state between the M reference base stations and the (N-M) base stations, such as TDOA, DDOA (FDOA), and other high-order differential motion state information. M is an integer greater than or equal to 1 and less than N.

[0100] The measurement model represents the mapping relationship between the state vector and the measurement vector, and is constructed based on a nonlinear function and measurement noise. That is, the measurement model describes how to obtain the measurement value from the state vector The measurement model at time k can be the following formula:

[0101] ; formula (8)

[0102] wherein, represents the measurement vector, represents the measurement function, represents the state vector at time k, and the measurement noise is assumed to be zero-mean Gaussian noise, and the covariance of the measurement noise is , is the measurement noise.

[0103] It should be understood that in the embodiments of the present application, the measurement vector is determined based on the position information of the N base stations and the state vector based on the measurement model, which reduces the dependence on absolute high-order motion state information and enhances the robustness of terminal positioning.

[0104] S14: predicting and updating the target terminal based on the state vector, the process noise covariance matrix, and the measurement vector using a Kalman filtering algorithm to obtain the target trajectory of the target terminal.

[0105] ​In a possible implementation, the Kalman filtering algorithm is an unscented Kalman filtering algorithm. It should be understood that the unscented Kalman filtering algorithm directly approximates the probability distribution of a nonlinear function through unscented transformation, reduces the probability of truncation error caused by linearization, and thus can provide higher estimation accuracy in terminal positioning; meanwhile, the unscented Kalman filtering algorithm has better stability in terminal positioning and is less likely to diverge because it does not need linearization processing.

[0106] In the embodiments of the present application, the Kalman filtering algorithm is used to predict and update the target terminal according to the state vector, the process noise covariance matrix, and the measurement vector, to obtain the target trajectory of the target terminal, effectively fuse the motion state information of the terminal and the high-order difference state information (i.e., the measurement vector), and improve the accuracy of terminal positioning.

[0107] Based on the terminal positioning method provided in the above embodiments, in a possible implementation, the method further includes Figure 2 As shown in FIG. 14, step S14 can include:

[0108] S21: determining a set of sample points according to the historical state estimation value and the historical state covariance matrix of the previous moment.

[0109] The state estimation value represents the optimal state estimation at a certain moment, i.e., the position, speed, etc. of the target terminal at a certain moment. The state covariance matrix is a symmetric positive definite matrix, which describes the uncertainty of the state estimation value.

[0110] It should be noted that if the current moment is the initial moment, the initial state estimation value and the initial state covariance matrix at the initialization time can be used as the historical state estimation value and the historical state covariance matrix. That is, in the initialization stage, the initial state estimation value and the initial state covariance matrix can be defined in advance, and when the current moment is the initial moment, the initial state estimation value and the initial state covariance matrix are used as the historical state estimation value and the historical state covariance matrix of the previous moment to determine the set of sample points.

[0111] The set of sample points (Sigma point set) is a concept in the unscented Kalman filtering (UKF) algorithm, which is used to approximate the probability distribution of a nonlinear function. The Sigma point set is a set of specifically selected points used to capture the main features of the state distribution.

[0112] In a possible implementation, step S21 can include:

[0113] A1: decomposing the historical state covariance matrix to obtain a first numerical value.

[0114] As an example, suppose the historical state estimate at the previous time step is... The covariance matrix of the historical states is Then, the historical state covariance matrix calculated by Cholesky decomposition is: The square root:

[0115] ;Formula (9)

[0116] in, Indicates the first value. Represents the Cholesky decomposition function. .

[0117] A2: Determine the sample point set based on the historical state estimate and the first value.

[0118] As an example, suppose the first value is Then it can generate Sigma points :

[0119] ;Formula (10)

[0120] in, Representation matrix The List, Indicates the relationship with the i-th Sigma point Relative to historical state estimates Symmetrical sample points.

[0121] S22: Propagate the sample point set through the motion model to determine the predicted state mean and predicted state covariance matrix.

[0122] A motion model is a mathematical expression that describes the change of the motion state of a target terminal over time. The motion model is constructed based on the state vector.

[0123] The predicted state mean is one of the outputs of the prediction step in Kalman filtering. It represents the mean of the state vector predicted for the next time step based on the current state and the motion model.

[0124] The predicted state covariance matrix describes the uncertainty of the predicted state mean. It is a symmetric, non-negative definite matrix whose elements represent the covariance between the components of the state vector.

[0125] The predicted state covariance matrix is ​​determined based on the predicted state mean and the process noise covariance matrix.

[0126] In one possible implementation, step S22 may include:

[0127] B1: obtaining a first weight and a second weight.

[0128] The first weight is used to determine the predicted state mean, and the second weight is used to determine the predicted state covariance matrix. In a possible implementation, the first weight and the second weight can be obtained in the following manner:

[0129] C1: obtaining a scaling parameter of the Kalman filtering algorithm and a total number of variables estimated by the Kalman filtering algorithm.

[0130] The scaling parameter can include . α is a small positive number ( ) used to control the degree of diffusion of the Sigma points. (optimally for Gaussian distribution). κ is an auxiliary parameter (usually 0 or 3−n) used to further adjust the distribution of the Sigma points.

[0131] The total number of variables estimated by the Kalman filtering algorithm means the total number of variables (such as n) that need to be monitored and estimated by the Kalman filter.

[0132] C2: calculating a composite scaling parameter according to the scaling parameter and the total number.

[0133] The composite scaling parameter is a parameter in the UKF algorithm, used to adjust the distribution of the Sigma points, so as to control the accuracy and stability of the nonlinear transformation. The calculation formula is usually:

[0134] Equation (11)

[0135] Wherein, λ represents the composite scaling parameter, represents the scaling parameter, and n represents the total number of variables estimated by the Kalman filtering algorithm.

[0136] C3: calculating the first weight and the second weight according to the composite scaling parameter.

[0137] As an example, the first weight and the second weight can be calculated in the following manner:

[0138] Equation (12)

[0139] Wherein, , represents the initial weight, represents the second weight, represents the first weight, λ represents the composite scaling parameter, represents the scaling parameter, and n represents the total number of variables estimated by the Kalman filtering algorithm.

[0140] B2: Propagate the sample point set through the motion model to obtain a propagated sample point set.

[0141] As an example, a propagated sample point in the propagated sample point set can be:

[0142] Equation (13)

[0143] wherein, denotes a propagated sample point, denotes a sample point, is a state transition function, .

[0144] B3: Determine a predicted state mean according to the propagated sample point set and a first weight.

[0145] As an example, the predicted state mean can be:

[0146] Equation (14)

[0147] wherein, denotes a predicted state mean, denotes a first weight, denotes a propagated sample point.

[0148] B4: Determine a predicted state covariance matrix according to a second weight, the propagated sample point set, the predicted state mean and a process noise covariance matrix.

[0149] As an example, the predicted state covariance matrix can be:

[0150] ;

[0151] Equation (15)

[0152] wherein, denotes a predicted state covariance matrix, denotes a propagated sample point, denotes a predicted state mean, denotes a second weight.

[0153] S23: Propagate the sample point set through the measurement model to determine a predicted measurement mean, an innovation covariance matrix and a state-measurement cross-covariance matrix.

[0154] The predicted measurement mean is the expected mean of the measurement at the next time step, calculated based on the predicted state and the measurement model. The innovation covariance matrix describes the statistical properties of the innovation, i.e., the covariance between the predicted and actual measurements. The state-measurement cross-covariance matrix describes the correlation between the state vector and the measurements. It is a matrix whose elements represent the covariance between each component of the state vector and the measurement.

[0155] As an example, step S23 may include:

[0156] Propagating the sample point set through the measurement model can be represented as:

[0157] ;Formula (16)

[0158] in, Indicates the measurement of propagation sample points. The measurement function represents the measurement model. , This represents the propagation sample point.

[0159] The predicted measurement mean can be expressed as:

[0160] ;Formula (17)

[0161] in, Indicates the predicted measurement mean. Indicates the first weight. This indicates the measurement propagation sample point.

[0162] The new covariance matrix can be represented as:

[0163] ;

[0164] Formula (18)

[0165] in, Represents the new information covariance matrix. Indicates the second weight. Indicates the measurement of propagation sample points. R represents the predicted measurement mean, and R represents the covariance matrix of the measurement noise.

[0166] The state-measurement cross-covariance matrix can be represented as:

[0167] ;

[0168] Formula (19)

[0169] in, Represents the state-measurement cross-covariance matrix. Indicates the measurement of propagation sample points. denotes a propagated sample point, denotes a predicted measurement mean, denotes a predicted state mean, denotes a second weight.

[0170] S24: determining a posterior state estimate and a posterior state covariance matrix in dependence on the predicted state mean, the predicted state covariance matrix, the predicted measurement mean, the innovation covariance matrix, the state-measurement cross-covariance matrix and the measurement vector.

[0171] In one possible implementation, step S24 can comprise:

[0172] D1 : determining a Kalman gain in dependence on the innovation covariance matrix and the state-measurement cross-covariance matrix.

[0173] As one example, the Kalman gain can be expressed as:

[0174] ; equation (20)

[0175] wherein, denotes the Kalman gain, denotes the state-measurement cross-covariance matrix, denotes the innovation covariance matrix.

[0176] D2: determining the posterior state estimate in dependence on the predicted state mean, the measurement vector, the predicted measurement mean and the Kalman gain.

[0177] As one example, the updated state estimate can be expressed as:

[0178] ; equation (21)

[0179] wherein, denotes the updated state estimate, i.e. denotes the posterior state estimate at time instant denotes the predicted state mean, denotes the Kalman gain, denotes the predicted measurement mean, denotes the actual measurement value at time instant k.

[0180] D3: determining the posterior state covariance matrix in dependence on the predicted state covariance matrix, the innovation covariance matrix and the Kalman gain.

[0181] As one example, the updated covariance estimate matrix can be expressed as:

[0182] ; equation (22) ​

[0183] wherein, represents updating the covariance estimation matrix, i.e., represents updating the posterior state covariance matrix at the time instant k, represents the predicted state covariance matrix, represents the Kalman gain, represents the innovation covariance matrix.

[0184] It should be noted that the posterior state estimation value and the posterior state covariance matrix will be taken as the historical state estimation value for the next iteration (i.e., the next time instant) as and the historical state covariance matrix as .

[0185] S25: obtaining the target trajectory of the target terminal according to the posterior state estimation value and the posterior state covariance matrix.

[0186] The target trajectory means the motion trajectory of the target terminal. It should be understood that since the posterior state estimation value represents the optimal estimation of the motion state (such as position, velocity, acceleration, etc.) of the target terminal at the time instant k. The posterior state covariance matrix describes the uncertainty of the posterior state estimation value, and the diagonal elements are the variances of the state components, and the non-diagonal elements reflect the correlation between the components. Therefore, the target trajectory of the target terminal can be obtained according to the posterior state estimation value and the posterior state covariance matrix, so as to realize the continuous positioning and monitoring of the target terminal.

[0187] As an example, the motion trajectory of the target terminal can be determined in a time sequence splicing manner, for example: starting from the initial time instant k=0, the posterior state is recorded; then, at each time instant k, the state at the next time instant is predicted , and is updated to the posterior state in combination with the new measurement value; finally, the posterior state at each time instant is stored in time sequence to form a trajectory sequence.

[0188] It should be noted that the embodiments of the present application are also applicable to one-dimensional terminal positioning scenarios. When the spatial dimension is reduced from two dimensions to one dimension, only the dimensions of the state vector and the process noise covariance matrix need to be simplified (as shown in formula (1) and formula (4)), and the remaining steps can be implemented according to the steps S11-S14 in the above embodiments, i.e., the positioning of the target terminal can be realized, as shown in formula (5), and the repeated description is not made herein. It can be seen from formula (5) that the target trajectory obtained by the terminal positioning provided by the embodiments of the present application is more consistent with the actual trajectory of the target terminal, and the accuracy and reliability of the terminal positioning are improved to a certain extent. Figure 3a Figure 3a

[0189] ​​​Based on the terminal positioning method provided in the above embodiments, as an example, assuming that the motion state information of the target terminal includes the distance difference and the speed difference, the flow of terminal positioning can be:

[0190] Step 1: Determine the state vector, motion model, measurement model and measurement vector.

[0191] Assuming the state vector At time step Generally containing the position and speed of the target terminal, the state dimension , the state vector can be expressed as:

[0192] Equation (23)

[0193] Wherein, represents the state vector, represents the position, represents the speed.

[0194] Assuming that the motion model is an evolution based on a nonlinear function with additive process noise , the motion model can be expressed as:

[0195] Equation (24)

[0196] Wherein, the function can be selected as a linear constant velocity (CV) model, which can be expressed as:

[0197] Equation (25)

[0198] Wherein, is the state transition matrix of the CV model, expressed as:

[0199] Equation (26)

[0200] Wherein, is the time step, the process noise is zero-mean Gaussian noise, and the covariance matrix of the process noise is , that is, .

[0201] Wherein, the measurement vector is composed of TDOA and FDOA measurements of base stations, assuming that any one base station is taken as a reference base station, and the measurement dimension is , the measurement vector can be expressed as:

[0202] Equation (27)

[0203] wherein, denotes the measurement vector, is the TDOA between the base station and the reference base station, is the FDOA between the base station and the reference base station, both at time .

[0204] The measurement model represents the mapping between the state vector and the measurement vector, and is given by a non-linear function with additive measurement noise , which can be expressed as:

[0205] Equation (28)

[0206] The function computes the expected TDOA and FDOA values from the target state (position and velocity) and the known base station positions. Due to the involved calculations of distances and relative velocity projections, the function is non-linear. The measurement noise is assumed to be zero-mean Gaussian noise, and the covariance of the measurement noise is , i.e. .

[0207] Step 2: Utilize the Kalman filter algorithm to predict and update the target terminal according to the state vector, the process noise covariance matrix, and the measurement vector, to obtain the target trajectory of the target terminal.

[0208] It should be noted that the step 2 provided by the embodiments of the present application is similar to the step of the above-mentioned step S14, and the difference lies in the definition (or value) of the state vector, the motion model, the measurement model, and the measurement vector, and therefore the related explanations and descriptions can be referred to the explanations and descriptions of the above-mentioned step S14, which will not be repeated here.

[0209] It should be understood that, based on the above-mentioned step 1 and step 2, the positioning of the target terminal can be realized under the condition that the motion state information of the target terminal includes the distance difference and the velocity difference (two-dimensional scene), as shown in Figure 3b . It can be seen from Figure 3b that, in the two-dimensional scene, the target trajectory obtained by the terminal positioning provided by the embodiments of the present application is more consistent with the actual trajectory of the target terminal, which improves the accuracy and reliability of the terminal positioning to a certain extent.

[0210] Based on the terminal positioning method provided in the above embodiments, as an example, assuming that the motion state information of the target terminal includes distance difference, speed difference and acceleration difference, the flow of terminal positioning can be:

[0211] Step one: determine the state vector, motion model, measurement model and measurement vector.

[0212] Assuming the state vector At time step Generally containing the position, speed and acceleration of the target, the state dimension , the state vector can be expressed as:

[0213] ; equation (29)

[0214] Wherein, is the position, is the speed, is the acceleration.

[0215] Assuming that the motion model is an evolution based on a nonlinear function , accompanied by additive process noise , the motion model can be expressed as:

[0216] ; equation (30)

[0217] Wherein, the constant acceleration (CA) model, the state transition is linear, is expressed as:

[0218] ; equation (31)

[0219] Wherein, is the state transition matrix of the CA model, expressed as:

[0220] ; equation (32)

[0221] Wherein, is the time step. The process noise is assumed to be zero-mean Gaussian noise, and the covariance of the process noise is , that is .

[0222] The measurement vector is composed of TDOA, DDOA and DORA (Difference of Radial Acceleration) measurements of base stations, usually taking any one of the base stations as a reference base station, and the measurement dimension is , the measurement vector can be expressed as:

[0223] ; equation (33)

[0224] where, is the TDOA between the target terminal and the reference base station, is the FDOA between the target terminal and the reference base station, is the projected acceleration difference between the target terminal and the reference base station, all at time measurements.

[0225] The measurement model is given by a nonlinear function with additive measurement noise , and can be expressed as:

[0226] ; equation (34)

[0227] where the function computes the expected TDOA, DDOA, and DORA values from the target state (position, velocity, and acceleration) and the known base station locations.

[0228] The TDOA is computed as:

[0229] ; equation (35)

[0230] where, is the distance from the target terminal to the base station, is the speed of light.

[0231] The DDOA is computed as:

[0232] ; equation (36)

[0233] where, is the carrier frequency, is the unit vector pointing from the target terminal to the base station.

[0234] The DORA is computed as:

[0235] ; equation (37)

[0236] where, is a proportionality constant, is the acceleration vector of the target terminal.

[0237] ​​​​​The function is nonlinear due to the calculations involving range, relative velocity projection, and relative acceleration projection. The measurement noise is assumed to be zero-mean Gaussian noise with covariance , i.e.

[0238] Step two: using the Kalman filtering algorithm, according to the state vector, the process noise covariance matrix, and the measurement vector, the target terminal is predicted and updated to obtain the target trajectory of the target terminal.

[0239] It should be noted that the step two provided by the embodiments of the present application is similar to the step of the above-mentioned step S14, and the difference lies in the definition (or value) of the state vector, the motion model, the measurement model, and the measurement vector, and therefore the related explanation and description can be referred to the explanation and description of the above-mentioned step S14, which will not be repeated here.

[0240] It should be understood that based on the above-mentioned step one and step two, the positioning of the target terminal can be realized under the condition that the motion state information of the target terminal includes the range difference and the velocity difference (three-dimensional scene), as shown in Figure 3c It can be seen from Figure 3c that in the three-dimensional scene, the target trajectory obtained by the terminal positioning provided by the embodiments of the present application is more consistent with the actual trajectory of the target terminal, which improves the accuracy and reliability of the terminal positioning to a certain extent.

[0241] Based on the terminal positioning method provided by the above-mentioned embodiments, the embodiments of the present application further provide a structural diagram of a terminal positioning device.

[0242] As shown in Figure 4 , the terminal positioning device 40 provided by the embodiments of the present application can include:

[0243] The acquisition module 41 is configured to acquire the motion state information of the target terminal; the target terminal is in communication with N base stations, and N is an integer greater than or equal to 2;

[0244] The determination module 42 is configured to determine the state vector and the process noise covariance matrix according to the motion state information; the state vector represents the state component of the target terminal when moving in a multi-dimensional space, and the process noise covariance matrix represents the change process of the state vector from the last time to the current time under the condition of including driving noise;

[0245] The measurement vector determination module 43 is configured to determine a measurement vector based on a measurement model and according to the position information of the N base stations and the state vector, the measurement model representing a mapping relationship between the state vector and the measurement vector, the measurement vector representing a differential motion state between the M reference base stations and the N-M base stations, M being an integer greater than or equal to 1 and less than N, and the measurement model being constructed based on a nonlinear function and measurement noise;

[0246] The terminal positioning module 44 is configured to predict and update the target terminal according to the state vector, the process noise covariance matrix, and the measurement vector by using a Kalman filtering algorithm, to obtain a target trajectory of the target terminal.

[0247] In a possible implementation, the terminal positioning module 44 includes:

[0248] The sample point set determination unit is configured to determine a sample point set according to a historical state estimation value and a historical state covariance matrix at a previous moment;

[0249] The first determination unit is configured to propagate the sample point set through a motion model to determine a predicted state mean value and a predicted state covariance matrix, wherein the predicted state covariance matrix is determined based on the predicted state mean value and the process noise covariance matrix, and the motion model is constructed based on the state vector;

[0250] The second determination unit is configured to propagate the sample point set through a measurement model to determine a predicted measurement mean value, an innovation covariance matrix, and a state-measurement cross covariance matrix;

[0251] The third determination unit is configured to determine a posterior state estimation value and a posterior state covariance matrix according to the predicted state mean value, the predicted state covariance matrix, the predicted measurement mean value, the innovation covariance matrix, the state-measurement cross covariance matrix, and the measurement vector;

[0252] The terminal positioning unit is configured to obtain a target trajectory of the target terminal according to the posterior state estimation value and the posterior state covariance matrix.

[0253] In a possible implementation, the third determination unit is configured to:

[0254] determine a Kalman gain according to the innovation covariance matrix and the state-measurement cross covariance matrix;

[0255] determine the posterior state estimation value according to the predicted state mean value, the measurement vector, the predicted measurement mean value, and the Kalman gain;

[0256] determine the posterior state covariance matrix according to the predicted state covariance matrix, the innovation covariance matrix, and the Kalman gain.

[0257] In a possible implementation, the sample point set determination unit is configured to:

[0258] decomposing the historical state covariance matrix to obtain a first value;

[0259] determining the sample point set according to the historical state estimate value and the first value.

[0260] In a possible implementation, the first determining unit is specifically configured to:

[0261] obtain the first weight and the second weight;

[0262] propagate the sample point set through the motion model to obtain a propagated sample point set;

[0263] determine a predicted state mean value according to the propagated sample point set and the first weight;

[0264] determine a predicted state covariance matrix according to the second weight, the propagated sample point set, the predicted state mean value, and the process noise covariance matrix.

[0265] In a possible implementation, the first determining unit is specifically configured to:

[0266] obtain a scaling parameter of the Kalman filtering algorithm and a total number of variables estimated by the Kalman filtering algorithm;

[0267] calculate a composite scaling parameter according to the scaling parameter and the total number;

[0268] calculate the first weight and the second weight according to the composite scaling parameter.

[0269] It should be noted that the terminal positioning apparatus provided by the embodiments of the present application has the same beneficial effects as the terminal positioning method provided by the above embodiments, and thus will not be described again.

[0270] In a possible implementation, referring to Figure 5 , the figure is a schematic diagram of a control apparatus provided by an embodiment of the present application.

[0271] The control apparatus can include a memory 511 and a processor 512. As Figure 5 shown, the memory can be a random access memory (RAM), a flash memory, a read only memory (ROM), an EPROM memory, a non-volatile read only memory (Electronic Programmable ROM, EPROM), a register, a hard disk, a removable disk, etc.

[0272] The memory 511 can store computer instructions, and when the computer instructions stored in the memory 511 are executed by the processor 512, the processor 512 can be configured to perform the terminal positioning method. The memory 511 can also store data, such as the target trajectory, historical motion state information and the like involved in the above embodiments.

[0273] In the above embodiments, the implementation can be achieved by software, hardware, firmware or any combination thereof, in whole or in part. When implemented by software, the implementation can be achieved in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the whole or part of the flow or function according to the embodiments of the present application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center and the like integrated with one or more available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), or a semiconductor medium (such as a solid state disk (SSD)) and the like.

[0274] The embodiments of the present application also provide a readable storage medium for storing the method provided by the above embodiments. For example, random access memory (RAM), flash memory, read only memory (ROM), EPROM memory, non-volatile read only memory (Electronic Programmable ROM, EPROM), register, hard disk, removable disk or any other form of storage medium in the art.

[0275] The "first", "second" and the like mentioned in the names of the "first", "second" in the embodiments of the present application are only used for name identification, and do not represent the first and second in order.

[0276] It should be noted that the various embodiments described in this specification are intended to be illustrative only and that the scope of the application is defined by the appended claims. Numerous modifications and adaptations thereof will be apparent to those skilled in the art without departing from the spirit and scope of the application. Accordingly, the application is not limited to the specific disclosed embodiments, but rather only by the following claims.

[0277] The above description discloses only typical embodiments. However, specific structural and functional details disclosed herein are not to be interpreted as limiting but merely as a basis for the claims and as a representative basis for teaching one skilled in the art to various embodiments. The description is intended to cover any and all modifications and equivalents. Therefore, the scope of the claims should not be limited by the preferred embodiments and numerous modifications described herein.

Claims

1. A terminal positioning method, characterized in that, The method includes: Obtain motion state information of the target terminal; the target terminal communicates with N base stations, where N is an integer greater than or equal to 2; The state vector and process noise covariance matrix are determined based on the motion state information; the state vector represents the state components of the target terminal when it moves in multi-dimensional space, and the process noise covariance matrix represents the change process of the state vector from the previous moment to the current moment, including driving noise. Based on the measurement model, a measurement vector is determined according to the location information of N base stations and the state vector; the measurement model represents the mapping relationship between the state vector and the measurement vector, and the measurement vector represents the differential motion state between M reference base stations and (NM) base stations, where M is an integer greater than or equal to 1 and less than N; the measurement model is constructed based on nonlinear functions and measurement noise. Using the Kalman filter algorithm, the target terminal is predicted and updated based on the state vector, the process noise covariance matrix, and the measurement vector to obtain the target trajectory of the target terminal.

2. The terminal positioning method according to claim 1, characterized in that, The step of using the Kalman filter algorithm to predict and update the target terminal based on the state vector, the process noise covariance matrix, and the measurement vector to obtain the target trajectory of the target terminal includes: The sample point set is determined based on the historical state estimate and historical state covariance matrix from the previous moment. The sample point set is propagated through a motion model to determine the predicted state mean and the predicted state covariance matrix; wherein, the predicted state covariance matrix is ​​determined based on the predicted state mean and the process noise covariance matrix, and the motion model is constructed based on the state vector; The sample point set is propagated through the measurement model to determine the predicted measurement mean, the information covariance matrix, and the state-measurement cross-covariance matrix; The posterior state estimate and the posterior state covariance matrix are determined based on the predicted state mean, the predicted state covariance matrix, the predicted measurement mean, the information covariance matrix, the state-measurement cross-covariance matrix, and the measurement vector. The target trajectory of the target terminal is obtained based on the posterior state estimate and the posterior state covariance matrix.

3. The terminal positioning method according to claim 2, characterized in that, The step of determining the posterior state estimate and the posterior state covariance matrix based on the predicted state mean, the predicted state covariance matrix, the predicted measurement mean, the innovation covariance matrix, the state-measurement cross-covariance matrix, and the measurement vector includes: The Kalman gain is determined based on the information covariance matrix and the state-measurement cross-covariance matrix. The posterior state estimate is determined based on the predicted state mean, the measurement vector, the predicted measurement mean, and the Kalman gain. The posterior state covariance matrix is ​​determined based on the predicted state covariance matrix, the innovation covariance matrix, and the Kalman gain.

4. The terminal positioning method according to claim 2, characterized in that, The step of determining the sample point set based on the historical state estimate and the historical state covariance matrix of the previous time step includes: The historical state covariance matrix is ​​decomposed to obtain the first value; The sample point set is determined based on the historical state estimate and the first value.

5. The terminal positioning method according to claim 2, characterized in that, The step of propagating the sample point set through a motion model to determine the predicted state mean and predicted state covariance matrix includes: Obtain the first and second weights; The sample point set is propagated through a motion model to obtain a propagated sample point set; The mean of the predicted state is determined based on the propagation sample point set and the first weight; The predicted state covariance matrix is ​​determined based on the second weight, the propagation sample point set, the predicted state mean, and the process noise covariance matrix.

6. The terminal positioning method according to claim 5, characterized in that, The process of obtaining the first weight and the second weight includes: Obtain the scaling parameters of the Kalman filter algorithm and the total number of variables estimated by the Kalman filter algorithm; Calculate the composite scaling parameter based on the scaling parameter and the total quantity; The first weight and the second weight are calculated based on the composite scaling parameters.

7. The terminal positioning method according to any one of claims 1-6, characterized in that, The Kalman filtering algorithm is an unscented Kalman filtering algorithm.

8. A terminal positioning device, characterized in that, The device includes: An acquisition module is used to acquire motion state information of a target terminal; the target terminal communicates with N base stations, where N is an integer greater than or equal to 2. The determination module is used to determine a state vector and a process noise covariance matrix based on the motion state information; the state vector represents the state components of the target terminal when it moves in multi-dimensional space, and the process noise covariance matrix represents the change process of the state vector from the previous moment to the current moment, including driving noise. A measurement vector determination module is used to determine a measurement vector based on a measurement model, according to the location information of N base stations and the state vector; the measurement model represents the mapping relationship between the state vector and the measurement vector, and the measurement vector represents the differential motion state between M reference base stations and (NM) base stations, where M is an integer greater than or equal to 1 and less than N; the measurement model is constructed based on a nonlinear function and measurement noise. The terminal positioning module is used to predict and update the target terminal based on the state vector, the process noise covariance matrix, and the measurement vector using the Kalman filter algorithm, thereby obtaining the target trajectory of the target terminal.

9. A control device, characterized in that, It includes a processor and a memory, the memory being used to store programs, instructions, or code, and the processor being used to execute the programs, instructions, or code in the memory to perform the terminal positioning method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The device contains a computer program that is loaded by a processor to execute the terminal positioning method as described in any one of claims 1-7.

Images