TDOA moving target positioning method based on SIDSNRZND-TZ discrete model

Abstract

The present invention discloses a TDOA mobile target positioning method based on a SIDSNRZND-TZ discrete model, comprising: S1, setting a trajectory equation of a signal source, and constructing an acoustic sensor coordinate equation based on the trajectory equation; S2, converting the acoustic sensor coordinate equation into a linear equation; S3, constructing a SINRZND model according to the original OZND equation; S4, constructing a SIDSNRZND-TZ discrete model based on the SINRZND model, and using it to solve the linear equation to obtain the TDOA mobile target position. The present invention designs and adopts a new variable time step in the SIDSNRZND-TZ discrete model, so that the model has a faster convergence speed and higher accuracy. At the same time, the designed strong initial state time-varying coefficient is combined with the integral term, so that the model can solve the time-varying algebraic Ricardi (TVARE) problem in various situations, especially in the time-varying noise environment in the TDOA scene. The convergence characteristics of the SIDSNRZND-TZ discrete model are improved, while maintaining a high convergence accuracy, and has the advantages of high accuracy and low delay.

Description

Technical Field

[0001] The present invention belongs to the technical field of TDOA positioning, and particularly relates to a TDOA mobile target positioning method based on a SIDSNRZND-TZ discrete model; wherein, TDOA positioning refers to positioning using time difference, the SIDSNRZND-TZ model is a strong initial state attenuable step-length noise-resistant zeroing neural dynamics model, and the SINRZND model, which is the construction basis of the SIDSNRZND-TZ model, is a strong initial state noise-resistant zeroing neural dynamics model. Background Art

[0002] In TDOA positioning estimation, the time difference between the acoustic emission source signal reaching different acoustic emission sensors is mainly used for positioning. If the receivers are located on a straight line, there are many optimization processing methods. However, if the receivers are randomly distributed in space, the situation is more complicated, and nonlinear problems will be encountered when solving the hyperbolic equations. Some literatures give an exact solution with the same number of measurement parameters as the number of source signal coordinates. However, when the number of measurement parameters is redundant, this method cannot fully utilize the statistical information given by the redundant measurement parameters to improve positioning accuracy. For the case of redundant measurement parameters, some closed solutions are given in the relevant literature, but these solutions are not optimal. There are also literatures that give an iterative algorithm using Fourier series. This iteration requires a good initial value, otherwise it is easy to fall into the local minimum point, and convergence cannot be guaranteed. There is often no effective global peak, converges to several local maximum values, and is extremely sensitive to the initial search value.

[0003] To solve this problem, the prior art literature proposes a two-step weighted LS method. When the measurement parameter error is small, the performance is close to the optimal value. However, this method introduces the square term of the measurement parameter. When the measurement error is large, the quadratic term of the noise cannot be ignored, and its performance will deteriorate. At the same time, the increase in computational cost and the great interference of noise make it difficult to solve in real time. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model provided by the present invention solves the technical problems existing in the above-mentioned background technology.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a TDOA mobile target positioning method based on a SIDSNRZND-TZ discrete model, comprising the following steps:

[0006] S1. Setting a trajectory equation of a signal source, and constructing a coordinate equation of an acoustic sensor based on the trajectory equation;

[0007] S2, converting the coordinate equation of the acoustic sensor into a linear equation;

[0008] S3. Construct the SINRZND model based on the original OZND equation;

[0009] S4. Based on the SINRZND model, the SIDSNRZND-TZ discrete model is constructed and used to solve the linear equation to obtain the TDOA moving target position.

[0010] Furthermore, in step S1, the coordinate equations of the n sensors constructed are expressed as:

[0011] ;

[0012] In the formula, represents the x-axis coordinate of the nth sensor, Indicates the x-axis coordinate of the main sensor, Indicates the x-axis coordinate of the second sensor, represents the y-axis coordinate of the nth sensor, Indicates the y-axis coordinate of the main sensor, Indicates the y-axis coordinate of the second sensor, represents the z-axis coordinate of the nth sensor, Indicates the z-axis coordinate of the main sensor, Represents the z-axis coordinate of the second sensor, ( , , ) represents the moving target position, is the calculation parameter related to the distance from the second sensor to the origin of the virtual three-dimensional coordinate system, is the calculation parameter related to the distance from the nth sensor to the origin of the virtual three-dimensional coordinate system, Indicates the distance between the sound source and the main sensor, Indicates the distance difference between the sound source and the second sensor and the main sensor. Indicates the distance difference between the sound source from the nth sensor and the main sensor.

[0013] Furthermore, in step S2, the linear equation converted from the coordinate equation of the acoustic sensor is expressed as:

[0014] ;

[0015] The error function corresponding to the linear equation It is expressed as:

[0016] ;

[0017] in, , , .

[0018] Furthermore, in step S3, the constructed SINRZND model includes new strong initial state time-varying coefficients and a variable time step. It is expressed as:

[0019] ;

[0020] In the formula, Represents the SINRZND model The new strong initial state time-varying coefficients, represents the error function of the linear equation, and represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters. express When is the integrand, represents the substitution of time t;

[0021] Among them, the new strong initial state identification coefficient The expression is:

[0022] ;

[0023] The expression for the variable time step is:

[0024] ;

[0025] In the formula, , Indicates control The size of the first and second parameters controls the convergence of the model. represents the initial state enhancement function, , represents a decaying function, , ⌊ ⌋ means rounding down, exp means the power operation of e, e is a natural constant, ] are fixed parameters.

[0026] Furthermore, the step S4 is specifically as follows:

[0027] S41, reconstructing the error function of the linear equation using the constructed SINRZND model;

[0028] S42, based on the reconstruction of the error function, using the constructed SINRZND model to solve the linear equation and determine the preliminary solution of the target position;

[0029] S43, introduce ADZND-TZ variable time step discretization algorithm to further construct discrete model;

[0030] S44. Based on the constructed discrete model, the original variable time step in the ADZND-TZ variable time step discrete algorithm is changed to The discrete form of the function , and combined with the SINRZND model, the SIDSNRZND-TZ discrete model is obtained, and then the discrete solution corresponding to the preliminary solution of the target position is determined to obtain the TDOA moving target position.

[0031] Furthermore, in step S41, the expression for reconstructing the error function of the linear equation is:

[0032] ;

[0033] In the formula, express The first derivative of express The first derivative of express The first derivative of , represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters.

[0034] Furthermore, in step S42, the expression of the preliminary solution of the target position is:

[0035] ;

[0036] In the formula, express The generalized inverse matrix of the error is obtained by taking the first-order derivative, that is, the gradient, so that x, y, and z move along the negative gradient direction. The final result reached along this direction is the position of the moving target.

[0037] Furthermore, in step S43, the ADZND-TZ variable time step discrete algorithm introduced is expressed as:

[0038] when hour:

[0039] ;

[0040] when hour:

[0041] ;

[0042] In the formula, express The discrete form of the function, express The discrete form of express The iterative update value of express The iteration value at the previous moment, represents the iterative value of the first two moments, and the subscript k represents the signal at time Sampling at Represents the original variable time step in the ADZND-TZ variable time step discretization algorithm.

[0043] Furthermore, in step S43, the linear equation is solved using the SIDSNRZND-TZ discrete model, and the obtained TDOA moving target position is expressed as:

[0044] when hour:

[0045] ;

[0046] when hour:

[0047] ;

[0048] In the formula, , , express The discretization of express The discrete form of , Both said The discrete form of , Both said The discrete form of express The discrete form of express The discrete form of the first-order derivative is express The discrete form of the first-order derivative, express The discrete form of express The discrete form of express The discrete form of the signal is The subscript j indicates the sampling of the signal at time Sampling at .

[0049] The beneficial effects of the present invention are:

[0050] (1) The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model proposed in the present invention designs and adopts a new variable time step in the SIDSNRZND-TZ discrete model, so that the model has a faster convergence speed and higher accuracy.

[0051] (2) The strong initial time-varying coefficient designed in the present invention is combined with the integral term, so that the convergence characteristics of the SIDSNRZND-TZ model are improved in solving time-varying ARE problems in various situations, especially in the time-varying noise environment in the TDOA scenario, while maintaining a high convergence accuracy. It has the advantages of high precision and low latency. The SIDSNRZND-TZ model achieves global convergence under different conditions and is superior to existing solutions in terms of convergence speed, accuracy and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of the TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model provided in an embodiment of the present invention.

[0053] Figure 2 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVNARE problem under noise-free conditions.

[0054] Figure 3 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVCARE problem under noise-free conditions.

[0055] Figure 4 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVSE problem under noise-free conditions.

[0056] Figure 5 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVCLE problem under noise-free conditions.

[0057] Figure 6 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVLME problem under noise-free conditions.

[0058] Figure 7 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVMIE problem under noise-free conditions;

[0059] Figure 8This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVNARE problem under constant noise conditions.

[0060] Fig. 9 This is a logarithmic comparison of the convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVNARE problem under linear time-varying noise conditions.

[0061] Fig.10 Comparison of the logarithmic convergence accuracy of the SIDSNRZND-TZ discrete model and other discrete models for solving the TVNARE problem under bounded random noise.

[0062] Fig.11 This is the logarithmic graph of the convergence accuracy of the SIDSNRZND-TZ discrete model in the dynamic target detection process in the noise-free case.

[0063] Fig.12 The figure shows the trajectory comparison between the target path and the calculated path when the TDOA method is used to predict the displacement of the sound source position, as well as a schematic diagram of the position coordinates of the sensor.

[0064] Fig.13 Schematic diagram of the comparison of x-axis error, y-axis error and z-axis error during dynamic target detection. DETAILED DESCRIPTION

[0065] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0066] The embodiment of the present invention provides a TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model, such as Figure 1 As shown, the following steps are included:

[0067] S1. Setting a trajectory equation of a signal source, and constructing a coordinate equation of an acoustic sensor based on the trajectory equation;

[0068] S2, converting the coordinate equation of the acoustic sensor into a linear equation;

[0069] S3. Construct the SINRZND model based on the original OZND equation;

[0070] S4. Based on the SINRZND model, the SIDSNRZND-TZ discrete model is constructed and used to solve the linear equation to obtain the TDOA moving target position.

[0071] In step S1 of the embodiment of the present invention, illustratively, the trajectory equation of the set signal source is expressed as:

[0072] ;

[0073] In the formula, is the theoretical trajectory of the sound source, t is the time value, is a real number with three rows and one column.

[0074] In this embodiment, the coordinates of the n acoustic sensors used to receive signals are assigned as follows:

[0075] ;

[0076] The time difference between the arrival time of each sub-sensor and the arrival time of the main sensor is set to , the distance difference between each sub-sensor and the target sensor is equal to the difference between the arrival time of each sub-sensor and the arrival time of the main sensor The speed of sound The product of , so the distance from each sub-sensor to the origin of the virtual three-dimensional coordinate system is , so the equation for n sensors is:

[0077] ;

[0078] In the formula, represents the x-axis coordinate of the nth sensor, Indicates the x-axis coordinate of the main sensor, Indicates the x-axis coordinate of the second sensor, represents the y-axis coordinate of the nth sensor, Indicates the y-axis coordinate of the main sensor, Indicates the y-axis coordinate of the second sensor, represents the z-axis coordinate of the nth sensor, Indicates the z-axis coordinate of the main sensor, Represents the z-axis coordinate of the second sensor, ( , , ) represents the moving target position, is the calculation parameter related to the distance from the second sensor to the origin of the virtual three-dimensional coordinate system, is the calculation parameter related to the distance from the nth sensor to the origin of the virtual three-dimensional coordinate system, Indicates the distance between the sound source and the main sensor, Indicates the distance difference between the sound source and the second sensor and the main sensor. Indicates the distance difference between the sound source from the nth sensor and the main sensor.

[0079] Exemplarily, the calculation parameters related to the distance from the sensor to the origin of the virtual three-dimensional coordinate system are The calculation formula is:

[0080] ;

[0081] In the formula, , is the distance from the main sensor to the origin of the virtual 3D coordinate system, Represents the distance from the i-th sensor to the virtual 3D coordinate origin.

[0082] In step S2 of the embodiment of the present invention, the linear equation converted from the coordinate equation of the acoustic sensor is expressed as:

[0083] ;

[0084] The error function corresponding to the linear equation It is expressed as:

[0085] ;

[0086] in, , , .

[0087] In step S3 of the embodiment of the present invention, the constructed SINRZND model includes new strong initial state time-varying coefficients and a variable time step. It is expressed as:

[0088] ;

[0089] In the formula, Represents the SINRZND model The new strong initial state identification coefficient, represents the error function of the linear equation, and represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters. express When is the integrand, represents the substitution of time t;

[0090] Among them, the new strong initial state identification coefficient The expression is:

[0091] ;

[0092] The expression for the variable time step is:

[0093] ;

[0094] In the formula, , Indicates control The size of the first and second parameters controls the convergence of the model. represents the initial state enhancement function, , represents a decaying function, , ⌊ ⌋ means rounding down, exp means the power operation of e, e is a natural constant, ] are fixed parameters.

[0095] Step S4 of the embodiment of the present invention is specifically as follows:

[0096] S41, reconstructing the error function of the linear equation using the constructed SINRZND model;

[0097] S42, based on the reconstruction of the error function, using the constructed SINRZND model to solve the linear equation and determine the preliminary solution of the target position;

[0098] S43, introduce ADZND-TZ variable time step discretization algorithm to further construct discrete model;

[0099] S44. Based on the constructed discrete model, the original variable time step in the ADZND-TZ variable time step discrete algorithm is changed to The discrete form of the function , and combined with the SINRZND model, the SIDSNRZND-TZ discrete model is obtained, and then the discrete solution corresponding to the preliminary solution of the target position is determined to obtain the TDOA moving target position.

[0100] In step S41 of this embodiment, the expression for reconstructing the error function of the linear equation is:

[0101] ;

[0102] In the formula, express The first derivative of express The first derivative of express The first derivative of , represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters.

[0103] In step S42 of this embodiment, the expression of the preliminary solution of the target position is:

[0104] ;

[0105] In the formula, express The generalized inverse matrix of the error is obtained by taking the first-order derivative, that is, the gradient, so that x, y, and z move along the negative gradient direction. The final result reached along this direction is the position of the moving target.

[0106] In step S43 of this embodiment, the introduced ADZND-TZ variable time step discretization algorithm is expressed as:

[0107] when hour:

[0108] ;

[0109] when hour:

[0110] ;

[0111] In the formula, express The discrete form of the function, express The discrete form of express The iterative update value of express The iteration value at the previous moment, represents the iterative value of the first two moments, and the subscript k represents the signal at time Sampling at Represents the original variable time step in the ADZND-TZ variable time step discretization algorithm.

[0112] In step S43 of the embodiment of the present invention, the discrete solution corresponding to the preliminary solution of the target position is determined by using the SIDSNRZND-TZ discrete model and is expressed as:

[0113] when hour:

[0114] ;

[0115] when hour:

[0116] ;

[0117] In the formula, , , express The discretization of express The discrete form of , Both said The discrete form of , Both said The discrete form of express The discrete form of express The discrete form of the first-order derivative, express The discrete form of the first-order derivative, express The discrete form of express The discrete form of express The discrete form of the signal is The subscript j indicates the sampling of the signal at time Sampling at .

[0118] In an embodiment of the present invention, the SIDSNRZND-TZ discrete model constructed above can not only realize TDOA mobile target positioning, but also solve the time-varying algebraic Riccati (TVARE) problem in various situations. On this basis, this embodiment provides six cases in which the SIDSNRZND-TZ discrete model is used to solve the TVARE problem.

[0119] (1) Time-varying asymmetric algebraic Riccati (TVNARE) problem:

[0120] ;

[0121] in, , , , are all smooth time-varying matrices, is a zero matrix, is the unknown matrix to be solved m and n represent arbitrary real numbers, and t represents a time variable.

[0122] On this basis, the error function of the TVNARE problem is constructed as follows:

[0123] ;

[0124] According to the above error equation, the original OZND model is defined as:

[0125] ;

[0126] In the formula, The time derivative of is a scale parameter that controls the convergence speed, and the appropriate This embodiment constructs a SINRZND model based on the original OZND model. In addition to the original integral term, the constructed SINRZND model also includes a new strong initial state time-varying coefficient and a variable time step, wherein the expression of the new strong initial state time-varying coefficient is:

[0127] ;

[0128] Among them, Ω is the initial state enhancement function, and its expression is:

[0129] ;

[0130] Among them, ⌊⌋ means rounding down, exp means the power operation of e, and e is a natural constant;

[0131] The expression of the new variable time step is:

[0132] ;

[0133] in, is a fixed parameter.

[0134] Therefore, the constructed SINRZND model is:

[0135] ;

[0136] Among them, the integral term It has negative feedback, which is mainly used to suppress noise. and are scale parameters that control the convergence speed and are fixed real parameters. represents the first-order derivative of E(t), τ represents the substitution of time t, express 's points.

[0137] The general solution process of the SINRZND model to the TVARE problem is introduced below.

[0138] First, the error equation Derivative on both sides of :

[0139] ;

[0140] In the formula and It is expressed as:

[0141] ;

[0142] According to the expression of SINRZND model, after substituting it into both sides of the error equation and taking the derivative, we get:

[0143] ;

[0144] Taking into account matrix vectorization and Kronecker product, the above formula can be rewritten as:

[0145] ;

[0146] Among them, the matrix ;

[0147] .

[0148] In the above matrix represents the Kronecker product, Pointing to quantization operation, represents the matrix transpose operator, and are all unit matrices.

[0149] Then the solution to the TVNARE problem can be expressed as:

[0150] ;

[0151] in, Represents the generalized inverse matrix operator of a matrix.

[0152] Considering the use on digital hardware, the ADZND-TZ variable time step discretization algorithm is used:

[0153] when When (ADZND-E algorithm):

[0154] ;

[0155] when hour,

[0156] ;

[0157] The subscript k represents the signal at time Sampling at is a variable time step.

[0158] The original variable time step of the ADZND-TZ discrete algorithm is changed to the variable time step proposed in the embodiment, and the SIDSNRZND-TZ discrete model is further created in combination with the SINRZND model. Then, the discrete solution of the TVNARE problem can be expressed as:

[0159] when When (SIDSNRZND-E algorithm):

[0160] ;

[0161] when hour,

[0162] ;

[0163] (2) Time-varying continuous algebraic Ricardi (TVCARE) problem:

[0164] ;

[0165] in, , , are all smooth time-varying matrices, is a zero matrix, is the unknown matrix to be solved, represents any real number, and t represents the time variable. That is, the coefficient matrix in the TVNARE problem .

[0166] On this basis, the error function of the TVCARE problem is constructed as:

[0167] ;

[0168] Repeating the solution steps of problem (1), we get:

[0169] ;

[0170] The solution to the TVCARE problem using the SINRZND model is:

[0171] ;

[0172] Among them, the matrix , .

[0173] Further discretization solves the TVCARE problem and the discrete solution is:

[0174] when When (SIDSNRZND-E algorithm):

[0175] ;

[0176] when hour,

[0177] ;

[0178] (3) Time-varying Sylvester (TVSE) problem:

[0179] ;

[0180] in, , , are all smooth time-varying matrices, is a zero matrix, is the unknown matrix to be solved, represents any real number, t represents the time variable, that is, the coefficient matrix in the TVNARE problem .

[0181] On this basis, the error function of the TVCARE problem is constructed as:

[0182] ;

[0183] Repeating the solution steps of problem (1), we get:

[0184] ;

[0185] The solution to the TVSE problem using the SINRZND model is:

[0186] ;

[0187] Among them, the matrix , .

[0188] Further discretization solves the TVCARE problem and the discrete solution is:

[0189] when When (SIDSNRZND-E algorithm):

[0190] ;

[0191] when hour,

[0192] ;

[0193] (4) Time-varying continuous Lyapunov (TVCLE) problem:

[0194] ;

[0195] in, , are all smooth time-varying matrices, is a zero matrix, is the unknown matrix to be solved, represents any real number, and t represents the time variable. That is, the coefficient matrix in the TVNARE problem , .

[0196] On this basis, the error function of the TVCLE problem is constructed as:

[0197] ;

[0198] Repeating the solution steps of problem (1), we get:

[0199] ;

[0200] The solution to the TVSE problem using the SINRZND model is:

[0201] ;

[0202] Among them, the matrix , .

[0203] Further discretization solves the TVCARE problem and the discrete solution is:

[0204] when When (SIDSNRZND-E algorithm):

[0205] ;

[0206] when hour,

[0207] ;

[0208] (5) Time-varying linear matrix (TVLME) problem:

[0209] ;

[0210] or

[0211] ;

[0212] in, , , are all smooth time-varying matrices, is a zero matrix, is the unknown matrix to be solved, represents any real number, t represents the time variable, that is, the coefficient matrix in the TVNARE problem , or , .

[0213] On this basis, the error function of the TVLME problem is constructed as:

[0214] ;

[0215] or

[0216] ;

[0217] Repeating the solution steps of problem (1), we get:

[0218] ;

[0219] or

[0220] ;

[0221] The solution to the TVLME problem using the SINRZND model is:

[0222] ;

[0223] Among them, the matrix , .

[0224] Further discretization solves the TVLME problem and obtains the discrete solution:

[0225] when When (SIDSNRZND-E algorithm):

[0226] ;

[0227] when hour,

[0228] ;

[0229] (6) Time-varying matrix inversion (TVMIE) problem:

[0230] ;

[0231] in, is a smooth time-varying matrix, The identity matrix, is a zero matrix, is the unknown matrix to be solved, represents any real number, and t represents the time variable. That is, the coefficient matrix in the TVNARE problem , , .

[0232] On this basis, the error function of the TVMIE problem is constructed as:

[0233] ;

[0234] Repeating the solution steps of problem (1), we get:

[0235] ;

[0236] The solution to the TVSE problem using the SINRZND model is:

[0237] ;

[0238] Among them, the matrix , .

[0239] Further discretization solves the TVCARE problem and the discrete solution is:

[0240] when When (SIDSNRZND-E algorithm):

[0241] ;

[0242] when hour,

[0243] .

[0244] The embodiments of the present invention provide examples for comparing the model of the present invention with other discrete models.

[0245] In this embodiment, the new discrete model proposed in the present invention is compared with other discrete models, such as TDZND-E, TDZND-TZ, ADZND-E and ADZND-TZ discrete models. Figure 2-Figure 10 shown. Figure 2-Figure 7 The figure shows the comparison of logarithmic convergence accuracy of different discrete models for solving TVARE problems in various situations in the noise-free case. Figure 8-Figure 10 The logarithmic convergence accuracy comparison diagram of different discrete models for solving TVNARE problems under different noise conditions is shown. The results show that the SIDSNRZND-TZ model is superior in solving TVARE problems in various situations under different noise conditions.

[0246] More specifically, Figure 2-Figure 7 Together, we show the residual accuracy of different models in solving the TVARE problem under noise-free conditions. Figure 2-Figure 7 It can be seen more intuitively that the convergence speed of the SIDSNRZND-TZ model is faster than that of other models. In terms of accuracy, the proposed main model SIDSNRZND-TZ and sub-model SIDSNRZND-E are far more accurate than the TDZND-E model, TDZND-TZ model, ADZND-E model and ADZND-TZ model in solving TVARE problems in various situations.

[0247] Figure 8 The residual accuracy of different models for solving the TVNARE problem under constant noise conditions is shown. Compared with the noise-free condition, the convergence speed of the SIDSNRZND-TZ model is not much different. At the same time, the convergence accuracy of the SIDSNRZND-TZ model is much higher than that of the other five models, and except for the sub-model SIDSNRZND-E model, the other models have divergence. Therefore, under constant noise conditions, the SIDSNRZND-TZ model has better performance than other models.

[0248] Fig. 9 The residual accuracy of different models for solving the TVNARE problem under time-varying linear noise is shown. Fig. 9 , we found that due to the influence of time-varying noise, except for the model SIDSNRZND-TZ and the sub-model SIDSNRZND-E of the present invention, other models diverge and fail to converge. It is worth noting that our SIDSNRZND-TZ model still converges to 10 -4 Therefore, the SIDSNRZND-TZ model has a significant denoising effect. Therefore, the performance of the SIDSNRZND-TZ model is more competitive than other state-of-the-art solutions.

[0249] Fig.10 The residual accuracy of different models for solving the TVNARE problem under bounded random noise is shown. Fig.10 , we find that in the case of bounded random noise, there is a significant drop in accuracy compared to all models. It is worth noting that our SIDSNRZND-TZ model still converges to 10 -3 Therefore, the performance of the SIDSNRZND-TZ model is still competitive with other state-of-the-art solutions.

[0250] The experimental simulation results of using SIDSNRZND-TZ model to solve TDOA problem are as follows: Figure 11-13 As shown. Fig.11It can be seen that the convergence accuracy of the SIDSNRZND-TZ model can still reach 10 when the parameters are moderate. -3 From Fig.12 It can be seen from that the target trajectory is highly consistent with the actual trajectory. No matter what the initial value is, the final trajectory will tend to the target trajectory. Then, from Fig.13 In the figure, we can observe the errors of the x-axis, y-axis, and z-axis. The system error converges to zero quickly within 0.1s and remains globally stable. The error reaches 10 -9 It can be seen that the SIDSNRZND-TZ model proposed in the present invention has the advantages of high precision and low delay in solving the TDOA problem.

[0251] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0252] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model is characterized by: The following steps are involved: S1. Setting a trajectory equation of a signal source, and constructing a coordinate equation of an acoustic sensor based on the trajectory equation; S2, converting the coordinate equation of the acoustic sensor into a linear equation; S3. Construct the SINRZND model based on the original OZND equation; S4. Based on the SINRZND model, the SIDSNRZND-TZ discrete model is constructed and used to solve the linear equation to obtain the TDOA moving target position; Among them, the SINRZND model is a strong initial state noise-resistant zeroing neural dynamics model, and the SIDSNRZND-TZ model is a strong initial state decayable step-length noise-resistant zeroing neural dynamics model; in step S3, the original OZND equation is: ； In the formula, The time derivative of represents the scale parameter that controls the convergence speed; The constructed SINRZND model includes new strong initial state time-varying coefficients and variable time steps. It is expressed as: ； In the formula, Represents the SINRZND model The new strong initial state time-varying coefficients, represents the error function of the linear equation, and represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters. express When is the integrand, represents the substitution of time t; Among them, the new strong initial state identification coefficient The expression is: ； The expression for the variable time step is: ； In the formula, , Indicates control The size of the first and second parameters controls the convergence of the model. represents the initial state enhancement function, , represents a decaying function, , ⌊ ⌋ means rounding down, exp means the power operation of e, e is a natural constant, ] is a fixed parameter; The step S4 is specifically as follows: S41, reconstructing the error function of the linear equation using the constructed SINRZND model; S42, based on the reconstruction of the error function, using the constructed SINRZND model to solve the linear equation and determine the preliminary solution of the target position; S43, introduce ADZND-TZ variable time step discretization algorithm to further construct discrete model; S44. Based on the constructed discrete model, the original variable time step in the ADZND-TZ variable time step discrete algorithm is changed to The discrete form of the function , and combined with the SINRZND model, the SIDSNRZND-TZ discrete model is obtained, and then the discrete solution corresponding to the preliminary solution of the target position is determined to obtain the TDOA moving target position.

2. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 1 is characterized in that: In step S1, the coordinate equations of the n sensors constructed are expressed as: ； In the formula, represents the x-axis coordinate of the nth sensor, Indicates the x-axis coordinate of the main sensor, Indicates the x-axis coordinate of the second sensor, represents the y-axis coordinate of the nth sensor, Indicates the y-axis coordinate of the main sensor, Indicates the y-axis coordinate of the second sensor, represents the z-axis coordinate of the nth sensor, Indicates the z-axis coordinate of the main sensor, Represents the z-axis coordinate of the second sensor, ( , , ) represents the moving target position, is the calculation parameter related to the distance from the second sensor to the origin of the virtual three-dimensional coordinate system, is the calculation parameter related to the distance from the nth sensor to the origin of the virtual three-dimensional coordinate system, Indicates the distance between the sound source and the main sensor, Indicates the distance difference between the sound source and the second sensor and the main sensor. Indicates the distance difference between the sound source from the nth sensor and the main sensor.

3. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 2 is characterized in that: In step S2, the linear equation converted from the coordinate equation of the acoustic sensor is expressed as: ； The error function corresponding to the linear equation It is expressed as: ； in, , , .

4. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 1 is characterized in that: In step S41, the expression for reconstructing the error function of the linear equation is: ； In the formula, express The first derivative of express The first derivative of express The first derivative of , represents the first scale parameter and the second scale parameter that control the convergence speed, and both are fixed real parameters.

5. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 4 is characterized in that: In step S42, the expression of the preliminary solution of the target position is: ； In the formula, express The generalized inverse matrix of the error is obtained by taking the first-order derivative, that is, the gradient, so that x, y, and z move along the negative gradient direction. The final result reached along this direction is the position of the moving target.

6. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 5 is characterized in that: In step S43, the introduced ADZND-TZ variable time step discrete algorithm is expressed as: when hour: ； when hour: ； In the formula, express The discrete form of the function, express The discrete form of express The iterative update value of express The iteration value at the previous moment, represents the iterative value of the first two moments, and the subscript k represents the signal at time Sampling at Represents the original variable time step in the ADZND-TZ variable time step discretization algorithm.

7. The TDOA mobile target positioning method based on the SIDSNRZND-TZ discrete model according to claim 6 is characterized in that: In step S43, the linear equation is solved using the SIDSNRZND-TZ discrete model, and the obtained TDOA moving target position is expressed as: when hour: ； when hour: ； In the formula, , , express The discretization of express The discrete form of , Both said The discrete form of , Both said The discrete form of express The discrete form of express The discrete form of the first-order derivative is express The discrete form of the first-order derivative is express The discrete form of express The discrete form of express The discrete form of the signal is The subscript j indicates the sampling of the signal at time Sampling at .

Images