CN114088086A - Multi-target robust positioning method for resisting measurement outlier interference - Google Patents

Multi-target robust positioning method for resisting measurement outlier interference Download PDF

Info

Publication number
CN114088086A
CN114088086A CN202111402405.4A CN202111402405A CN114088086A CN 114088086 A CN114088086 A CN 114088086A CN 202111402405 A CN202111402405 A CN 202111402405A CN 114088086 A CN114088086 A CN 114088086A
Authority
CN
China
Prior art keywords
target
sensor
measurement
robust
outlier
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
Application number
CN202111402405.4A
Other languages
Chinese (zh)
Other versions
CN114088086B (en
Inventor
罗成名
杨旭东
王璐雪
王彪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu University of Science and Technology
Original Assignee
Jiangsu University of Science and Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu University of Science and Technology filed Critical Jiangsu University of Science and Technology
Priority to CN202111402405.4A priority Critical patent/CN114088086B/en
Publication of CN114088086A publication Critical patent/CN114088086A/en
Application granted granted Critical
Publication of CN114088086B publication Critical patent/CN114088086B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20—Instruments for performing navigational calculations
    • 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
    • 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
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques

Abstract

The invention discloses a multi-target robust positioning method for resisting measurement outlier interference, which comprises the following steps: establishing a multi-target state model and an observation model according to a motion model of a target and distance difference information between the target and a sensor pair acquired by a sensor; detecting the observed value based on the prior information of the sensor to find a measurement wild value; introducing virtual points, an association gate and a joint probability to correct the detected measurement field value; deducing an unscented particle filter model with robust characteristics by using the corrected measured value; and the robust unscented particle filtering results of multiple sensors are fused to realize multi-target positioning. The method weakens the influence of the measured outlier and particle degradation on the target positioning result, weakens the influence of the poorer filtering result on the final positioning result, improves the accuracy of multi-target positioning, and can provide a certain basis for the multi-target positioning.

Description

Multi-target robust positioning method for resisting measurement outlier interference
Technical Field
The invention belongs to the technical field of mobile target positioning by utilizing multiple sensors, and particularly relates to a multi-target robust positioning method for resisting measurement outlier interference.
Background
The nonlinear filtering problem is a hot spot problem in signal processing and control theory. It has wide applications in many fields such as radar localization, signal processing, mobile robotics, and navigation, and in the fields of pattern recognition and image processing, filtering algorithms are successfully applied to problems of image matching, image segmentation, skeletonization of images, and contour extraction.
In practical applications, most systems are non-linear and non-gaussian, and for such systems, kalman filtering will fail. With the diversification of the actual state model and the environment, many scholars gradually start to research filtering algorithms capable of adapting to complex environments. In order to solve the problem, a method for approximating a nonlinear state space model by a Kalman filter is provided, namely Taylor series expansion is used for replacing a state transition equation and a measurement equation, but for a strong nonlinear system, the method brings larger truncation errors, and meanwhile, the processing of a Jacobian matrix is a complex calculation process. On the basis, an unscented Kalman filtering algorithm is provided, the mean value and the covariance are calculated by using a plurality of sigma points in a recursion mode, but the unscented Kalman filtering algorithm still can only use normal distribution to approach the real posterior probability. A particle filtering algorithm is now proposed based on sequential importance sampling, combining resampling techniques with monte carlo importance sampling. The algorithm is an optimal regression algorithm, combines Monte Carlo thought and recursive Bayesian filtering, and has good estimation effect when processing nonlinear and non-Gaussian systems.
The prior art has the following disadvantages:
1. the complex target motion environment can cause the wild value of the measurement sequence at a certain moment, which can cause the effect of particle filtering to be poor and has the problem of low positioning precision.
2. The conventional particle filtering cannot well select an optimal established density function to guide the particles to resample, and the problem of particle degradation exists.
3. Meanwhile, in an actual environment, more than one sensor is arranged, the positioning results obtained by filtering the measured values with different accuracies are different in accuracy, and if the final positioning result is deteriorated by simply summing and averaging the measured values, the problem that data fusion cannot be effectively carried out exists.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a multi-target robust positioning method for resisting measurement outlier interference, which solves the problem that the positioning result is influenced by the existing measurement outlier, particle degradation and poor filtering result.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a multi-target robust positioning method for resisting measurement outlier interference comprises the following specific steps: establishing a multi-target state model and an observation model according to a motion model of a target and distance difference information between the target and a sensor pair acquired by a sensor; detecting the observed value based on the prior information of the sensor to find a measurement wild value; introducing virtual points, an association gate and a joint probability to correct the detected measurement field value; deducing an unscented particle filter model with robust characteristics by using the corrected measured value; and the robust unscented particle filtering results of the multiple sensors are fused to realize multi-target positioning.
Further, the content and the method for establishing the state model and the observation model of the multiple targets according to the motion model of the targets and the distance difference information between the targets and the sensor pairs acquired by the sensors specifically comprise the following steps:
(1) multiple target states of
Figure BDA0003368959420000021
Wherein mxj(k)、myj(k) And mzj(k) Respectively represent the position coordinates of the moving object,
Figure BDA0003368959420000022
and
Figure BDA0003368959420000023
representing the X-axis, y-axis and z-axis velocities of the moving object, respectively, so that the state model of the multiple objects can be expressed as Xj(k)＝FjXj(k-1)+Γjuj(k-1)，uj(k-1) Process noise in accordance with Gaussian distribution, FjBeing a state transition matrix, ΓjIs a state noise matrix;
(2) moving object mj(k)＝[mxj(k),myj(k),mzj(k)]TAnd a sensor ai＝[axi,ayi,azi]TThe distance relationship between them is
Figure BDA0003368959420000024
Where | · | | represents a two-norm, moving target mj(k) And a sensor aiAnd a1The distance difference between is expressed as
Figure BDA0003368959420000025
Based on distance differences
Figure BDA0003368959420000026
And measuring the noise vj(k) Can be expressed as Zj(k)＝H(Xj(k))+vj(k) Wherein
Figure BDA0003368959420000027
Further, the method for detecting the observed value based on the prior information of the sensor to find the content and the method of the measurement outlier specifically includes the following steps:
(1) when moving object mj(k) And a sensor aiGeometric distance between
Figure BDA0003368959420000028
Less than the communication radius R of the sensorcThe mobile node can sense the wireless signal of the sensing node;
(2) because the measured data is distance difference
Figure BDA0003368959420000029
Therefore, the distance difference relationship between the moving target and the sensor pair is used to judge whether the field value exists in the measurement sequence, namely when
Figure BDA00033689594200000210
Considering that the observed value has no outlier; otherwise, the measurement is present in the outlier.
Further, the content and method for introducing the virtual observation point, the association gate and the joint probability to correct the detected measurement outlier specifically include the following steps:
(1) the associated gate of an ellipsoid is equivalent to a sphere, the radius of which is expressed as
Figure BDA00033689594200000211
Where V is the volume of an ellipsoid, so that the observed value is predicted
Figure BDA0003368959420000031
Generating virtual observation points near the center that follow a Gaussian distribution with mean l/2 and variance l
Figure BDA0003368959420000032
Wherein N ispThe number of the virtual observation points is;
(2) the innovation of a virtual observation point can be represented as
Figure BDA0003368959420000033
According to the set threshold value g and the covariance of the predicted innovation
Figure BDA0003368959420000034
To screen effective virtual observation points when
Figure BDA0003368959420000035
When the number is less than g, the virtual observation point is an effective virtual observation point; otherwise, the virtual observation point is eliminated, and the effective virtual observation point is represented as
Figure BDA0003368959420000036
Wherein N isvThe number of effective virtual observation points is;
(3) joint probability beta of valid virtual observation pointsm2As its weight, finally according to
Figure BDA0003368959420000037
And correcting the observed value with the outlier.
Further, the content and method for deriving the unscented particle filter model with robust characteristics by using the corrected measurement values specifically include the following steps: mean values of particles obtained from unscented Kalman filtering
Figure BDA0003368959420000038
And corresponding covariance
Figure BDA0003368959420000039
Constructing a Gaussian suggestion density function as an importance density function in particle filtering to guide the resampling of the particles, wherein the expression is
Figure BDA00033689594200000310
To prevent particle degradation, particle filtering is then aggregated to obtain a minimum variance estimate of the state
Figure BDA00033689594200000311
Further, the content and the method for realizing multi-target positioning by fusing the robust unscented particle filter result of the multi-sensor specifically comprise the following steps:
(1) measuring field value correction method is carried out on the measured data obtained by all the sensors, and then robust unscented particle filtering is carried out by using the corrected measured values to obtain all the filtering result sets
Figure BDA00033689594200000312
Simultaneously collecting corresponding covariance sets
Figure BDA00033689594200000313
(2) According to
Figure BDA00033689594200000314
All covariance is normalized, wherein |, refers to Hadamard product, and then normalized covariance is used
Figure BDA00033689594200000315
And corresponding filtering results
Figure BDA00033689594200000316
Carrying out weighted summation to obtain the final filtering result of the moving target
Figure BDA00033689594200000317
Will be provided with
Figure BDA00033689594200000318
As a result of the positioning of the moving object
Figure BDA00033689594200000319
Namely, it is
Figure BDA00033689594200000320
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. when disturbance is added to a measurement sequence at a certain moment, the weight distribution of the particles is disturbed, and the disturbance with different sizes can generate different weight distributions on the particles, so that the weights of the poorer particles can be larger.
2. The problem of particle degradation exists in the common particle filter, and the problem of particle degradation is weakened by using the Gaussian suggested density obtained by the unscented Kalman filter as an importance density function in the particle filter.
3. And meanwhile, data fusion is carried out on all final filtering results, so that the problem that the multi-target positioning is influenced by the poor filtering results is solved.
Drawings
FIG. 1 is a block diagram of a multi-target robust positioning process for resisting measurement outlier interference according to the present invention.
FIG. 2 is a detailed step diagram of multi-target robust positioning for resisting measurement outlier interference according to the present invention.
FIG. 3 is a schematic diagram of a sensor deployment configuration used in the present invention.
FIG. 4 is a diagram of a multi-target robust positioning error resisting the interference of measurement outliers in accordance with the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments.
The invention is described in detail with reference to the accompanying fig. 1-4.
As shown in fig. 1, a multi-target robust positioning method for resisting measurement outlier interference includes the steps: establishing a multi-target state model and an observation model according to a motion model of a target and distance difference information between the target and a sensor pair acquired by a sensor; detecting the observed value based on the prior information of the sensor to find a measurement wild value; introducing a virtual observation point, an association gate and a joint probability to correct the detected measurement outlier; deducing an unscented particle filter model with robust characteristics by using the corrected measured value; and the robust unscented particle filtering results of multiple sensors are fused to realize multi-target positioning.
As shown in fig. 2, the content and method for establishing the state model and the observation model of the multiple targets specifically include the following steps:
(1) multiple target states of
Figure BDA0003368959420000041
Multiple targetsIs represented as Xj(k)＝FjXj(k-1)+Γjuj(k-1)，uj(k-1) Process noise in accordance with Gaussian distribution, FjBeing a state transition matrix, ΓjIs a state noise matrix;
(2) moving object mj(k)＝[mxj(k),myj(k),mzj(k)]TAnd a sensor ai＝[axi,ayi,azi]TThe distance relationship between them is
Figure BDA0003368959420000042
Where | · | | represents a two-norm, moving target mj(k) And a sensor aiAnd a1Distance difference between
Figure BDA0003368959420000043
Based on distance differences
Figure BDA0003368959420000044
And measuring the noise vj(k) Can be expressed as Zj(k)＝H(Xj(k))+vj(k) Wherein
Figure BDA0003368959420000045
The method for detecting the observed value based on the prior information of the sensor to find the content and the method of the measured outlier specifically comprises the following steps:
(1) when moving object mj(k) And a sensor aiGeometric distance between
Figure BDA0003368959420000051
Smaller than communication radius sensor RcThe mobile node can sense the wireless signal of the sensing node;
(2) because the measured data are
Figure BDA0003368959420000052
So that the distance difference between the moving object and the sensor pair is usedIs used to determine whether the measured sequence has a outlier, i.e., when
Figure BDA0003368959420000053
Considering that the observed value has no outlier; otherwise, the measurement has a outlier.
The content and the method for correcting the observation outlier specifically comprise the following steps:
(1) the associated gate of an ellipsoid is equivalent to a sphere, the radius of which is expressed as
Figure BDA0003368959420000054
Where V is the volume of an ellipsoid, so that the observed value is predicted
Figure BDA0003368959420000055
Generating virtual observation points near the center that follow a Gaussian distribution with mean l/2 and variance l
Figure BDA0003368959420000056
Wherein N ispThe number of the virtual observation points is;
(2) the innovation of a virtual observation point can be represented as
Figure BDA0003368959420000057
According to the set threshold value g and the covariance of the predicted innovation
Figure BDA0003368959420000058
To screen the effective virtual observation points
Figure BDA0003368959420000059
When the number is less than g, the virtual observation point is an effective virtual observation point; otherwise, the virtual observation point is eliminated, and the effective virtual observation point is represented as
Figure BDA00033689594200000510
Wherein N isvThe number of effective virtual observation points is;
(3) joint probability beta of valid virtual observation pointsm2As its weightFinally according to
Figure BDA00033689594200000511
And correcting the observed value with the outlier.
The content and the method for deducing the unscented particle filter model with the robust characteristic by using the corrected measured value specifically comprise the following steps:
(1) forming a new observation set Z from the corrected observations of claim 4j(k) Binding particles
Figure BDA00033689594200000512
Sum covariance
Figure BDA00033689594200000513
To construct a scaled Sigma point set, and the mean weight and covariance weight corresponding to each Sigma point are represented as WmAnd WcThe Sigma point set is predicted in one step according to the state equation, and the particle prediction mean value is calculated according to the predicted Sigma point set
Figure BDA00033689594200000514
And prediction covariance
Figure BDA00033689594200000515
(2) Reconstructing a Sigma point set based on the predicted mean and covariance in step (5.1)
Figure BDA00033689594200000516
Then, the prediction observed quantity is calculated according to the measurement model
Figure BDA00033689594200000517
And according to WmAnd
Figure BDA00033689594200000518
calculating the observed mean
Figure BDA00033689594200000519
Calculating to obtain an autocovariance matrix PZ(k),Z(k)And cross covariance matrix PX(k),Z(k)While the update formula of the Kalman gain is expressed as
Figure BDA0003368959420000061
Then updating the mean value of the particles
Figure BDA0003368959420000062
And its covariance
Figure BDA0003368959420000063
(3) According to
Figure BDA0003368959420000064
And
Figure BDA0003368959420000065
the Gaussian suggestion density function is constructed as an importance density function in particle filtering, and the expression is
Figure BDA0003368959420000066
Particle filtering is then aggregated to obtain a minimum variance estimate of the state.
The content and the method for realizing multi-target positioning by combining the robust unscented particle filtering results of the multiple sensors specifically comprise the following steps:
(1) using the corrected measurements for the robust unscented particle filtering of claim 5 to obtain all sets of filter results
Figure BDA0003368959420000067
Then, corresponding covariance sets are collected simultaneously
Figure BDA0003368959420000068
(2) According to
Figure BDA0003368959420000069
All covariance values are normalized, wherein [ ] refers to Hadamard product, and then the normalized values are usedCovariance
Figure BDA00033689594200000610
And corresponding filtering results
Figure BDA00033689594200000611
Carrying out weighted summation to obtain the final filtering result of the moving target
Figure BDA00033689594200000612
Will be provided with
Figure BDA00033689594200000613
As a result of the positioning of the moving object
Figure BDA00033689594200000614
Namely, it is
Figure BDA00033689594200000615
As shown in fig. 3, sensor a1、a2、a3And a4Respectively arranged at four corners of a two-dimensional positioning space, and the movable trolley moves along the dotted line in the figure in the positioning area, and the sensor a1The other 3 sensors are base stations and obtain distance difference information corresponding to different moments in the moving process of the moving trolley.
As shown in fig. 4, the positioning error of the moving object in the present invention within 100 seconds of sampling time is described when the moving object performs a uniform motion.
The invention discloses a multi-target robust positioning method for resisting measurement outlier interference, which deploys and coordinates a sensor according to the geometric dimension of a positioning area, and establishes a state model and an observation model based on a motion model of a target and a distance difference between the sensor and a moving target; detecting and distinguishing the measurement field value by using the prior information of the sensor; then correcting the measured outlier; the unscented particle filter with the robust characteristic is fused for multi-target positioning, and the influence of a poor filtering result on the positioning is weakened. The invention can correct the measurement outlier, and can combine unscented Kalman filtering and particle filtering to obtain a better multi-target positioning result, thereby providing stable service for multi-target positioning.
The foregoing shows and describes the general principles and broad features of the present invention and advantages thereof. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are described in the specification and illustrated only to illustrate the principle of the present invention, but that various changes and modifications may be made therein without departing from the spirit and scope of the present invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims (6)

1. A multi-target robust positioning method for resisting measurement outlier interference is characterized in that: establishing a multi-target state model and an observation model according to a motion model of a target and distance difference information between the target and a sensor pair acquired by a sensor; detecting the observed value based on the prior information of the sensor to find a measurement wild value; introducing virtual points, an association gate and a joint probability to correct the detected measurement field value; deducing an unscented particle filter model with robust characteristics by using the corrected measured value; and the robust unscented particle filtering results of the multiple sensors are fused to realize multi-target positioning.
2. The multi-target robust positioning method for resisting measurement outlier interference according to claim 1, wherein the method comprises the following steps: the content and the method for establishing the state model and the observation model of the multiple targets according to the motion model of the targets and the distance difference information between the targets and the sensor pairs acquired by the sensors specifically comprise the following steps:
(1) multiple target states of
Figure FDA0003368959410000011
Wherein mxj(k)、myj(k) And mzj(k) Respectively represent the position coordinates of the moving object,
Figure FDA0003368959410000012
and
Figure FDA0003368959410000013
representing the X-axis, y-axis and z-axis velocities of the moving object, respectively, so that the state model of the multiple objects can be expressed as Xj(k)＝FjXj(k-1)+Γjuj(k-1)，uj(k-1) Process noise in accordance with Gaussian distribution, FjBeing a state transition matrix, ΓjIs a state noise matrix;
(2) moving object mj(k)＝[mxj(k),myj(k),mzj(k)]TAnd a sensor ai＝[axi,ayi,azi]TThe distance relationship between them is
Figure FDA0003368959410000014
Where | · | | represents a two-norm, moving target mj(k) And a sensor aiAnd a1The distance difference between is expressed as
Figure FDA0003368959410000015
Based on distance differences
Figure FDA0003368959410000016
And measuring the noise vj(k) Can be expressed as Zj(k)＝H(Xj(k))+vj(k) In which
Figure FDA0003368959410000017
3. The multi-target robust positioning method for resisting measurement outlier interference according to claim 1, which is characterized in that: the method for detecting the observed value based on the prior information of the sensor to find the content and the method of the measurement wild value specifically comprises the following steps:
(1) when moving an objectmj(k) And a sensor aiGeometric distance d betweeni jLess than the communication radius R of the sensorcThe mobile node can sense the wireless signal of the sensing node;
(2) because the measured data is distance difference
Figure FDA0003368959410000018
Therefore, the distance difference relationship between the moving target and the sensor pair is used to judge whether the field value exists in the measurement sequence, namely when
Figure FDA0003368959410000019
Considering that the observed value has no outlier; otherwise, the measurement has a outlier.
4. The multi-target robust positioning method for resisting measurement outlier interference according to claim 1, which is characterized in that: the content and the method for correcting the detected measurement outlier by introducing the virtual observation point, the association gate and the joint probability specifically comprise the following steps:
(1) the associated gate of an ellipsoid is equivalent to a sphere, the radius of which is expressed as
Figure FDA0003368959410000021
Where V is the volume of an ellipsoid, so that the observed value is predicted
Figure FDA0003368959410000022
Virtual observation points for the vicinity of the center yielding a Gaussian distribution with mean l/2 and variance l
Figure FDA0003368959410000023
Wherein N ispThe number of the virtual observation points is;
(2) the innovation of a virtual observation point can be represented as
Figure FDA0003368959410000024
According to the set threshold value g and the prediction innovation partyDifference (D)
Figure FDA0003368959410000025
To screen effective virtual observation points when
Figure FDA0003368959410000026
When the number is less than g, the virtual observation point is an effective virtual observation point; otherwise, the virtual observation point is eliminated, and the effective virtual observation point is represented as
Figure FDA0003368959410000027
Wherein N isvThe number of effective virtual observation points is;
(3) joint probability beta of valid virtual observation pointsm2As its weight, finally according to
Figure FDA0003368959410000028
And correcting the observed value with the outlier.
5. The multi-target robust positioning method for resisting measurement outlier interference according to claim 1, which is characterized in that: the content and the method for deducing the unscented particle filter model with the robust characteristic by using the corrected measured value specifically comprise the following steps: mean values of particles obtained from unscented Kalman filtering
Figure FDA0003368959410000029
And its corresponding covariance
Figure FDA00033689594100000210
Constructing a Gaussian suggestion density function as an importance density function in particle filtering to guide the resampling of the particles, wherein the expression is
Figure FDA00033689594100000211
To prevent particle degradation, particle filtering is then aggregated to obtain a minimum variance estimate of the state
Figure FDA00033689594100000212
6. The multi-target robust positioning method for resisting measurement outlier interference according to claim 1, which is characterized in that: the content and the method for realizing multi-target positioning by fusing the robust unscented particle filtering result of the multi-sensor specifically comprise the following steps:
(1) the method of claim 4 for correcting the measurement outliers of the measurement data obtained from all sensors, followed by robust unscented particle filtering using the corrected measurements to obtain all sets of filter results
Figure FDA00033689594100000213
Simultaneously collecting corresponding covariance sets
Figure FDA00033689594100000214
(2) According to
Figure FDA00033689594100000215
All covariance values are normalized, wherein [ ] refers to Hadamard product, and then normalized covariance values are used
Figure FDA00033689594100000216
And corresponding filtering results
Figure FDA00033689594100000217
Carrying out weighted summation to obtain the final filtering result of the moving target
Figure FDA0003368959410000031
Will be provided with
Figure FDA0003368959410000032
The elements of the first row and the first column, the third row and the first column and the fifth row and the first column in the mobile objectResult of positioning
Figure FDA0003368959410000033
Namely that
Figure FDA0003368959410000034
CN202111402405.4A 2021-11-23 2021-11-23 Multi-target robust positioning method for resisting measurement wild value interference Active CN114088086B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111402405.4A CN114088086B (en) 2021-11-23 2021-11-23 Multi-target robust positioning method for resisting measurement wild value interference

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111402405.4A CN114088086B (en) 2021-11-23 2021-11-23 Multi-target robust positioning method for resisting measurement wild value interference

Publications (2)

Publication Number Publication Date
CN114088086A true CN114088086A (en) 2022-02-25
CN114088086B CN114088086B (en) 2023-11-24

Family

ID=80303855

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111402405.4A Active CN114088086B (en) 2021-11-23 2021-11-23 Multi-target robust positioning method for resisting measurement wild value interference

Country Status (1)

Country Link
CN (1) CN114088086B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107607107A (en) * 2017-09-14 2018-01-19 斯坦德机器人（深圳）有限公司 A kind of Slam method and apparatus based on prior information
CN108983215A (en) * 2018-05-25 2018-12-11 哈尔滨工程大学 A kind of method for tracking target based on maximum cross-correlation entropy adaptively without mark particle filter
CN111047627A (en) * 2019-11-14 2020-04-21 中山大学 Smooth constraint unscented Kalman filtering method and target tracking method
CN111896008A (en) * 2020-08-20 2020-11-06 哈尔滨工程大学 Improved robust unscented Kalman filtering integrated navigation method
CN111983927A (en) * 2020-08-31 2020-11-24 郑州轻工业大学 Novel maximum entropy ellipsoid collective filtering method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107607107A (en) * 2017-09-14 2018-01-19 斯坦德机器人（深圳）有限公司 A kind of Slam method and apparatus based on prior information
CN108983215A (en) * 2018-05-25 2018-12-11 哈尔滨工程大学 A kind of method for tracking target based on maximum cross-correlation entropy adaptively without mark particle filter
CN111047627A (en) * 2019-11-14 2020-04-21 中山大学 Smooth constraint unscented Kalman filtering method and target tracking method
CN111896008A (en) * 2020-08-20 2020-11-06 哈尔滨工程大学 Improved robust unscented Kalman filtering integrated navigation method
CN111983927A (en) * 2020-08-31 2020-11-24 郑州轻工业大学 Novel maximum entropy ellipsoid collective filtering method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
LONG CHENG, ET AL: "An Indoor Localization Algorithm Based on Modified Joint Probabilistic Data Association for Wireless Sensor Network", 《IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS》, vol. 17, no. 1, pages 63 - 72, XP011818864, DOI: 10.1109/TII.2020.2979690 *
XINNAN FAN，ET AL: "A Single-Way Ranging Localization of AUVs Based on PSO of Outliers Elimination", 《JOURNAL OF ROBOTICS》, pages 1 - 15 *

Also Published As

Publication number Publication date
CN114088086B (en) 2023-11-24

Similar Documents

Publication Publication Date Title
CN1940591B (en) System and method of target tracking using sensor fusion
CN106443622B (en) A kind of distributed object tracking based on improvement joint probability data association
CN113074739B (en) UWB/INS fusion positioning method based on dynamic robust volume Kalman
CN107315171B (en) Radar networking target state and system error joint estimation algorithm
CN111178385A (en) Target tracking method for robust online multi-sensor fusion
CN110865343B (en) LMB-based particle filter tracking-before-detection method and system
CN107066806B (en) Data Association and device
CN111007454B (en) Extended target tracking method based on cooperative target information
CN111291471B (en) Constraint multi-model filtering method based on L1 regular unscented transformation
Mallick et al. Out-of-sequence measurement processing for tracking ground target using particle filters
CN115061139A (en) Multi-sensor fusion method and system for intelligent driving vehicle
CN111259332B (en) Fuzzy data association method and multi-target tracking method in clutter environment
CN115840221A (en) Method for realizing target feature extraction and multi-target tracking based on 4D millimeter wave radar
CN112328959A (en) Multi-target tracking method based on adaptive extended Kalman probability hypothesis density filter
CN115204212A (en) Multi-target tracking method based on STM-PMBM filtering algorithm
CN109509207B (en) Method for seamless tracking of point target and extended target
CN111735443B (en) Dense target track correlation method based on assignment matrix
CN114035187A (en) Perception fusion method of automatic driving system
CN111274529B (en) Robust Gao Sini Weisal PHD multi-expansion target tracking algorithm
CN117392215A (en) Mobile robot pose correction method based on improved AMCL and PL-ICP point cloud matching
CN116953692A (en) Track association method under cooperative tracking of active radar and passive radar
CN114088086B (en) Multi-target robust positioning method for resisting measurement wild value interference
CN115950414A (en) Adaptive multi-fusion SLAM method for different sensor data
CN114705223A (en) Inertial navigation error compensation method and system for multiple mobile intelligent bodies in target tracking
CN112285697A (en) Multi-sensor multi-target space-time deviation calibration and fusion method

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