CN109447133A - A kind of elimination method of the location information outlier based on SVR algorithm - Google Patents
A kind of elimination method of the location information outlier based on SVR algorithm Download PDFInfo
- Publication number
- CN109447133A CN109447133A CN201811182293.4A CN201811182293A CN109447133A CN 109447133 A CN109447133 A CN 109447133A CN 201811182293 A CN201811182293 A CN 201811182293A CN 109447133 A CN109447133 A CN 109447133A
- Authority
- CN
- China
- Prior art keywords
- auv
- svr
- value
- model
- 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
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
A kind of elimination method of the location information outlier based on SVR algorithm, the present invention use structural risk minimization, and in the process, the historical position resolved data at l moment first before acquisition AUV current time is as training dataset.Then it constructs SVR regression model and determines the parameter of SVR model to be built.Then the relative position of AUV is resolved by pure Furthest Neighbor and the conditional parameter of current time AUV substitution SVR model is obtained into position prediction.It determines whether the positional value calculated is outlier finally by setting degree of belief, rejects current calculated value if difference is greater than the degree of belief of setting, and return to the positional value for calculating subsequent time, otherwise retain the value.The present invention realizes the pretreatment of the position metric data in hydrolocation, reduces the error generated by outlier and is conducive to next filtering processing.
Description
Technical field
The present invention relates to the elimination method fields of location information outlier, more particularly to a kind of position based on SVR algorithm
The elimination method of information outlier.
Background technique
Pure Furthest Neighbor in USBL system, which is located through, measures distance in multiple positions, carries out position in conjunction with AUV course information
Confidence breath resolves positioning.Pass through the course speed v of vertical range d, AUV of AUV initial position p, AUV and underwater sound source transmitter
And course angleTo carry out real-time resolving to the position of AUV.When USBL is positioned, due to by underwater noise, reverberation,
The influence of the factors such as acoustic propagation multi-path effect, Doppler effect, velocity of sound nonlinear Distribution inevitably generates positioning and misses
Difference, to influence AUV navigation and positioning accuracy.Sometimes 1%~5% even up to 10%~20% serious offset target of data is true
Value, to become outlier.
In practical engineering applications, when carrying out positioning calculation to AUV with pure Furthest Neighbor, due to measurement system or
The reason of data transmission etc., it will lead to and occur outlier in metric data, directly apply to karr if without screening
Graceful filtering is modified, and does not simply fail to error existing for amendment state estimation, it is unstable to will cause filtering system instead, when serious
It will lead to filtering divergence.Therefore, measurement information is pre-processed, the position error as caused by outlier and filtering hair can be reduced
Dissipate
To solve the above-mentioned problems, scholars propose the method for many removal outlier.Wherein have based on least square
Method of abnormal value removing and correction, the method for abnormal value removing and correction based on fitting of a polynomial and the method for abnormal value removing and correction based on Kalman filtering.This
A little methods can reject outlier to a certain extent, but if parameter selection is bad, it is likely that make data processed result because
For be distorted too seriously without convincingness or be not achieved effectively reject outlier result.
Summary of the invention
In order to solve problem above, the present invention provides a kind of elimination method of location information outlier based on SVR algorithm,
The thought of SVR algorithm is to replace traditional empirical risk minimization principle using structural risk minimization, and sample is mapped
To higher-dimension nuclear space, nonlinear problem is mapped to the linear problem of higher dimensional space, to carry out in high-dimensional feature space linear
Return, thus using SVR algorithm in hydrolocation because caused by various errors position outlier reject, for this purpose,
The present invention provides a kind of elimination method of location information outlier based on SVR algorithm, method includes the following steps:
(1) to the position where different moments AUV, the historical position solution at l moment before obtaining current time counts
According to obtaining training dataset, wherein the data set at each moment includes existing condition data, including AUV and underwateracoustic
The course speed v of vertical range d, AUV of source transmitter, course angleAnd the positional value calculated;
(2) a SVR regression model is constructed, determines the regular parameter and Gauss nuclear parameter of SVR model to be built, then
Each training dataset is trained, corresponding support vector regression SVR model is finally obtained, it can be to present moment
AUV carries out position prediction;
(3) according to the vertical range difference d of current time AUV, course speed v, course angleAUV is resolved by pure Furthest Neighbor
Absolute position;
(4) carry out the position prediction value of pre-acquired subsequent time using support vector regression model.I.e. by current time AUV's
Vertical range difference d, course speed v, course angleSVR model in input step 2, obtains the prediction of current time AUV position
Value;
(5) judge the position resolving value for the AUV that subsequent time is obtained according to pure Furthest Neighbor and positional value that SVR is predicted
Whether difference meets condition, and the positional value that the resolving obtains if meeting is available, otherwise weeds out the value and recalculates.
Further improvement of the present invention in the step 1, obtains training sample set are as follows:
S={ (xi, yi), i=1,2 ..., l };
WhereinIt is the state parameter at i-th of moment, for the factor for influencing the position AUV, yi∈ R is
Model sample output is the position resolving value at i-th of moment.
Further improvement of the present invention, in the step 2, the SVR regression model of building is referred to:
F (x)=ωTφ(xk)+b
Wherein b is amount of bias, and φ is Nonlinear Mapping.The purpose of model training is to calculate optimized parameter ωTAnd b, it is of equal value
In the solution for solving following convex quadratic programming problem:
Constraint condition:
In formula, C > 0 is penalty coefficient, indicates the tolerance to the sample point for being more than maximization boundary, the bigger expression pair of C
The big sample punishment of error is bigger, and the on the contrary then sample big to error shows higher tolerance, and model can be changed in adjustment C
Generalization ability;ε is insensitive loss function;ξi,For slack variable.Using method of Lagrange multipliers, primal problem is converted into
Following dual problem:
Constraint condition:
In formula, Qij=φ (xi)Tφ(xj)=K (xi, xj), K () is kernel function here, can be to avoid by kernel function
Explicit use Nonlinear Mapping φ (), so as to avoid dimension disaster, αi,For the solution of above-mentioned dual problem, required is returned
Return function that can rewrite are as follows:
In formula, x is the factor for influencing the position AUV;xiFor the factor of the impact position of i-th of sample in l sample;K(xi,
X) it is kernel function, for radial basis function, RBF kernel function is shown below the kernel function used:
K (x, y)=exp |-γ | | x-y | |2|
In formula, γ is the width of kernel function.
Further improvement of the present invention is referred to using the location formula that pure Furthest Neighbor resolves AUV in the step 3:
In formula,
x3=X0
y3=Y0
Wherein source emission device psPosition be known as (X0, Y0), p1, p2, p3For three points on AUV track, three points with
psHorizontal distance be respectively r1, r2, r3。t13, t23AUV is distributed as from p1Position is to p3Time of position and from p2Position is to p3Position
The time set.
Further improvement of the present invention, in the step 5, described judges whether position resolving value and position prediction value are full
Sufficient condition specifically includes: setting degree of belief δ, is p (t) in the position resolving value that t moment obtains, pre- in the position that t moment obtains
Measured value isWhen p (t) meets:
(1)When, p (t) is received;
(2)When, p (t) is outlier, is removed, and p (t) is recalculated;
The value of δ is chosen according to the precision characteristic and confidence interval of position.
A kind of elimination method of the location information outlier based on SVR algorithm of the present invention, use background of the invention be
In USBL hydrolocation.Model is obtained by the location information training sample data of moment AUV before extracting.Practical Project is answered
With with important application value.It is compared with other methods, does not depend on design experiences, therefore more convenient.It mentions through the invention
Method out realizes the pretreatment of location position data, can effectively eliminate outlier reduce the error as caused by outlier and
Filtering divergence.Be conducive to next filtering processing.
Detailed description of the invention
Fig. 1 is that the present invention is based on the flow diagrams of the elimination method of the location information outlier of SVR algorithm.
Specific embodiment
Present invention is further described in detail with specific embodiment with reference to the accompanying drawing:
The present invention provides a kind of elimination method of location information outlier based on SVR algorithm, and the thought of SVR algorithm is to use
Structural risk minimization replaces traditional empirical risk minimization principle, and sample is mapped to higher-dimension nuclear space, will be non-thread
Property problem be mapped to the linear problem of higher dimensional space, so that linear regression is carried out in high-dimensional feature space, to utilize SVR algorithm
In hydrolocation because caused by various errors position outlier reject.
Step 1, training data is obtained first, to the position where different moments AUV, obtains 4 before current time
The historical position resolved data at moment, obtains training dataset.Wherein, the data set at each moment includes existing conditional number
According to including AUV and the course speed v of vertical range d, AUV of underwater sound source transmitter, course angleAnd calculate
Positional value.It is as follows to obtain training sample set:
S={ (xi, yi), i=1,2,3,4 }
WhereinThe factor of the position AUV, y are influenced for model sample inputi∈ R is model sample
Output is the position resolving value at i-th of moment.
Step 2, construction SVR regression model, determine the regular parameter and Gauss nuclear parameter of SVR model to be built, then
Each training dataset is trained, corresponding support vector regression SVR model is finally obtained, constructs linear regression function,
The function concrete form are as follows:
F (x)=ωTφ(xk)+b
Wherein b is amount of bias.The purpose of model training is to calculate optimized parameter ωTAnd b, it is equivalent to solve following convex secondary
The solution of planning problem:
Constraint condition:
In formula, C > 0 is penalty coefficient, indicates the tolerance to the sample point for being more than maximization boundary, the bigger expression pair of C
The big sample punishment of error is bigger, and the on the contrary then sample big to error shows higher tolerance, and model can be changed in adjustment C
Generalization ability;ε is insensitive loss function;ξi,For slack variable.Using method of Lagrange multipliers, primal problem is converted into
Following dual problem:
Constraint condition:
In formula, αi,For the solution of above-mentioned dual problem, so, in primal-dual optimization problem, it is only necessary to solve αiWithSubtract
Small calculation amount.Primal-dual optimization problem is solved, α is obtainediWithSolution.Qij=φ (xi)Tφ(xj)=K (xi, xj) it is core letter
Number, for radial basis function, RBF kernel function is shown below the kernel function used:
K (x, y)=exp |-γ | | x-y | |2|
γ is the width of kernel function.Required regression function can be rewritten are as follows:
In formula, x is the factor for influencing the position AUV;xiFor the factor of the impact position of i-th of sample in l sample.Therefore
Obtained trained SVR model form are as follows:
It in primal-dual optimization problem, needs to select two parameters C and γ, the optimal value for choosing C and γ can
Obtain an optimal models.By by training sample S={ (xt-i, p (t-i)), i=1,2,3,4 }, whereinBring model solution parameter alpha intoiWithDuring this using GA algorithm find optimal C and
γ, wherein the mean square error generated when using training SVR model is as GA algorithm fitness function.It can be obtained by and train in this way
SVR model.
Step 3, according to the vertical range difference d of current time AUV, course speed v, course angleIt is resolved by pure Furthest Neighbor
The absolute position of AUV.Pure Furthest Neighbor resolves the location formula of AUV are as follows:
In formula,
x3=X0
y3=Y0
Wherein source emission device psPosition be known as (X0, Y0), p1, p2, p3For three points on AUV track, three points with
psHorizontal distance be respectively r1, r2, r3。r13, t23AUV is distributed as from p1Position is to p3Time of position and from p2Position is to p3Position
The time set.
Step 4, the position prediction value of pre-acquired subsequent time is carried out using support vector regression model.I.e. by current time
The state parameter of AUVI.e. vertical range is poor, course is fast, the SVR model in boat input step 2, obtains current
The predicted value of moment AUV position obtains the predicted value of current time position
Step 5, judge the position resolving value of subsequent time AUV and whether the difference of positional value that SVR is predicted meets item
Degree of belief δ is arranged in part, is p (t) in the position resolving value that t moment obtains, is in the position prediction value that t moment obtainsWork as p
(t) meet:
(1)、When, p (t) is received;
(2)、When, p (t) is outlier, is removed, and p (t) is recalculated.
The value of δ is chosen according to the precision characteristic and confidence interval of position.The precision of model is related with the selection of δ value,
Selecting suitable δ value is to obtain the essential condition of good model effect.Reasonable threshold value (0.4- is chosen by priori knowledge
0.8m/s), 97% or more outlier can be rejected.
The object of the present invention is achieved like this:
In hydrolocation, for the location information of the AUV calculated, since AUV mobility is weak, acceleration is small, usually
It is driven at a constant speed with a certain steady rate along a direction, location information changes not too large in a short time, and AUV is in operational process
In, position data before can use predicts the position of subsequent time AUV.Outlier generally arises from transmission process generation
The successive observation point of location information can be regarded as consecutive variations, in conjunction with statistical models, use the conjunction of data by error value
Rationality test criterion carries out unruly-value rejecting.In order to effectively reject the outlier problem as caused by various errors in positioning, this
Invention provides a kind of location information method of abnormal value removing and correction based on support vector regression.Support vector machines uses structure risk most
Smallization principle replaces traditional empirical risk minimization principle, can preferably solve small sample, non-linear and high-dimensional ask
Topic.Relative to traditional mode of learning, it is generally not present Local Minimum problem, has very strong generalization ability, and not excessively
Rely on the quality and quantity of sample.The present invention predicts the next position variation of AUV by support vector regression model, from existing
Location information data concentrate choose historical data, predict subsequent time AUV position simultaneously made the difference with the position being calculated come
Reject the location information outlier after resolving.
The above described is only a preferred embodiment of the present invention, being not the limit for making any other form to the present invention
System, and made any modification or equivalent variations according to the technical essence of the invention, still fall within present invention model claimed
It encloses.
Claims (5)
1. a kind of elimination method of the location information outlier based on SVR algorithm, which is characterized in that method includes the following steps:
(1) to the position where different moments AUV, the historical position resolved data at l moment before current time is obtained, is obtained
To training dataset, wherein the data set at each moment includes existing condition data, is sent out including AUV and underwater sound source
The course speed v of vertical range d, AUV of emitter, course angleAnd the positional value calculated;
(2) a SVR regression model is constructed, the regular parameter and Gauss nuclear parameter of SVR model to be built are determined, then to every
A training dataset is trained, and finally obtains corresponding support vector regression SVR model, can AUV to present moment into
Row position prediction;
(3) according to the vertical range difference d of current time AUV, course speed v, course angleThe exhausted of AUV is resolved by pure Furthest Neighbor
To position;
(4) carry out the position prediction value of pre-acquired subsequent time using support vector regression model.I.e. by the vertical of current time AUV
Range difference d, course speed v, course angleSVR model in input step 2, obtains the predicted value of current time AUV position;
(5) judge the position resolving value for the AUV that subsequent time is obtained according to pure Furthest Neighbor and the difference of positional value that SVR is predicted
Whether condition is met, the positional value that the resolving obtains if meeting is available, otherwise weeds out the value and recalculates.
2. a kind of elimination method of location information outlier based on SVR algorithm according to claim 1, it is characterised in that:
In the step 1, training sample set is obtained are as follows:
S={ (xi, yi), i=1,2 ..., l };
WhereinIt is the state parameter at i-th of moment, for the factor for influencing the position AUV, yi∈ R is model
Sample output is the position resolving value at i-th of moment.
3. a kind of elimination method of location information outlier based on SVR algorithm according to claim 1, it is characterised in that:
In the step 2, the SVR regression model of building is referred to:
F (x)=ωTφ(xk)+b
Wherein b is amount of bias, and φ is Nonlinear Mapping.The purpose of model training is to calculate optimized parameter ωTAnd b, it is equivalent to solve
The solution of following convex quadratic programming problem:
Constraint condition
In formula, C > 0 is penalty coefficient, indicates that the bigger expression of C is to error to being more than the tolerance for maximizing the sample point on boundary
Big sample punishment is bigger, and the on the contrary then sample big to error shows higher tolerance, and the extensive of model can be changed in adjustment C
Ability;ε is insensitive loss function;ξi,For slack variable.Using method of Lagrange multipliers, primal problem is converted into as follows
Dual problem:
Constraint condition
In formula, Qij=φ (xi)Tφ(xj)=K (xi, xj), K () is kernel function here, can be to avoid explicit by kernel function
Use Nonlinear Mapping φ (), so as to avoid dimension disaster, αi,For the solution of above-mentioned dual problem, required recurrence letter
Number can be rewritten are as follows:
In formula, x is the factor for influencing the position AUV;xiFor the factor of the impact position of i-th of sample in l sample;K(xi, x) be
Kernel function, for radial basis function, RBF kernel function is shown below the kernel function used:
K (x, y)=exp |-γ | | x-y | |2|
In formula, γ is the width of kernel function.
4. a kind of elimination method of location information outlier based on SVR algorithm according to claim 1, it is characterised in that:
It is referred in the step 3 using the location formula that pure Furthest Neighbor resolves AUV:
In formula,
x3=X0
y3=Y0
Wherein source emission device psPosition be known as (X0, Y0), p1, p2, p3For three points on AUV track, three points and psWater
Flat distance is respectively r1, r2, r3。t13, t23AUV is distributed as from p1Position is to p3Time of position and from p2Position is to p3Position when
Between.
5. a kind of elimination method of location information outlier based on SVR algorithm according to claim 1, it is characterised in that:
In the step 5, described judges that resolving value in position is specifically included with whether position prediction value meets condition: setting degree of belief δ,
It is p (t) in the position resolving value that t moment obtains, is in the position prediction value that t moment obtainsWhen p (t) meets:
(1)When, p (t) is received;
(2)When, p (t) is outlier, is removed, and p (t) is recalculated;
The value of δ is chosen according to the precision characteristic and confidence interval of position.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811182293.4A CN109447133B (en) | 2018-10-11 | 2018-10-11 | SVR algorithm-based method for eliminating position information outliers |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811182293.4A CN109447133B (en) | 2018-10-11 | 2018-10-11 | SVR algorithm-based method for eliminating position information outliers |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109447133A true CN109447133A (en) | 2019-03-08 |
CN109447133B CN109447133B (en) | 2021-09-24 |
Family
ID=65545430
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811182293.4A Active CN109447133B (en) | 2018-10-11 | 2018-10-11 | SVR algorithm-based method for eliminating position information outliers |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109447133B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110057365A (en) * | 2019-05-05 | 2019-07-26 | 哈尔滨工程大学 | A kind of depth AUV dive localization method latent greatly |
CN110673148A (en) * | 2019-10-25 | 2020-01-10 | 海鹰企业集团有限责任公司 | Active sonar target real-time track resolving method |
CN110807027A (en) * | 2019-11-05 | 2020-02-18 | 上海眼控科技股份有限公司 | Outlier processing method, device, equipment and storage medium |
CN111076728A (en) * | 2020-01-13 | 2020-04-28 | 东南大学 | DR/USBL-based deep submersible vehicle combined navigation method |
CN111291927A (en) * | 2020-01-20 | 2020-06-16 | 河北环铁技术开发有限公司 | Annular RGV (reduced-size-vector graphics) trolley scheduling method based on SVR (support vector regression) model prediction |
CN112114287A (en) * | 2020-09-21 | 2020-12-22 | 东南大学 | Outlier real-time eliminating method for azimuth observation data |
CN113298138A (en) * | 2021-05-21 | 2021-08-24 | 西安建筑科技大学 | Radar radiation source individual identification method and system |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104853434A (en) * | 2015-01-13 | 2015-08-19 | 中山大学 | Indoor positioning method based on SVM and K mean value clustering algorithm |
CN107092744A (en) * | 2017-04-16 | 2017-08-25 | 北京工业大学 | Subsidence value Forecasting Methodology based on EMD SVR |
CN107576328A (en) * | 2017-08-22 | 2018-01-12 | 西北工业大学 | A kind of automatic Tracking monitoring system of water surface for submarine navigation device |
CN108489495A (en) * | 2018-03-29 | 2018-09-04 | 东南大学 | A kind of RFID tag indoor orientation method and equipment based on SVR and PSO |
-
2018
- 2018-10-11 CN CN201811182293.4A patent/CN109447133B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104853434A (en) * | 2015-01-13 | 2015-08-19 | 中山大学 | Indoor positioning method based on SVM and K mean value clustering algorithm |
CN107092744A (en) * | 2017-04-16 | 2017-08-25 | 北京工业大学 | Subsidence value Forecasting Methodology based on EMD SVR |
CN107576328A (en) * | 2017-08-22 | 2018-01-12 | 西北工业大学 | A kind of automatic Tracking monitoring system of water surface for submarine navigation device |
CN108489495A (en) * | 2018-03-29 | 2018-09-04 | 东南大学 | A kind of RFID tag indoor orientation method and equipment based on SVR and PSO |
Non-Patent Citations (1)
Title |
---|
徐晓苏等: "基于SALDE-UKF-SVM算法的WLAN室内定位方法", 《中国惯性技术学报》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110057365A (en) * | 2019-05-05 | 2019-07-26 | 哈尔滨工程大学 | A kind of depth AUV dive localization method latent greatly |
CN110057365B (en) * | 2019-05-05 | 2022-06-21 | 哈尔滨工程大学 | Large-submergence-depth AUV submergence positioning method |
CN110673148A (en) * | 2019-10-25 | 2020-01-10 | 海鹰企业集团有限责任公司 | Active sonar target real-time track resolving method |
CN110807027A (en) * | 2019-11-05 | 2020-02-18 | 上海眼控科技股份有限公司 | Outlier processing method, device, equipment and storage medium |
CN111076728A (en) * | 2020-01-13 | 2020-04-28 | 东南大学 | DR/USBL-based deep submersible vehicle combined navigation method |
CN111291927A (en) * | 2020-01-20 | 2020-06-16 | 河北环铁技术开发有限公司 | Annular RGV (reduced-size-vector graphics) trolley scheduling method based on SVR (support vector regression) model prediction |
CN111291927B (en) * | 2020-01-20 | 2022-11-04 | 河北环铁技术开发有限公司 | Annular RGV (reduced-size-vector graphics) trolley scheduling method based on SVR (support vector regression) model prediction |
CN112114287A (en) * | 2020-09-21 | 2020-12-22 | 东南大学 | Outlier real-time eliminating method for azimuth observation data |
CN112114287B (en) * | 2020-09-21 | 2021-04-20 | 东南大学 | Outlier real-time eliminating method for azimuth observation data |
CN113298138A (en) * | 2021-05-21 | 2021-08-24 | 西安建筑科技大学 | Radar radiation source individual identification method and system |
CN113298138B (en) * | 2021-05-21 | 2024-04-23 | 西安建筑科技大学 | Individual identification method and system for radar radiation source |
Also Published As
Publication number | Publication date |
---|---|
CN109447133B (en) | 2021-09-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109447133A (en) | A kind of elimination method of the location information outlier based on SVR algorithm | |
CN106595834B (en) | A method of obtaining the horizontal longitudinal correlation of the big depth sound field in deep-sea | |
CN106093849B (en) | A kind of Underwater Navigation method based on ranging and neural network algorithm | |
CN102707276B (en) | The AIS of based target topology information and radar track robust association algorithm | |
CN108734725B (en) | Probability data correlation filtering extended target tracking method based on Gaussian process | |
CN103759732B (en) | A kind of centralized multisensor multiple hypotheis tracking method of angle information auxiliary | |
CN110263075B (en) | LADCP and USBL combined observation and data processing method | |
CN110673148A (en) | Active sonar target real-time track resolving method | |
CN106896363A (en) | A kind of submarine target active tracing track initiation method | |
CN104777469B (en) | A kind of radar node selecting method based on error in measurement covariance matrix norm | |
CN106323272B (en) | A kind of method and electronic equipment obtaining track initiation track | |
CN110555398A (en) | Fault diagnosis method for determining first arrival moment of fault based on optimal filtering smoothness | |
CN112946698A (en) | Satellite signal cycle slip detection method based on reinforcement learning | |
ZA202305525B (en) | Method and device for generating water quality report based on distributed nodes, electronic device, and storage medium | |
CN109031314A (en) | A kind of underwater node positioning method towards Sound speed profile | |
CN104298650B (en) | Multi-method fusion based Kalman filtering quantization method | |
CN108759846B (en) | Method for establishing self-adaptive extended Kalman filtering noise model | |
CN106570536B (en) | A kind of positioning using TDOA aims of systems high precision tracking filtering method | |
CN110045363B (en) | Multi-radar track association method based on relative entropy | |
CN116027320A (en) | Radar and AIS data fusion method based on multi-factor Euclidean distance correlation | |
Xu et al. | Survey of track association of radar and AIS | |
CN107271974B (en) | Space-time error solving method based on stable angular points | |
CN108267743A (en) | Iteratively faster Underwater Navigation method based on fitting | |
CN110850366B (en) | Positioning method based on received signal strength under mixed line-of-sight and non-line-of-sight environment | |
Tsyganov et al. | Adaptive eetimation of a moving object trajectory using sequential hypothesis testing |
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 |