CN117216482A - Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy - Google Patents
Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy Download PDFInfo
- Publication number
- CN117216482A CN117216482A CN202311465494.6A CN202311465494A CN117216482A CN 117216482 A CN117216482 A CN 117216482A CN 202311465494 A CN202311465494 A CN 202311465494A CN 117216482 A CN117216482 A CN 117216482A
- Authority
- CN
- China
- Prior art keywords
- covariance
- representing
- kalman filter
- calculating
- sample
- 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
- 238000000034 method Methods 0.000 title claims abstract description 54
- 230000002708 enhancing effect Effects 0.000 title claims abstract description 19
- 239000011159 matrix material Substances 0.000 claims abstract description 25
- 238000012952 Resampling Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims description 20
- 238000005070 sampling Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 230000001902 propagating effect Effects 0.000 claims description 6
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 230000036039 immunity Effects 0.000 claims 1
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000012545 processing Methods 0.000 abstract description 2
- 238000001914 filtration Methods 0.000 description 9
- 238000012417 linear regression Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000011109 contamination Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Abstract
The invention relates to the technical field of electric digital processing, in particular to a method for enhancing the wild value interference resistance of a Kalman filter by means of an iteration strategy, which comprises the following steps: generating a volume sample; obtaining a prediction sample, and calculating a priori estimated mean value and a priori estimated covariance; generating a resampling status sample; obtaining an observed quantity prediction sample and predicting an observed quantity mean value and an observed quantity covariance; calculating a cross covariance between the state and the observation; obtaining a linearization observation model; calculating linearization error covariance, total observation noise covariance and normalized residual error; calculating a weight matrix, and carrying out weight weighting correction on the covariance of the total observed noise; calculating Kalman gain, posterior covariance and posterior state mean; updating the original observed noise covariance after re-weighting until the iteration update reaches the set times. The method provided by the invention is beneficial to improving the precision and the robustness of the Kalman filter and preventing the Kalman filter from diverging.
Description
Technical Field
The invention relates to the technical field of electric digital processing, in particular to a method for enhancing the wild value interference resistance of a Kalman filter by means of an iteration strategy.
Background
Since the unscented Kalman filtering technique was introduced into nonlinear state estimation, nonlinear filtering algorithms based on deterministic sampling have been developed rapidly, but the observed data provided by the sensors are susceptible to outlier contamination, causing the observed noise to appear as non-Gaussian noise, which in turn causes the filter estimation accuracy to drop or even diverge.
The mutual interference between the reflections of the target surface usually results in non-gaussian radar observations, and in order to cope with non-gaussian noise, various robust updating methods based on Huber functions are widely used in non-linear non-gaussian systems. In order to improve the robust filtering efficiency, iterative weighted least squares (Iteratively reweighted least-squares, IRLS) is added in the nonlinear filtering based on Huber, and Karlgaard and Schaub combine the Huber technology with the SPKF for the first time to cope with non-Gaussian noise. The prediction step based on kinematics and dynamics in the traditional SPKF is reserved, and the difference is that the update step based on an observation model is that specifically, a nonlinear observation model is converted into a linear model through a statistical linearization technology, and then a linear regression problem is solved; in order to weaken the influence of non-Gaussian noise on an estimation result, a Huber function is adopted to limit a normalized residual error, then a cost function is solved, and a posterior estimation value is obtained through iteration.
The iterative solution technique is an iterative weighted least square method, in each iteration, only the weight matrix is updated by using the updated posterior estimation value, and the transfer matrix in the linear regression model is fixed to an initial value, that is, the nonlinear observation model is linearized around the prior estimation value, which can lead to the reduction of the estimation accuracy of the filter. On the other hand, the Huber function commonly used at present does not have the "re-dropping" capability, so that the influence of a larger deviation wild value on the filter cannot be removed, and therefore the robustness of the Huber function type filter is not strong enough. In contrast, a gaussian kernel function based on the maximum correlation entropy criterion has a "re-descent" capability, which can make such a filter obtain stronger robustness, but because such a function does not appear as a convex function in the full domain, the filter is prone to be locally optimal, resulting in filter divergence.
Disclosure of Invention
The invention aims to solve the technical problem of providing a method for enhancing the wild value interference resistance of a Kalman filter by means of an iteration strategy, wherein a statistical linear regression technology is introduced into a nonlinear model to compensate linearization errors, and a re-weighted original observed noise covariance is added in a subsequent iteration process, so that the existing robust iteration strategy is improved, the Kalman filter can obtain higher estimation precision and consistency under non-Gaussian noise, the precision and the robustness of the Kalman filter are improved, and the Kalman filter divergence is prevented.
The invention is realized by the following technical scheme:
a method for enhancing the outlier interference resistance of a kalman filter by means of an iterative strategy, comprising the steps of:
s1: pair Kalman filterSampling the posterior estimation mean value and the posterior estimation covariance at the moment and generating a volume sample;
s2: propagating the volume sample through a process model to obtain a predicted sample, calculating a priori estimated mean value and a priori estimated covariance according to the predicted sample, and initializing a robust iterative updating process;
s3: resampling the Kalman filter estimation mean and covariance matrix to generate a resampling state sample;
s4: propagating the resampled state sample through a nonlinear observation model to obtain an observed quantity prediction sample, and predicting an observed quantity mean value and an observed quantity covariance according to the observed quantity prediction sample;
s5: calculating the cross covariance between the state and the observation according to the resampling state sample generated in the step S3 and the observed quantity prediction sample obtained in the step S4;
s6: carrying out statistical linearization treatment on the nonlinear observation model according to the observed quantity mean value and observed quantity covariance predicted in the step S4 and the cross covariance between the state and the observation calculated in the step S5 to obtain a linearization observation model;
s7: calculating linearization error covariance according to the linearization observation model, calculating total observation noise covariance according to the linearization error covariance, and calculating normalized residual according to the linearization observation model and the total observation noise covariance;
s8: calculating a weight and a weight matrix corresponding to the normalized residual error according to the normalized residual error and the Huber function, and weighting and correcting the covariance of the total observed noise according to the calculated weight matrix;
s9: calculating the Kalman gain of the Kalman filter according to the re-weighted corrected total observed noise covariance obtained in the step S8 and the linearization observed model obtained in the step S6, and calculating the posterior covariance and the posterior state mean value according to the Kalman gain of the Kalman filter;
s10: and (3) updating the weighted matrix obtained in the step (S8) to obtain the original observed noise covariance after re-weighting, judging whether the iterative updating reaches the set times, ending the iterative updating if the iterative updating reaches the set times, and switching to the step (S3) to continue the iterative updating if the iterative updating does not reach the set times.
Preferably, the Kalman filter is applied in step S1 according to equation (1)Sampling the posterior estimation mean value and the posterior estimation covariance at the moment, and resampling according to the formula (2) in the step S3 to obtain a resampling state sample:
（1）；
（2）；
wherein:representing a volume sample, +_>Represents sample number, ++>Representing construction matrix->Ordinal number of->Representing construction matrix->Is>Column vector,/->Representing Kalman filter->Time posterior estimated mean value ++>Representing Kalman filter->Time-of-day posterior estimation covariance, +.>Representing status dimension +.>Representing a resample status sample, +.>Representing Kalman filter->Time posterior estimated mean value ++>Representing Kalman filter->The posterior of time of day estimates covariance.
Further, the process model in step S2 is equation (3), the a priori estimated mean is calculated according to equation (4), and the a priori estimated covariance is calculated according to equation (5):
（3）；
（4）；
（5）；
wherein:representing prediction samples, ++>Representing a process model->Representing the a priori estimated mean value,representing a priori estimated covariance,/->Representing process noise covariance,/->Representing the volumetric sample weight.
Optimally, the nonlinear observation model in the step S4 is represented by a formula (6), and the observation mean value is predicted according to a formula (7)Predicting observed quantity covariance +.>：
（6）；
（7）；
（8）；
Wherein:predicting samples for observance +_>Representing a nonlinear observation model.
Further, in step S5, the cross-covariance between the state and the observation is calculated according to equation (9)：
（9）。
Preferably, the linear model in step S6 is of formula (10):
（10）；
wherein:representing the observation vector +.>Representing the true state vector, +.>Representing the observation matrix +.>Representing a constant matrix>Representing the total observed noise.
Further, in step S7, a linearization error covariance is calculated according to equation (11)Calculating the total observed noise covariance +.>Calculating normalized residual +.>：
（11）；
（12）；
（13）；
Wherein:representing the weighted modified original noise covariance.
Further, in step S8, the weight is calculated according to equation (14)Calculating a weight matrix according to formula (15)>Weight-corrected total observed noise covariance according to equation (16)>：
（14）；
（15）；
（16）；
Wherein:represents the regulatory parameters->Representing normalized residual->Is>A component.
Further, in step S9, a Kalman gain is calculated according to equation (17)：
（17）。
Further, in step S10, a posterior covariance is calculated according to equation (18)Calculating a posterior state mean value according to formula (19)：
（18）；
（19）。
Further, step S11 updates the re-weighted original observed noise covariance according to (20)：
（20）。
The invention has the beneficial effects that:
according to the method for enhancing the outlier interference resistance of the Kalman filter by means of the iteration strategy, the statistical linear regression technology is introduced into the nonlinear model approximation to compensate linearization errors, and the original observed noise covariance after re-weighting is added in the subsequent iteration process to improve the iteration strategy, so that the iteration strategy does not have significant influence on the Kalman filter under Gaussian noise, the Kalman filter can obtain higher estimation precision and consistency under non-Gaussian noise, and along with the increasing of disturbance probability, the advantage is more and more obvious, the iteration strategy can improve the filtering precision and robustness of the Kalman filter under the premise of not bringing heavy calculation load, and the method is more suitable for robust state estimation under non-Gaussian noise.
Drawings
FIG. 1 is a schematic flow chart of the method of the invention.
Detailed Description
The invention relates to a method for enhancing the wild value interference resistance of a Kalman filter by means of an iteration strategy, a specific flow chart is shown in figure 1, and the method comprises the following steps:
s1: pair Kalman filterSampling the posterior estimation mean value and the posterior estimation covariance at the moment and generating a volume sample;
s2: propagating the volume sample through a process model to obtain a predicted sample, calculating a priori estimated mean value and a priori estimated covariance according to the predicted sample, and initializing a robust iterative updating process;
s3: resampling the Kalman filter estimation mean and covariance matrix to generate a resampling state sample;
s4: propagating the resampled state sample through a nonlinear observation model to obtain an observed quantity prediction sample, and predicting an observed quantity mean value and an observed quantity covariance according to the observed quantity prediction sample;
s5: calculating the cross covariance between the state and the observation according to the resampling state sample generated in the step S3 and the observed quantity prediction sample obtained in the step S4;
s6: carrying out statistical linearization treatment on the nonlinear observation model according to the observed quantity mean value and observed quantity covariance predicted in the step S4 and the cross covariance between the state and the observation calculated in the step S5 to obtain a linearization observation model; thus, a statistical linear regression technology is introduced into nonlinear model approximation, and linearization errors can be compensated;
s7: calculating linearization error covariance according to the linearization observation model, calculating total observation noise covariance according to the linearization error covariance, and calculating normalized residual according to the linearization observation model and the total observation noise covariance;
s8: calculating a weight and a weight matrix corresponding to the normalized residual error according to the normalized residual error and the Huber function, and weighting and correcting the covariance of the total observed noise according to the calculated weight matrix;
s9: calculating the Kalman gain of the Kalman filter according to the re-weighted corrected total observed noise covariance obtained in the step S8 and the linearization observed model obtained in the step S6, and calculating the posterior covariance and the posterior state mean value according to the Kalman gain of the Kalman filter;
s10: and (3) updating the weighted matrix obtained in the step (S8) to obtain the original observed noise covariance after re-weighting, judging whether the iterative updating reaches the set times, ending the iterative updating if the iterative updating reaches the set times, and switching to the step (S3) to continue the iterative updating if the iterative updating does not reach the set times.
The heavy tail observation noise can cause the navigation accuracy to be obviously reduced, and although the widely applied Huber-based M estimation can cope with non-Gaussian noise, the adopted iteration strategy cannot utilize the posterior estimation value to re-linearize the observation model, which can cause the filtering accuracy to be reduced. In consideration of the defect of research on the iterative robust strategy, the invention deduces an iterative robust strategy based on optimization. In order to approximate the nonlinear equation more accurately, a linearization method based on a statistical linear regression theory is introduced in the gauss-newton method, thereby compensating uncertainty caused by approximation errors in observed noise covariance. In addition, the invention improves iteration and robustness strategies by means of a new cost function, namely, the corrected covariance matrix in the last iteration is incorporated again in the next iteration loop, so that the filter can be helped to obtain higher filtering precision, stronger robustness and better consistency.
A statistical linear regression technology is introduced into nonlinear model approximation to compensate linearization errors, the method is embodied in step S6, and the original observed noise covariance after re-weighting is added in the subsequent iteration process to improve an iteration strategy, the iteration strategy does not have obvious influence on a Kalman filter under Gaussian noise, and under non-Gaussian noise, the Kalman filter can obtain higher estimation precision and consistency, and the method is embodied in step S8 and step S10 along with the increasing of disturbance probability.
The iterative strategy can improve the filtering efficiency on the premise of not bringing heavy calculation load, and is more suitable for robust state estimation under non-Gaussian noise.
In step S1, the Kalman filter is subjected to the method of formula (1)Sampling the posterior estimation mean value and the posterior estimation covariance at the moment, and resampling according to the formula (2) in the step S3 to obtain a resampling state sample:
（1）；
（2）；
wherein:representing a volume sample, +_>Represents sample number, ++>Representing construction matrix->Ordinal number of->Representing construction matrix->Is>Column vector,/->，/>Representing Kalman filter->Time posterior estimated mean value ++>Representing Kalman filter->Time-of-day posterior estimation covariance, +.>Representing status dimension +.>Representing a resample status sample, +.>Representing KalmanFilter->Time posterior estimated mean value ++>Representing Kalman filter->The posterior of time of day estimates covariance.
Further, the process model in step S2 is equation (3), the a priori estimated mean is calculated according to equation (4), and the a priori estimated covariance is calculated according to equation (5):
（3）；
（4）；
（5）；
wherein:representing prediction samples, ++>Representing a process model->Representing the a priori estimated mean value,representing a priori estimated covariance,/->Representing process noise covariance,/->Representing the volumetric sample weight.
Optimally, the nonlinear observation model in the step S4 is represented by a formula (6), and the observation mean value is predicted according to a formula (7)Predicting observed quantity covariance +.>：
（6）；
（7）；
（8）；
Wherein:predicting samples for observance +_>Representing a nonlinear observation model.
Further, in step S5, the cross-covariance between the state and the observation is calculated according to equation (9)：
（9）。
Preferably, the linear model in step S6 is of formula (10):
（10）；
wherein:representing the observation vector +.>Representing the true state vector, +.>Representing the observation matrix of the image of the object,，/>representing a constant matrix>，/>Representing the total observed noise of the device,，/>representing linearization errors, +.>Representing the original observed noise.
Further, in step S7, a linearization error covariance is calculated according to equation (11)Calculating the total observed noise covariance +.>Calculating normalized residual +.>：
（11）；
（12）；
（13）；
Wherein:representing the weighted modified original noise covariance.
Further, in step S8, the weight is calculated according to equation (14)Calculating a weight matrix according to formula (15)>Weight-corrected total observed noise covariance according to equation (16)>：
（14）；
（15）；
（16）；
Wherein:represents the regulatory parameters->Representing normalized residual->Is>A component.
Further, in step S9, a Kalman gain is calculated according to equation (17)In step S10, a posterior covariance is calculated according to equation (18)>Calculating posterior state mean +.>：
（17）；
（18）；
（19）。
Further, step S10 updates the re-weighted original observed noise covariance as shown in FIG. 20：
（20）。
Through the specific implementation mode, the original observed noise covariance improvement iteration strategy after re-weighting is added in the iteration process, so that the Kalman filter can obtain higher estimation accuracy and consistency under non-Gaussian noise, and the advantages are more and more obvious along with the increase of disturbance probability. The specific iteration mode provided by the invention does not bring heavy calculation load, is superior to the traditional fixed point iteration strategy, is beneficial to improving the filtering precision, estimation consistency and robustness of the Kalman filter, and prevents the Kalman filter from diverging.
In summary, the method for enhancing the outlier interference resistance of the Kalman filter by means of the iteration strategy introduces a statistical linear regression technology into the nonlinear model to compensate linearization errors, and adds the original observed noise covariance after re-weighting in the subsequent iteration process to improve the existing iteration strategy, so that the Kalman filter can obtain higher estimation precision and consistency under non-Gaussian noise, the precision of the Kalman filter is improved, and the Kalman filter divergence is prevented.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. The method for enhancing the wild value interference resistance of the Kalman filter by means of the iterative strategy is characterized by comprising the following steps of: the method comprises the following steps:
s1: pair Kalman filterSampling the posterior estimation mean value and the posterior estimation covariance at the moment and generating a volume sample;
s2: propagating the volume sample through a process model to obtain a predicted sample, calculating a priori estimated mean value and a priori estimated covariance according to the predicted sample, and initializing a robust iterative updating process;
s3: resampling the Kalman filter estimation mean and covariance matrix to generate a resampling state sample;
s4: propagating the resampled state sample through a nonlinear observation model to obtain an observed quantity prediction sample, and predicting an observed quantity mean value and an observed quantity covariance according to the observed quantity prediction sample;
s5: calculating the cross covariance between the state and the observation according to the resampling state sample generated in the step S3 and the observed quantity prediction sample obtained in the step S4;
s6: carrying out statistical linearization treatment on the nonlinear observation model according to the observed quantity mean value and observed quantity covariance predicted in the step S4 and the cross covariance between the state and the observation calculated in the step S5 to obtain a linearization observation model;
s7: calculating linearization error covariance according to the linearization observation model, calculating total observation noise covariance according to the linearization error covariance, and calculating normalized residual according to the linearization observation model and the total observation noise covariance;
s8: calculating a weight and a weight matrix corresponding to the normalized residual error according to the normalized residual error and the Huber function, and weighting and correcting the covariance of the total observed noise according to the calculated weight matrix;
s9: calculating the Kalman gain of the Kalman filter according to the re-weighted corrected total observed noise covariance obtained in the step S8 and the linearization observed model obtained in the step S6, and calculating the posterior covariance and the posterior state mean value according to the Kalman gain of the Kalman filter;
s10: and (3) updating the weighted matrix obtained in the step (S8) to obtain the original observed noise covariance after re-weighting, judging whether the iterative updating reaches the set times, ending the iterative updating if the iterative updating reaches the set times, and switching to the step (S3) to continue the iterative updating if the iterative updating does not reach the set times.
2. The method for enhancing the outlier interference resistance of a kalman filter by means of an iterative strategy according to claim 1, characterized in that: in step S1, the Kalman filter is subjected to the method of formula (1)Sampling the posterior estimation mean value and the posterior estimation covariance at the moment, and resampling according to the formula (2) in the step S3 to obtain a resampling state sample:
（1）；
（2）；
wherein:representing a volume sample, +_>Represents sample number, ++>Representing construction matrix->Ordinal number of->Representing construction matrix->Is>Column vector,/->Representing Kalman filter->Time posterior estimated mean value ++>Representing Kalman filter->Time-of-day posterior estimation covariance, +.>Representing status dimension +.>Representing a resample status sample, +.>Representing Kalman filter->Time posterior estimated mean value ++>Representing Kalman filter->The posterior of time of day estimates covariance.
3. The method for enhancing the outlier interference resistance of a kalman filter by means of an iterative strategy according to claim 2, characterized in that: the process model in step S2 is equation (3), the prior estimation mean is calculated according to equation (4), and the prior estimation covariance is calculated according to equation (5):
（3）；
（4）；
（5）；
wherein:representing prediction samples, ++>Representation process modelIs (1)>Representing a priori estimated mean,/->Representing a priori estimated covariance,/->Representing process noise covariance,/->Representing the volumetric sample weight.
4. A method for enhancing the immunity of a kalman filter to outlier interference by means of an iterative strategy according to claim 3, characterized in that: the nonlinear observation model in step S4 is represented by formula (6), and the observed quantity average value is predicted according to formula (7)Predicting observed quantity covariance +.>：
（6）；
（7）；
（8）；
Wherein:predicting samples for observance +_>Representing a nonlinear observation model.
5. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 4, wherein: in step S5, the cross-covariance between the state and the observations is calculated according to equation (9)：
（9）。
6. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 5, wherein: the linear model in step S6 is of formula (10):
（10）；
wherein:representing the observation vector +.>Representing the true state vector, +.>Representing the observation matrix +.>A constant matrix is represented and is used to represent,representing the total observed noise.
7. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 6, wherein: in step S7, a linearization error covariance is calculated according to equation (11)Calculating the total observed noise covariance +.>Calculating normalized residual +.>：
（11）；
（12）；
（13）；
Wherein:representing the weighted modified original noise covariance.
8. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 7, wherein: in step S8, the weight is calculated according to equation (14)Calculating a weight matrix according to formula (15)>Root of Chinese characterAccording to (16) weight correction total observed noise covariance +.>：
（14）；
（15）；
（16）；
Wherein:represents the regulatory parameters->Representing normalized residual->Is>A component.
9. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 8, wherein: in step S9, kalman gain is calculated according to equation (17)In step S10, a posterior covariance is calculated according to equation (18)>Calculating posterior state mean +.>：
（17）；
（18）；
（19）。
10. The method for enhancing the antioutlier interference capability of a kalman filter by means of an iterative strategy according to claim 9, wherein: step S10 updating the re-weighted original observed noise covariance according to (20)：
（20）。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311465494.6A CN117216482B (en) | 2023-11-07 | 2023-11-07 | Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311465494.6A CN117216482B (en) | 2023-11-07 | 2023-11-07 | Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117216482A true CN117216482A (en) | 2023-12-12 |
| CN117216482B CN117216482B (en) | 2024-03-01 |
Family
ID=89041221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311465494.6A Active CN117216482B (en) | 2023-11-07 | 2023-11-07 | Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117216482B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106487358A (en) * | 2016-09-30 | 2017-03-08 | 西南大学 | A kind of maximal correlation entropy volume kalman filter method based on statistical linear regression |
| CN107547067A (en) * | 2017-09-15 | 2018-01-05 | 北京航空航天大学 | A kind of multi-model self calibration EKF method |
| CN109388778A (en) * | 2018-09-11 | 2019-02-26 | 东南大学 | A kind of iteration volume point Unscented kalman filtering method |
| CN109459705A (en) * | 2018-10-24 | 2019-03-12 | 江苏理工学院 | A kind of power battery SOC estimation method of anti-outlier robust Unscented kalman filtering |
| CN114124033A (en) * | 2021-10-11 | 2022-03-01 | 北京川速微波科技有限公司 | Kalman filter implementation method, device, storage medium and equipment |
-
2023
- 2023-11-07 CN CN202311465494.6A patent/CN117216482B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106487358A (en) * | 2016-09-30 | 2017-03-08 | 西南大学 | A kind of maximal correlation entropy volume kalman filter method based on statistical linear regression |
| CN107547067A (en) * | 2017-09-15 | 2018-01-05 | 北京航空航天大学 | A kind of multi-model self calibration EKF method |
| CN109388778A (en) * | 2018-09-11 | 2019-02-26 | 东南大学 | A kind of iteration volume point Unscented kalman filtering method |
| CN109459705A (en) * | 2018-10-24 | 2019-03-12 | 江苏理工学院 | A kind of power battery SOC estimation method of anti-outlier robust Unscented kalman filtering |
| CN114124033A (en) * | 2021-10-11 | 2022-03-01 | 北京川速微波科技有限公司 | Kalman filter implementation method, device, storage medium and equipment |
Non-Patent Citations (3)
| Title |
|---|
| SHOUPENGLI: ""Optimization-based iterative and robust strategy for spacecraft relative navigation in elliptical orbit"", 《AEROSPACE SCIENCE AND TECHNOLOGY》 * |
| 姜浩楠;蔡远利;: "鲁棒高斯和集合卡尔曼滤波及其在纯角度跟踪中的应用", 控制理论与应用, no. 02 * |
| 秦康;董新民;陈勇;: "基于Huber的鲁棒高阶容积卡尔曼滤波算法", 计算机工程与应用, no. 07 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117216482B (en) | 2024-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108759833B (en) | Intelligent vehicle positioning method based on prior map | |
| CN111178385B (en) | Target tracking method for robust online multi-sensor fusion | |
| CN110702095A (en) | Data-driven high-precision integrated navigation data fusion method | |
| CN110443832B (en) | Evidence filtering target tracking method based on observation interval value | |
| CN107994885B (en) | Distributed fusion filtering method for simultaneously estimating unknown input and state | |
| CN112432644B (en) | Unmanned ship integrated navigation method based on robust adaptive unscented Kalman filtering | |
| WO2020052213A1 (en) | Iterative cubature unscented kalman filtering method | |
| CN115981148A (en) | Unmanned aerial vehicle ground moving target tracking method | |
| CN109582916A (en) | A kind of method for adaptive kalman filtering based on observation noise Unknown Variance | |
| CN112083457A (en) | IMM satellite positioning and navigation method based on neural network optimization | |
| CN117216482B (en) | Method for enhancing wild value interference resistance of Kalman filter by means of iterative strategy | |
| CN115457152A (en) | External parameter calibration method and device, electronic equipment and storage medium | |
| CN110702093B (en) | Particle filter-based positioning method and device, storage medium and robot | |
| CN108846427A (en) | The single out-of-sequence measurement central fusion method of any delay step of nonlinear system | |
| CN113191427B (en) | Multi-target vehicle tracking method and related device | |
| CN108680162B (en) | Human body target tracking method based on progressive unscented Kalman filtering | |
| CN109253727B (en) | Positioning method based on improved iteration volume particle filter algorithm | |
| CN113960645A (en) | High-precision satellite navigation positioning resolving method based on Doppler frequency assisted smoothing | |
| CN110912535B (en) | Novel non-pilot Kalman filtering method | |
| CN113759364A (en) | Millimeter wave radar continuous positioning method and device based on laser map | |
| CN111274529B (en) | Robust Gao Sini Weisal PHD multi-expansion target tracking algorithm | |
| CN107966697B (en) | Moving target tracking method based on progressive unscented Kalman | |
| CN116482579A (en) | Robust noise amount self-correction electric connector residual life prediction method | |
| Bhaumik et al. | Risk-sensitive formulation of unscented Kalman filter | |
| CN106814608B (en) | Predictive control adaptive filtering algorithm based on posterior probability distribution |
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 |