UWB positioning method based on clustering identification and NLOS mitigation

Abstract

The invention relates to the technical field of wireless sensor positioning, in particular to a UWB positioning method based on clustering identification and NLOS mitigation. According to the algorithm, two-stage clustering and Kalman filtering are combined, UWB ranging data under the NLOS and LOS mixed condition are distinguished by using two-stage clustering of KDE kernel density estimation and BSCAN clustering to distinguish ranging data under the NLOS state and discard the ranging data, and improved adaptive Kalman filtering is used for resolving, so that the positioning performance is ensured, and the positioning accuracy is improved. Meanwhile, the UWB positioning precision under the NLOS propagation condition is greatly improved, and the UWB positioning reliability is effectively improved. The reliability is crucial in the fields of military affairs, logistics, health monitoring and the like, and particularly when the system needs to maintain high-precision positioning under different conditions (such as different physical environments or mobile scenes), the application and popularization of the UWB disclosed by the invention play a positive role.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless sensor positioning, and in particular to a UWB positioning method based on clustering recognition and NLOS mitigation. Background Art

[0002] With the rapid development of wireless sensor network technology, wireless sensor positioning technology has become a research hotspot, possessing significant significance and promising applications across various fields, including commerce, industry, and the military. In recent years, UWB has experienced rapid development. Its advantages, such as high transmission rate, low power consumption, strong anti-interference performance, robust multipath resolution, and strong penetration, have made it an excellent choice for wireless sensor positioning. However, in complex environments (such as indoors and underground tunnels), dense obstacles can cause wireless sensors to randomly switch between Line-of-Sight (LOS) and Non-Low-Sight (NLOS) propagation. In NLOS environments, the lack of a direct signal path results in propagation time delays, generating positive deviations that increase UWB ranging errors, significantly impacting positioning accuracy.

[0003] Correctly identifying and mitigating NLOS in mixed LOS and NLOS environments is a current research hotspot. Current research algorithms are generally divided into two categories: identification algorithms, which distinguish between measurements under LOS and NLOS conditions and then use only LOS measurements for positioning. This approach often results in positioning failures in severe NLOS conditions, significantly impacting positioning performance. Mitigation algorithms, on the other hand, feed all data into the mitigation algorithm for processing. These algorithms typically have high computational complexity, and positioning performance deteriorates as the number of NLOS measurements increases. Summary of the Invention

[0004] This invention aims to address the UWB positioning problem under NLOS propagation conditions. It proposes a UWB positioning method based on clustering identification and NLOS mitigation. This algorithm utilizes clustering to effectively identify UWB ranging signals in NLOS conditions and simultaneously employs an improved adaptive Kalman filter to resolve positioning information, effectively mitigating the errors introduced by these conditions. This method creatively combines NLOS identification and mitigation algorithms to significantly improve positioning accuracy.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] In a first aspect, the present invention provides a UWB positioning method based on clustering identification and NLOS mitigation, which includes:

[0007] Obtain ranging data of n base stations in a measurement cycle; divide the n ranging data into groups of m (n>m) and traverse all possible groups, where each ranging data can be divided into different groups at the same time;

[0008] For each set of ranging data, a preliminary coarse positioning point is calculated respectively; for each base station, all preliminary coarse positioning points involved in the ranging data of the base station are divided into one group;

[0009] For all the preliminary positioning coarse measurement points corresponding to each base station, KDE kernel density estimation is performed to obtain the corresponding coarse measurement sites; BSCAN density clustering is performed on all coarse measurement sites to obtain the density center;

[0010] For each base station, calculating the Euclidean distance between the coarse measurement point corresponding to the base station and the density center; the base station corresponding to the coarse measurement point whose Euclidean distance exceeds a specified threshold is considered to be in the NLOS state;

[0011] The nonlinear ranging model and the improved adaptive Kalman filter are used to solve the ranging information after discarding the NLOS state base station to obtain the positioning information of the tag to be found.

[0012] As a preferred embodiment, the preliminary positioning coarse measurement points corresponding to each set of ranging data are respectively calculated based on the least square positioning method.

[0013] As a preferred embodiment, if m=3, the base stations are numbered 1, 2…N, and the numbered set is C∈{1, 2…N}; the formula for calculating the preliminary coarse positioning point corresponding to each set of ranging data includes:

[0014]

[0015] X abc =(H T H) -1 H T b

[0016]

[0017] Among them, X abc Indicates the coordinates of the initial positioning coarse measurement point corresponding to the base station combination numbered a, b, c; a, b, c represent the base station numbers; H is the coefficient matrix; b is the constant term vector; x a ,y a 、x b ,y b 、x c ,y c is the location of the corresponding base station, d a d b dc is the theoretical distance between the target point and each base station, and x, y are the coordinates of the tag to be found.

[0018] As a preferred embodiment, the formula for the Euclidean distance between the coarse measurement point corresponding to the base station and the density center is:

[0019]

[0020] Among them, D i is the distance between the rough measurement point of the base station numbered i and the density center. The base station corresponding to the point whose distance exceeds the specified threshold is determined to be a NLOS state base station.

[0021] As a preferred embodiment, the calculation formula of the threshold value is specified as follows:

[0022]

[0023] Among them, δ represents the preset threshold, w i D i The initial weight, w i norm is the weight after normalization of the initial weight; σ is the adjustment factor.

[0024] As a preferred embodiment, the formula for obtaining positioning information by using a nonlinear ranging model and an improved adaptive Kalman filter to calculate ranging information after discarding NLOS state base stations includes:

[0025]

[0026] P0=I

[0027]

[0028] P k - =FP k-1 F T +Q

[0029]

[0030] d v =[d i ], i=1, 2…n

[0031]

[0032] R k '=R k +(y k y k T )GDOP k

[0033]

[0034] in, is the state estimate updated by the Kalman gain and measurement residual at time k, is the predicted state at time k, K k is the Kalman gain, y k is the measurement residual, is the initial state, P0 is the initial covariance matrix, F is the state transfer matrix, is the predicted state at the previous moment, is the prediction covariance matrix, H k is the observation matrix, R k is the noise covariance matrix, R k ′ is adaptive noise, P k-1 is the covariance matrix of the previous moment, Q is the process noise covariance matrix, d v To remove the ranging data of NLOS state base stations, d i Indicates d v The ranging data of the i-th base station, h(x,y) is the nonlinear measurement model, z i is the distance value from the tag to be determined to the i-th base station; is the measured value based on the predicted state, GDOP k is the GDOP value formed by the predicted position and the base station at the current moment, H v Design matrix formed for LOS anchor and target state estimation.

[0035] In a second aspect, the present invention provides a UWB positioning system based on clustering identification and NLOS mitigation, comprising:

[0036] An acquisition module is used to acquire ranging data of n base stations in a measurement cycle; divide the n ranging data into groups of m (n>m) and traverse all possible groups, and each ranging data can be divided into different groups at the same time;

[0037] The first grouping module is configured to calculate a preliminary coarse positioning point for each group of ranging data; for each base station, all preliminary coarse positioning points calculated using the ranging data of the base station are grouped into one group;

[0038] The density center acquisition module is used to perform KDE kernel density estimation on all preliminary positioning coarse measurement points corresponding to each base station to obtain the coarse measurement site corresponding to each base station; and perform BSCAN density clustering on all coarse measurement sites to obtain the density center;

[0039] an NLOS identification module configured to calculate, for each base station, a Euclidean distance between a coarse measurement point corresponding to the base station and the density center; and to regard a base station corresponding to a coarse measurement point whose Euclidean distance exceeds a specified threshold as being in an NLOS state;

[0040] The positioning information acquisition module is used to use the nonlinear ranging model and the improved adaptive Kalman filter to solve the ranging information after discarding the NLOS state base station to obtain the required tag positioning information.

[0041] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the UWB positioning method based on clustering identification and NLOS mitigation is implemented.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention combines two-stage clustering with Kalman filtering. The two-stage clustering of UWB ranging data under mixed NLOS and LOS conditions uses KDE kernel density estimation and BSCAN clustering to distinguish and discard ranging data under NLOS conditions. The improved adaptive Kalman filter is then used for solution. This not only ensures positioning performance, but also greatly improves the positioning accuracy of UWB under NLOS propagation conditions, effectively enhancing the reliability of UWB positioning. This reliability is crucial in fields such as military, logistics, and health monitoring, especially when the system needs to maintain high-precision positioning under different conditions (e.g., different physical environments or mobile scenarios). Therefore, the present invention plays a positive role in promoting the application of UWB. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of a UWB positioning method based on clustering identification and NLOS mitigation provided by an embodiment of the present invention;

[0045] Figure 2 The flowchart of the clustering-based NLOS identification method for UWB signals is shown in Figure 2.

[0046] Figure 3 This is the flow chart of the improved adaptive Kalman filter algorithm.

[0047] Figure 4 This is a schematic diagram comparing the position solution effects of the solution proposed in this invention and the standard Kalman filter. DETAILED DESCRIPTION

[0048] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0049] Example 1

[0050] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0051] The present invention aims to address the deficiencies of the existing technology and propose a UWB positioning method based on clustering identification and NLOS mitigation to solve the problem of inaccurate positioning caused by the NLOS effect in UWB positioning in complex environments.

[0052] The object of the present invention is to achieve the following technical solution: a UWB positioning method based on clustering identification and NLOS mitigation, the method comprising the following steps:

[0053] Step 1: For a UWB positioning system with n (n>3) positioning base stations, number each base station sequentially. During a measurement cycle during system operation, group the n ranging data into groups of three. All possible groups are permuted and distinguished by base station number combinations. For each group of ranging data, use the least squares method to calculate a point, resulting in several preliminary coarse positioning points.

[0054] Step 2: Regroup the coarse measurement points according to the principle of each base station participating in the solution, and divide them into n groups, which is the number of base stations. For each group, perform KDE (Kernel Density Estimation) on it to obtain the coarse measurement points corresponding to each base station, for a total of n.

[0055] Step 3: Perform BSCAN (Boundary Scan) density clustering on the coarse measurement points corresponding to the n base stations to obtain the density center corresponding to the N base stations at that moment. Calculate the Euclidean distance between each coarse measurement point and the density center. Base stations corresponding to points exceeding the threshold are considered to be in the NLOS state and their measured data are discarded.

[0056] Step 4: Apply the above NLOS identification and data processing to a continuous dynamic process. The processed data is used to perform path calculation using an improved adaptive Kalman filter. The state observation matrix is adjusted based on the available data, and the noise covariance matrix is updated based on the Euclidean distance between the coarse measurement points corresponding to the available base stations and the density center to obtain precise positioning information.

[0057] Figure 2 The present invention provides a flowchart of a UWB positioning method based on cluster identification and NLOS mitigation. Figure 2 As shown, the first aspect of the present invention relates to a clustering-based UWB signal NLOS identification method, which includes steps 1 to 3.

[0058] Furthermore, in step 1, the detailed steps of obtaining the rough determination site include:

[0059] In measurement period k, the coordinates of the positioning tag to be determined are [x(k), y(k)], then the relationship between the tag node and the anchor node (UWB base station) is:

[0060]

[0061] Among them, {(x1,y1),...,(x n ,y n )} are the coordinates of n base stations, d1(k) to d n (k) is the distance between the label node and the anchor node.

[0062] Divide n ranging data into groups of 3 (can be expanded to more than 3), and the total can be divided into Different groups are formed and distinguished by the base station number combination. The base stations are numbered 1, 2…N, and the number set is C∈{1,2…N}.

[0063] For each set of data in a measurement cycle, a point can be calculated using geometric methods. To simplify the formula, the (k) representing the ranging cycle will be omitted below. Among them, for each set of data, the theoretical distance formula from the base station to the tag position to be determined is as follows:

[0064]

[0065] Among them, a, b, c refer to the base station number, x a ,y a 、x b ,y b 、x c ,y c is the location of the corresponding base station, d a d b d c is the theoretical distance between the target point and each base station, and x, y are the coordinates to be determined.

[0066] For each set of data, the corresponding preliminary rough measurement point is solved by geometric method according to the distance formula. The distance formula is transformed into the following form:

[0067]

[0068] Here, i represents the base station number. For the distance difference between the target point and multiple anchor points, we can determine the position of the target point by solving the equations. Taking anchor point a as the reference point, the distance difference with anchor point b can be expressed as:

[0069] da 2 -d b 2 =(xx b ) 2 +(yy b ) 2 -(xx a ) 2 -(yy a ) 2

[0070] Next, a linear equation system is constructed based on the distance difference, and the coefficient matrix H can be expressed as:

[0071]

[0072] The distance difference expansion constructs the constant term vector b in the linear equation system, which can be expressed as:

[0073]

[0074] The constructed linear equations are:

[0075] H.X abc =b

[0076] in, Represents the coordinates of the point to be found corresponding to the base station combination numbered a, b, c.

[0077] X abc =(H T H) -1 H T b

[0078] Solving X can obtain the corresponding preliminary coarse measurement points of this group. The obtained preliminary coarse measurement point set is R abc =(x,y) abc , a,b,c∈C, the number is Each base station corresponds to indivual.

[0079] Furthermore, the step 2 specifically includes the following contents:

[0080] To obtain The initial coarse measurement points are grouped repeatedly according to the principle that each base station participates in the solution. Each base station is allocated the points it participates in the solution. Then, KDE kernel density estimation is performed on each group separately to obtain further coarse measurement points corresponding to each base station.

[0081] The basic formula for kernel density estimation is:

[0082]

[0083] in, is the density estimate at point (x, y). n is the number of sample points. h is the bandwidth (or smoothing parameter) that controls the smoothness of the kernel function. In the two-dimensional case, the same bandwidth is usually used for smoothing in the direction. K(u, v) is a two-dimensional kernel function, defined in two-dimensional space, taking the Gaussian kernel function

[0084] To find the center of maximum density for a two-dimensional kernel density estimate, the goal is to find The largest point is the mode of the density. The gradient ascent method is used to solve it, and the formula is:

[0085]

[0086] in, and Represents the gradient of the density estimation function in the x and y directions at (x,y).

[0087] According to the gradient information, update the values of x and y. The formula is:

[0088]

[0089] Among them, x new and y new Represents the horizontal and vertical coordinates of the updated point, α is the learning rate, which controls the update amplitude of each step and is set to 0.01.

[0090] The final KDE kernel density center is K i =(X i ,Y i ), i=1,2,...,N.

[0091] Furthermore, the step 3 specifically includes the following contents:

[0092] Perform a BSCAN clustering on the obtained N kernel density centers. The definition of the cluster center p conforms to the formula:

[0093] N(p)={q∈K i |d(p,q)≤∈}

[0094] Among them, d(p,q) represents the Euclidean distance between the cluster center p and the point q in the neighborhood, N(p) is the neighborhood of point p, and ∈ is the set distance threshold, which means all points whose distance from p is within the range of ∈.

[0095] Let the minimum point number threshold be minPts. If there is a point p and |N(p)|≥minPts, indicating that the neighborhood of p contains at least minPts points, then p can be considered a cluster core point. Here, minPts = 3, ∈ = 1.

[0096] Obtain all cluster core points, calculate their geometric center B = (X0, Y0), and regard it as the cluster center.

[0097] By calculating the kernel density center K corresponding to each base station i The distance from the cluster center B is used to determine the base station in the NLOS state. The distance formula is:

[0098]

[0099] Among them, D i is the distance between the core density center and the BSCAN density center of the base station numbered i. The base station corresponding to the point whose distance exceeds the specified threshold is determined to be a NLOS state base station. The calculation formula of the threshold is:

[0100]

[0101] Among them, w i D i The initial weight, w i norm is the weight after normalization of the initial weight. σ is the adjustment factor, which can be adjusted according to the application scenario. It is usually set to 0.30m.

[0102] The distance between KDE and density center of each base station is determined. If D i >δ, then the base station numbered i can be regarded as NLOS state, and its corresponding D i with d i .

[0103] Figure 3 The flowchart of the improved adaptive Kalman filter algorithm provided by the embodiment of the present invention is as follows. Figure 3 As shown, the second aspect of the present invention relates to an improved adaptive Kalman filter algorithm for adapting the UWB data processed by the first aspect of the present invention to perform position solution, and the method includes step 4.

[0104] Furthermore, the step 4 specifically includes the following contents:

[0105] Motion process in a certain plane, initial position x0, y0 and initial x and y axis speed v x0 and v y0 for:

[0106]

[0107] The initial covariance matrix is the identity matrix:

[0108] P0=I

[0109] Furthermore, in the Kalman filter described in step 3, for each moment k in the motion process, a prediction is made based on the state estimation and state transfer matrix of the previous moment:

[0110]

[0111] in, is the predicted state. F is the state transition matrix, which describes the evolution of the system from the previous moment to the current moment:

[0112]

[0113] Furthermore, in the Kalman filter described in step 3, the covariance prediction formula is:

[0114] P k - =FP k-1 F T +Q

[0115] Among them, P k-1 is the covariance matrix of the previous moment, and Q is the process noise covariance matrix. k - is the prediction covariance matrix.

[0116] Furthermore, in the Kalman filter described in step 3, in the updating step, the current measurement value z is used k To adjust the predicted state estimate, the formula is:

[0117]

[0118] in, is the measured value based on the predicted state. k is the observation matrix, which is used to map from the state space to the measurement space. Further, in the Kalman filter described in step 3, H k Calculated based on the location of the valid base station:

[0119] In each measurement cycle, after NLOS identification, the distance data after removing the NLOS anchor point is recorded as d v :

[0120] d v =[d i ], i=1, 2…n

[0121] Among them, d iRepresents the ranging data of the remaining LOS anchor points after removing the NLOS anchor points. n is the number of LOS anchor points at the current moment.

[0122] Furthermore, in the step of calculating the dynamic state observation matrix described in step 3, the nonlinear ranging model and the coordinates of the LOS base station and its ranging data are used to obtain the observation matrix within the measurement period to update the estimated position at the next moment. The measurement model is:

[0123]

[0124] Where i represents the number of the LOS anchor point. For the nonlinear measurement model h(x,y), we obtain the Jacobian matrix H by calculating the partial derivative of the measurement function k , this matrix is used to map the changes in the state space to the measurement space.

[0125] For the ranging model, its Jacobian matrix is:

[0126]

[0127] Calculate the partial derivatives and get:

[0128]

[0129] Furthermore, in the step of calculating the dynamic state observation matrix described in step 3, the form of the observation matrix is:

[0130]

[0131] Since the state observation only involves position information, the elements in the last two columns of the observation matrix are 0.

[0132] Furthermore, in the Kalman filtering process described in step 3, the measurement noise covariance matrix R driven by historical data is introduced. k , use the D obtained in step 2 after NLOS determination i Building R k Since the base stations work independently, the correlation of measurement noise between them can be ignored. At the time k of Kalman filtering, R k The settings are:

[0133]

[0134] Among them, D i is the distance between the kernel density center of the anchor point numbered i in the two-stage clustering process and the BSCAN density center, where i = 1, 2...n, meaning that only the data of the anchor point that is valid (LOS state) at this moment is taken.

[0135] Furthermore, in the Kalman filtering process described in step 3, the measurement residual of the update step is:

[0136]

[0137] Among them, d i Represents the distance between the i-th anchor point and the target position, where (x i ,y i ), i=1,2,...,N are the coordinates of each base station. (x k ,y k ) represents the predicted position of k at the current moment.

[0138] Furthermore, in the Kalman filtering process described in step 3, the adaptive measurement noise R k 'for:

[0139] R k '=R k +(y k y k T )GDOP k

[0140] Among them, y k is the difference between the observed and predicted measurements, GDOP k Represents the GDOP value formed by the current predicted position and the base station. When the target is moving, the GDOP formed by the target and the anchor point has an amplifying effect on the positioning error. Some locations with large GDOPs will increase the error. Therefore, an adaptive measurement noise covariance matrix based on GDOP adjustment is introduced:

[0141] The GDOP formed by the predicted position at time k during motion and the LOS anchor point can be expressed as:

[0142]

[0143] Among them, H v The design matrix formed by the LOS anchor point and the target state estimated position is calculated as:

[0144]

[0145] Furthermore, in the Kalman filtering process described in step 3, the Kalman gain is:

[0146]

[0147] Furthermore, in the Kalman filtering process described in step 3, the state estimate is updated by the Kalman gain and the measurement residual, and the formula is:

[0148]

[0149] Furthermore, in the Kalman filtering process described in step 3, the covariance matrix is updated, and the formula is:

[0150]

[0151] Among them, P k is the updated covariance matrix, representing the new estimation uncertainty.

[0152] The ranging data after NLOS recognition processing can be solved through the improved adaptive Kalman filter to finally obtain accurate positioning information.

[0153] Figure 4 This is a schematic diagram comparing the position solution effects of the solution proposed in the present invention and the standard Kalman filter. A scenario as shown in the figure is set, and the receiver moves according to the green dotted line position. During the movement, each base station measures the real-time position of the receiver. When there is an obstacle between the base station and the receiver, the ranging data is interfered with. The data of the complete movement process are path-solved by the standard Kalman filter and the solution proposed in the present invention, and compared. The root mean square error of the solution proposed in the present invention is 0.076549m, and the average errors in the x-direction and y-direction are 0.035354m and 0.044015m respectively; while the root mean square error of the standard Kalman filter is 0.42824m, and the average errors in the x-direction and y-direction are 0.20744m and 0.23455m respectively. It can be verified that this solution has very good positioning accuracy.

[0154] Beneficial effects of the present invention: The present invention combines two-stage clustering with Kalman filtering, and uses two-stage clustering of KDE kernel density estimation and BSCAN clustering for UWB ranging data under mixed NLOS and LOS conditions to distinguish and discard ranging data under NLOS state, and uses improved adaptive Kalman filtering for solution, which not only ensures the positioning performance, but also greatly improves the positioning accuracy of UWB under NLOS propagation conditions, effectively improving the reliability of UWB positioning. In the fields of military, logistics, health monitoring, etc., this reliability is crucial, especially when the system needs to maintain high-precision positioning under different conditions (such as different physical environments or mobile scenarios), so the application and promotion of UWB of the present invention has a positive effect.

[0155] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0156] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0157] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0159] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which are all protected by the present invention.

Claims

1. A UWB positioning method based on clustering identification and NLOS mitigation, characterized in that: include: Obtain ranging data of n base stations in a measurement cycle; Divide n ranging data into groups of m (n>m) and traverse all possible groups. Each ranging data can be divided into different groups at the same time. For each set of ranging data, a preliminary coarse positioning point is calculated respectively; for each base station, all preliminary coarse positioning points involved in the ranging data of the base station are divided into one group; For all the preliminary positioning coarse measurement points corresponding to each base station, KDE kernel density estimation is performed to obtain the corresponding coarse measurement sites; BSCAN density clustering is performed on all coarse measurement sites to obtain the density center; For each base station, calculating the Euclidean distance between the coarse measurement point corresponding to the base station and the density center; the base station corresponding to the coarse measurement point whose Euclidean distance exceeds a specified threshold is considered to be in the NLOS state; The nonlinear ranging model and the improved adaptive Kalman filter are used to solve the ranging information after discarding the NLOS state base station to obtain the positioning information of the tag to be found.

2. The UWB positioning method based on clustering identification and NLOS mitigation according to claim 1, characterized in that: The preliminary positioning coarse measurement points corresponding to each set of ranging data are respectively calculated based on the least square positioning method.

3. The UWB positioning method based on cluster identification and NLOS mitigation according to claim 1, characterized in that: If m = 3, the base stations are numbered 1, 2…N, and the set of numbers is C∈{1, 2…N}; the formulas for calculating the initial coarse positioning points for each set of ranging data include: X abc =(H T H) -1 H T b Among them, X abc Indicates the coordinates of the initial positioning coarse measurement point corresponding to the base station combination numbered a, b, c; a, b, c represent the base station numbers; H is the coefficient matrix; b is the constant term vector; x a ,y a 、x b ,y b 、x c ,y c is the location of the corresponding base station, d a d b d c is the theoretical distance between the target point and each base station, and x, y are the coordinates of the tag to be found.

4. The UWB positioning method based on clustering identification and NLOS mitigation according to claim 3, characterized in that: The formula for the Euclidean distance between the coarse measurement point corresponding to the base station and the density center is: Among them, D i is the distance between the rough measurement point of the base station numbered i and the density center. The base station corresponding to the point whose distance exceeds the specified threshold is determined to be a NLOS state base station.

5. The UWB positioning method based on clustering identification and NLOS mitigation according to claim 4, characterized in that: The calculation formula for the threshold value specified here is: Among them, δ represents the preset threshold, w i D i The initial weight of is the weight after normalization of the initial weight; σ is the adjustment factor.

6. The UWB positioning method based on clustering identification and NLOS mitigation according to claim 4, characterized in that: The formula for obtaining positioning information by using the nonlinear ranging model and the improved adaptive Kalman filter to solve the ranging information after discarding the NLOS state base station include: P0=I P k - =FP k-1 F T +Q d v =[d i ],i=1,2…n in, is the state estimate updated by the Kalman gain and measurement residual at time k, is the predicted state at time k, K k is the Kalman gain, y k is the measurement residual, is the initial state, P0 is the initial covariance matrix, F is the state transfer matrix, is the predicted state of the previous moment, P k - is the prediction covariance matrix, H k is the observation matrix, R k is the noise covariance matrix, R k ′ is adaptive noise, P k-1 is the covariance matrix of the previous moment, Q is the process noise covariance matrix, d v To remove the ranging data of NLOS state base stations, d i Indicates d v The ranging data of the i-th base station, h(x,y) is the nonlinear measurement model, z i is the distance from the tag to be found to the i-th base station; is the measured value based on the predicted state, GDOP k is the GDOP value formed by the predicted position and the base station at the current moment, H v Design matrix formed for LOS anchor and target state estimation.

7. A UWB positioning system based on clustering identification and NLOS mitigation, characterized in that: include: An acquisition module is used to obtain ranging data of n base stations in a measurement cycle; Divide n ranging data into groups of m (n>m) and traverse all possible groups. Each ranging data can be divided into different groups at the same time; The first grouping module is configured to calculate a preliminary coarse positioning point for each group of ranging data; for each base station, all preliminary coarse positioning points calculated using the ranging data of the base station are grouped into one group; The density center acquisition module is used to perform KDE kernel density estimation on all preliminary positioning coarse measurement points corresponding to each base station to obtain the coarse measurement site corresponding to each base station; and perform BSCAN density clustering on all coarse measurement sites to obtain the density center; an NLOS identification module configured to calculate, for each base station, a Euclidean distance between a coarse measurement point corresponding to the base station and the density center; and to regard a base station corresponding to a coarse measurement point whose Euclidean distance exceeds a specified threshold as being in an NLOS state; The positioning information acquisition module is used to use the nonlinear ranging model and the improved adaptive Kalman filter to solve the ranging information after discarding the NLOS state base station to obtain the required tag positioning information.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the computer program implements the UWB positioning method based on clustering identification and NLOS mitigation according to any one of claims 1 to 6.

Images