CN113407909A - Tasteless algorithm for non-analytic complex nonlinear system - Google Patents
Tasteless algorithm for non-analytic complex nonlinear system Download PDFInfo
- Publication number
- CN113407909A CN113407909A CN202110801131.XA CN202110801131A CN113407909A CN 113407909 A CN113407909 A CN 113407909A CN 202110801131 A CN202110801131 A CN 202110801131A CN 113407909 A CN113407909 A CN 113407909A
- Authority
- CN
- China
- Prior art keywords
- covariance
- algorithm
- tasteless
- signal
- complex
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 42
- 230000009967 tasteless effect Effects 0.000 title claims abstract description 18
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 230000003321 amplification Effects 0.000 claims abstract description 9
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 17
- 239000000126 substance Substances 0.000 claims description 8
- 238000012886 linear function Methods 0.000 claims description 2
- 230000003190 augmentative effect Effects 0.000 description 8
- 238000000034 method Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000000342 Monte Carlo simulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 230000003416 augmentation Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Theoretical Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Analysis (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Computing Systems (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a tasteless algorithm for a non-analytic nonlinear complex system, which comprises the following steps: calculating second-order statistics of the input signals in the amplification form, namely amplification variance and covariance, inputting the second-order statistics into a system, and calculating sigma points; the tasteless algorithm can obtain first-order and second-order statistics of the output points corresponding to the nonlinear system through the calculation of a small number of sigma characteristic points; and by researching the relation between the error and the non-roundness, beta parameters are adjusted, and the algorithm error is reduced. The invention can obtain approximate second-order statistic of the output signal by calculating the statistical property of a small number of sigma points aiming at the condition that the input signal is a complex signal and the system function is nonlinear non-analytic, and reduce the system error by researching the relation between beta parameters and non-roundness when the signal is non-circular.
Description
Technical Field
The invention relates to the technical field of complex signal processing, in particular to a tasteless algorithm for a non-analytic complex nonlinear system.
Background
An Unscented Transform (UT) is an algorithm that handles non-linear transfer of mean and covariance. Finite parameters are used to approximate the probabilistic statistics of random quantities by determining finite sigma points. A large amount of computation and data storage can be avoided. The method is often applied to actual scenes such as automatic control, navigation, tracking, artificial intelligence, fault estimation and the like by combining with a Kalman filtering algorithm, and state estimation under a nonlinear system function is carried out.
But the traditional tasteless algorithm mainly aims at real signals, the research on the complex signals is still less, especially the research on the non-analytic mapping function is blank,
disclosure of Invention
It is therefore an objective of the claimed invention to provide a non-odor algorithm for non-analytic complex nonlinear systems to solve the above-mentioned problems.
In order to achieve the purpose, the invention adopts the following technical scheme:
a tasteless algorithm for non-analytic complex nonlinear systems, comprising the steps of:
step S1, calculating second-order statistics of the input signal in the amplification form, inputting the second-order statistics into a nonlinear system, and calculating sigma points;
step S2, according to the sigma point calculated in the step S1, a tasteless algorithm is adopted to obtain first-order and second-order statistics of the output point corresponding to the nonlinear system;
and step S3, when the input signal is non-circular, calculating the optimal beta parameter by considering the relation between the error and the non-circular degree, and reducing the system error.
Further, the step S1 specifically includes:
step S101, the input signal is a complex signal, and the complex signal and the conjugate signal thereof are combined into an amplification form:x=[xT,xH]Twherein x is a complex signal, xTExpressed as transposes of complex signals, xHExpressed as a conjugate transpose of the complex signal;
step S102, calculating second-order statistics according to the input signal, wherein the second-order statistics comprise a covariance matrix and a pseudo covariance matrix, and the expression of the covariance matrix is as follows:
wherein the content of the first and second substances,expressed as a covariance matrix of the input signal, RxExpressed as the covariance, P, of the input signalxExpressed as the pseudo-covariance of the input signal, i.e. as the conjugate signAndrepresenting the covariance and conjugate of the pseudo-covariance, respectively.
The expression of the pseudo covariance matrix is:
step S103, sigma point calculation:
wherein the content of the first and second substances,is a sigma point, especiallyIs a mean point, where p ═ L is the dimension of x,Is the mean value of x and lambda is a scale parameter.
Further, the step S2 specifically includes:
step S201, the weight of the tasteless algorithm is calculated as:
wherein, Wi (m)And Wi (c)Weight coefficients respectively representing mean value and covariance calculation, L is input dimension, alpha and beta respectively represent scale parameters, alpha is used for controlling the distance between a sigma point and the mean value, and the value of alpha is generally smaller and is about 10-3β includes a part of the input distribution information, and when the distribution is gaussian, the value is generally 2.
In step S202, the statistic of the output y is calculated as:
wherein the content of the first and second substances,respectively input and output, in particularIs a mean value point,Is the mean, R 'of the output'yAnd P'yCovariance and pseudo-covariance of y, respectively.
Further, in the step S3, the optimal beta parameter includes a parameter β that minimizes a pseudo covariance errorPAnd a parameter beta that minimizes the covariance errorR;
The parameter betaPThe expression of (a) is:
wherein the content of the first and second substances,H xxin order to be a Hessian matrix,for the i-th derivative to the non-linear function,
the parameter betaRThe expression of (a) is:
the invention has the beneficial effects that:
for an analytic nonlinear system, the algorithm used by the method is basically consistent with the result of actual output after the Monte Carlo algorithm, the mean value, the variance and the covariance of the output can be accurately calculated, and the method has better accuracy compared with the traditional augmentation algorithm.
Drawings
FIG. 1 is a schematic diagram illustrating the effect of changing the non-circular coefficient on the systematic error in example 2 when β is 2.
Fig. 2 shows an example 2 in which f ═ x is analytically converted when a complex circle signal (non-circularity ρ ═ 0) is obtained2The results of the traditional augmented UT algorithm and the algorithm are shown in the figure.
Fig. 3 shows the case where the complex non-circular signal (non-circularity ρ ═ 0.6) is analytically transformed into f ═ x in example 22The results of the traditional augmented UT algorithm and the algorithm are shown in the figure. .
FIG. 4 is a flow chart of an algorithm for a tasteless algorithm for a non-analytic complex nonlinear system in example 2.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The embodiment provides a tasteless algorithm for an unresolved complex nonlinear system, comprising the following steps:
step 101, forming an amplification form from the complex signal and its conjugate signalx=[xT,xH]T
102, calculating second-order statistics, wherein the covariance is as follows:
the pseudo covariance is:
the augmented covariance matrix and the augmented pseudo covariance both include the covariance R of the original xxAnd pseudo covariance PxAnd (4) information.
Step 103, sigma point calculation:
can be abbreviated as:
wherein sigmaiThe ith column root mean square matrix representing the augmented covariance or augmented pseudocovariance.
Wherein the system function of the non-analytic complex system is related to x and its conjugate, and for the non-analytic complex nonlinear system, the Taylor expansion is:
the mean value is:
the sigma point Taylor expansion of the UT algorithm is:
in step 201, the weight of the Unscented Transform (UT) is calculated as:
step 202, substituting weights, the statistics of the output y can be calculated as:
the approximate mean of the UT algorithm is calculated as:
odd positive and negative signs cancel each other
Wherein
So that there are
Note that in the actual operation of the UT the function is still considered as an analytic function, and the derivation for the complex number is introduced into the formula only by the augmented vector, so that in averaging, the pseudo-covariance is still brought inSigma point of (d):
The use of the augmented input thus enables the UT algorithm to still guarantee the first-second order accuracy for non-analytic functions.
Step 301, the UT algorithm pseudo covariance is calculated as:
wherein due to
And is
The fourth order can be written as:
for the actual covariance, the mean of the taylor-expanded odd terms at the mean is all 0, while the odd terms and higher orders are controlled in UT by a coefficient α close to 0, while the actual result of the second order is consistent with the result of the UT calculation.
And the error term of the fourth order is
When in useH xxIf the matrix is diagonal, then a minimum error can be obtained with a beta of 2, however, when the matrix is diagonalH xxWhen the diagonal matrix is off, i.e. f is not resolved, simply defining β as 2, the error of the previous term cannot be removed.
When β is 1+ ρ*And UT is consistent with the calculation result of the Monte Carlo method, wherein rho is a non-circular coefficient,
step 302, for UT covariance, should be brought inPoint calculation covariance, calculated as the optimum value of β
E.g. when f ═ x2=xx*Is provided with
When β is ρ*+ρ+ρ*UT is consistent with the calculation result of Monte Carlo method at-1.
Example 2
Referring to fig. 1-4, the present embodiment also provides a non-odor algorithm for non-analytic complex nonlinear systems, which can be used for, but not limited to, mean and second order statistic analysis after complex nonlinear analytic transformation, and also for real and analytic function transformation. For the non-analytic function, the embodiment takes f ═ x tint2=xx*For example, the influence of input non-roundness on system errors is analyzed, and the calculation of the algorithm after the parameters are optimized is shownThe accuracy of the calculation. For the analytic function, for convenience of comparison, the present example uses a complex number to undergo analytic transformation of f ═ x2For example, the superiority of the algorithm of the present invention is shown by the error result.
As shown in fig. 4, the algorithm of the present invention includes the following steps:
step S1, calculating second-order statistics of the input signal in the amplification form, namely amplification variance and covariance, inputting the second-order statistics into a system, and calculating sigma points;
step S2, the tasteless algorithm can obtain first-order and second-order statistics of the corresponding output points of the nonlinear system through the calculation of a small number of sigma characteristic points;
and step S3, adjusting beta parameters by researching the relation between the error and the non-roundness, and reducing the error of the algorithm.
The calculated sigma point satisfies the following equation:
the weight of the Unscented Transform (UT) is calculated as:
the statistic of output y can be calculated as:
parameter beta that minimizes the pseudo-covariance errorPSatisfy the requirement of
Parameter beta that minimizes covariance errorRSatisfy the requirement of
Specifically, there are different optimal β parameters for different nonlinear functions, but all relate to the non-circularity of the signal itself. As shown in fig. 1, the non-circularity of the input affects the error of the system, which is why the optimal parameters found in many papers are not constant. As shown in Table 1, the effect of non-circularity is balanced, and the error of the system is greatly reduced after parameters changing along with the input non-circularity are adopted.
TABLE 1
For the analytic nonlinear system, as shown in fig. 2 and fig. 3, the result of the algorithm used by the invention is basically consistent with the result of the actual output after the Monte Carlo algorithm, and the mean value, the variance and the covariance of the output can be accurately calculated. Has better accuracy than the prior augmentation algorithm.
The invention is not described in detail, but is well known to those skilled in the art.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.
Claims (4)
1. A tasteless algorithm for use in a non-analytic complex nonlinear system, comprising the steps of:
step S1, calculating second-order statistics of the input signal in the amplification form, inputting the second-order statistics into a nonlinear system, and calculating sigma points;
step S2, according to the sigma point calculated in the step S1, a tasteless algorithm is adopted to obtain first-order and second-order statistics of the output point corresponding to the nonlinear system;
and step S3, when the input signal is non-circular, calculating the optimal beta parameter by considering the relation between the error and the non-circular degree, and reducing the system error.
2. The tasteless algorithm for an unresolved complex nonlinear system according to claim 1, wherein the step S1 specifically comprises:
step S101, the input signal is a complex signal, and the complex signal and the conjugate signal thereof are combined into an amplification form:x=[xT,xH]Twherein x is a complex signal, xTExpressed as transposes of complex signals, xHExpressed as a conjugate transpose of the complex signal;
step S102, calculating second-order statistics according to the input signal, wherein the second-order statistics comprise a covariance matrix and a pseudo covariance matrix, and the expression of the covariance matrix is as follows:
wherein the content of the first and second substances,expressed as a covariance matrix of the input signal, RxExpressed as the covariance, P, of the input signalxExpressed as the pseudo-covariance of the input signal, i.e. as the conjugate signAndrepresenting the covariance and conjugate of the pseudo-covariance, respectively.
The expression of the pseudo covariance matrix is:
step S103, sigma point calculation:
3. The tasteless algorithm for an unresolved complex nonlinear system according to claim 2, wherein the step S2 specifically comprises:
step S201, the weight of the tasteless algorithm is calculated as:
wherein, Wi (m)And Wi (c)Weight coefficients respectively representing mean and covariance calculations, L being an input dimension, α and β respectively representing scale parameters, α being 10-3Beta contains part of input distribution information, and when the distribution is Gaussian distribution, the value is 2;
in step S202, the statistic of the output y is calculated as:
4. The tasteless algorithm for an unresolved complex nonlinear system of claim 3, wherein in step S3, the optimal beta parameter comprises a parameter β that minimizes a pseudo-covariance errorPAnd a parameter beta that minimizes the covariance errorR;
The parameter betaPThe expression of (a) is:
wherein the content of the first and second substances,H xxin order to be a Hessian matrix,for the i-th derivative to the non-linear function,
the parameter betaRThe expression of (a) is:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110801131.XA CN113407909B (en) | 2021-07-15 | 2021-07-15 | Calculation method of odorless algorithm for non-analytic complex nonlinear system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110801131.XA CN113407909B (en) | 2021-07-15 | 2021-07-15 | Calculation method of odorless algorithm for non-analytic complex nonlinear system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113407909A true CN113407909A (en) | 2021-09-17 |
CN113407909B CN113407909B (en) | 2024-01-09 |
Family
ID=77686403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110801131.XA Active CN113407909B (en) | 2021-07-15 | 2021-07-15 | Calculation method of odorless algorithm for non-analytic complex nonlinear system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113407909B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110070211A (en) * | 2009-12-18 | 2011-06-24 | 두산디에스티주식회사 | Signal processing system and method using random scaled unscented transformation |
CN110231636A (en) * | 2019-04-29 | 2019-09-13 | 深圳市联和安业科技有限公司 | The adaptive Unscented kalman filtering method of GPS and BDS dual mode satellite navigation system |
CN110501686A (en) * | 2019-09-19 | 2019-11-26 | 哈尔滨工程大学 | A kind of method for estimating state based on NEW ADAPTIVE high-order Unscented kalman filtering |
CN111506873A (en) * | 2020-04-13 | 2020-08-07 | 深圳埃瑞斯瓦特新能源有限公司 | Nonlinear system parameter estimation method based on extended Kalman filtering algorithm |
CN111985093A (en) * | 2020-08-03 | 2020-11-24 | 哈尔滨工程大学 | Adaptive unscented Kalman filtering state estimation method with noise estimator |
-
2021
- 2021-07-15 CN CN202110801131.XA patent/CN113407909B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110070211A (en) * | 2009-12-18 | 2011-06-24 | 두산디에스티주식회사 | Signal processing system and method using random scaled unscented transformation |
CN110231636A (en) * | 2019-04-29 | 2019-09-13 | 深圳市联和安业科技有限公司 | The adaptive Unscented kalman filtering method of GPS and BDS dual mode satellite navigation system |
CN110501686A (en) * | 2019-09-19 | 2019-11-26 | 哈尔滨工程大学 | A kind of method for estimating state based on NEW ADAPTIVE high-order Unscented kalman filtering |
CN111506873A (en) * | 2020-04-13 | 2020-08-07 | 深圳埃瑞斯瓦特新能源有限公司 | Nonlinear system parameter estimation method based on extended Kalman filtering algorithm |
CN111985093A (en) * | 2020-08-03 | 2020-11-24 | 哈尔滨工程大学 | Adaptive unscented Kalman filtering state estimation method with noise estimator |
Non-Patent Citations (1)
Title |
---|
李恒;王光宇;秦健;: "无味卡尔曼滤波算法分析及其应用仿真研究", 兵工学报, no. 2 * |
Also Published As
Publication number | Publication date |
---|---|
CN113407909B (en) | 2024-01-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Sandmann et al. | Estimation of stochastic volatility models via Monte Carlo maximum likelihood | |
Rao | On the theory of bilinear time series models | |
Wan et al. | The unscented Kalman filter | |
Hansen et al. | Finite-sample properties of some alternative GMM estimators | |
Chuang et al. | Hybrid robust support vector machines for regression with outliers | |
Tao et al. | Large volatility matrix inference via combining low-frequency and high-frequency approaches | |
Diewert et al. | Frank Knight's theorem in linear programming revisited | |
Kultur et al. | Ensemble of neural networks with associative memory (ENNA) for estimating software development costs | |
Arize et al. | Import demand behavior in Africa: Some new evidence | |
Bodurtha et al. | Non-parametric estimation of an implied volatility surface | |
Fossà et al. | Multifidelity orbit uncertainty propagation using Taylor polynomials | |
Dufour et al. | Practical methods for modeling weak VARMA processes: identification, estimation and specification with a macroeconomic application | |
Noh | Posterior inference on parameters in a nonlinear DSGE model via Gaussian-based filters | |
CN113407909A (en) | Tasteless algorithm for non-analytic complex nonlinear system | |
Brankart et al. | Efficient local error parameterizations for square root or ensemble Kalman filters: application to a basin-scale ocean turbulent flow | |
Carriero et al. | No arbitrage priors, drifting volatilities, and the term structure of interest rates | |
Hansen et al. | Option pricing with time-varying volatility risk aversion | |
Klaus et al. | Evans function for Lax operators with algebraically decaying potentials | |
Cahyono | On the linear relation of signals | |
Jin et al. | Expectation-maximization algorithm based identification of Hammerstein nonlinear ARMAX systems with Gaussian mixture noises | |
Stangebye | Long-term sovereign debt: A steady state analysis | |
Takada | Simulated minimum Hellinger distance estimation of stochastic volatility models | |
Lim | Testing for the fundamental determinants of the long run real exchange rate | |
Andreasen et al. | The extended perturbation method: With applications to the New Keynesian model and the zero lower bound | |
Kim et al. | A software for correcting systematic biases in RCM input boundary conditions |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |