CN108549221A - A kind of filtering method and relevant apparatus of linear stochaastic system - Google Patents
A kind of filtering method and relevant apparatus of linear stochaastic system Download PDFInfo
- Publication number
- CN108549221A CN108549221A CN201810293455.5A CN201810293455A CN108549221A CN 108549221 A CN108549221 A CN 108549221A CN 201810293455 A CN201810293455 A CN 201810293455A CN 108549221 A CN108549221 A CN 108549221A
- Authority
- CN
- China
- Prior art keywords
- noise
- value expression
- filter
- observation value
- obtains
- 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
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Abstract
This application discloses a kind of filtering methods of linear stochaastic system, including：Item is superimposed to linear stochaastic system addition noise, obtains noise superposition linear stochaastic system；It is superimposed linear stochaastic system according to noise and is observed value expression Construction treatment, obtains observation value expression；According to optimal filter method, noise superposition linear stochaastic system and observation value expression are filtered as parameter, obtain filter result.It is superimposed item by adding noise to linear stochaastic system, linear stochaastic system is set not only to be interfered by the system noise of last moment, also interfered by the system noise at current time, therefore the environment of noise jamming in actual conditions can be more completely fitted, the precision that subsequent filtering can further be improved makes to be filtered more accurate.Disclosed herein as well is a kind of linear stochaastic system filter, another filter and computer readable storage mediums, have above-mentioned advantageous effect.
Description
Technical field
This application involves automation control area, more particularly to a kind of filtering method of linear stochaastic system, filter,
Another filter and computer readable storage medium.
Background technology
It, constantly will control, the involvement of Communications And Computer technology in present industrial system as information technology continues to develop
In being operated to the mechanical device of information processing and different levels.Using after these technologies can to automatic control system realize with
Machine controls, and has obtained stochastic control system.Wherein, stochastic control system is the dynamical system influenced by enchancement factor.Continuous
Research in, mainly three of stochastic control system aspects are improved, are modeling, filtering and random adaptation control respectively
System.
In general, mainly realizing the fitting to stochastic regime, then root to constructing linear stochaastic system during modeling
It is filtered operation according to the linear stochaastic system.But with the rising of accuracy of detection, the noise in actual environment is gradually found
It is especially complex with interference, and with application development, the design requirement of system becomes more sophisticated so that the prior art can not
Reach the filtering accuracy requirement of design.
Therefore, how to solve the filtering problem of the linear stochaastic system in the case where noise jamming is complex is ability
Field technique personnel Important Problems of interest.
Invention content
The purpose of the application is to provide a kind of filtering method of linear stochaastic system, filter, another filter
And computer readable storage medium, by linear stochaastic system add noise be superimposed item, make linear stochaastic system not only by
System noise to last moment is interfered, and is also interfered by the system noise at current time, therefore can more completely be fitted reality
The environment of noise jamming in the situation of border, can further improve the precision of subsequent filtering, make to be filtered more accurate.
In order to solve the above technical problems, the application provides a kind of filtering method of linear stochaastic system, including：
Item is superimposed to linear stochaastic system addition noise, obtains noise superposition linear stochaastic system；Wherein, it is described linearly with
Machine system carries out modeling and handles to obtain；
It is superimposed linear stochaastic system according to the noise and is observed value expression Construction treatment, obtains observation expression
Formula；
According to optimal filter method, the noise is superimposed linear stochaastic system and the observation value expression as parameter
It is filtered, obtains filter result.
Optionally, linear stochaastic system is superimposed according to the noise and is observed value expression Construction treatment, observed
Value expression, including：
It is superimposed linear stochaastic system according to the noise and carries out initializer Construction treatment, obtains the expression of initial observation value
Formula；
Communication bound term is added to the initial observation value expression, obtains communication constraint observation value expression；
Observation value expression is constrained to the communication and adds packet loss coefficient entry, obtains the observation value expression.
Optionally, communication bound term is added to the initial observation value expression, obtains communication constraint observation value expression,
Including：
Bound term is communicated to initial observation value expression addition Markov, obtains the communication constraint observation table
Up to formula.
Optionally, according to optimal filter method, the noise is superimposed linear stochaastic system and the observation value expression
It is filtered as parameter, obtains filter result, including：
According to optimal filter method, the noise is superimposed linear stochaastic system and the observation value expression as parameter
Construct optimal filter equation；
Determine error co-variance matrix according to primary condition and the optimal filter equation, by the optimal filter equation and
Determining error co-variance matrix carries out iteration calculating, obtains filter result.
The application also provides a kind of filter of linear stochaastic system, including：
Noise laminating module obtains noise superposition linear random system for being superimposed item to linear stochaastic system addition noise
System；Wherein, the linear stochaastic system carries out modeling and handles to obtain；
Observation acquisition module is observed for being superimposed linear stochaastic system according to the noise at value expression construction
Reason obtains observation value expression；
Filter module, for according to optimal filter method, the noise to be superimposed linear stochaastic system and the observation
Expression formula is filtered as parameter, obtains filter result.
Optionally, the observation acquisition module includes：
Initial construction unit carries out initializer Construction treatment for being superimposed linear stochaastic system according to the noise,
Obtain initial observation value expression；
Communication constraint unit obtains communication constraint and sees for adding communication bound term to the initial observation value expression
Measured value expression formula；
Packet loss constraint element adds packet loss coefficient entry for constraining observation value expression to the communication, obtains the sight
Measured value expression formula.
Optionally, the communication constraint element is specifically used for initial observation value expression addition Markov communication
Bound term obtains the communication constraint observation value expression.
Optionally, the filter module includes：
Optimal filter equation acquiring unit, for according to optimal filter method, the noise to be superimposed linear stochaastic system
Go out optimal filter equation as parametric configuration with the observation value expression；
Iterative filtering computing unit, for determining error covariance square according to primary condition and the optimal filter equation
The optimal filter equation and the error co-variance matrix determined are carried out iteration calculating, obtain filter result by battle array.
The application also provides a kind of filter of linear stochaastic system, including：
Memory, for storing computer program；
Processor, the step of filtering method as described above is realized when for executing the computer program.
The application also provides a kind of computer readable storage medium, and calculating is stored on the computer readable storage medium
The step of machine program, the computer program realizes filtering method as described above when being executed by processor.
A kind of filtering method of linear stochaastic system provided herein, including：Noise is added to linear stochaastic system
It is superimposed item, obtains noise superposition linear stochaastic system；Wherein, the linear stochaastic system carries out modeling and handles to obtain；Root
It is observed value expression Construction treatment according to noise superposition linear stochaastic system, obtains observation value expression；According to optimal
The noise is superimposed linear stochaastic system and the observation value expression is filtered as parameter, obtained by filtering method
To filter result.
It is superimposed item by adding noise to linear stochaastic system, makes linear stochaastic system not only by the system of last moment
Noise jamming is also interfered by the system noise at current time, or makes linear stochaastic system system not only by this moment
System noise interfere, also by the system noise of subsequent time interfere, due in actual conditions Complex Noise interference usually by
It is caused in noise superposition, therefore can more completely be fitted the environment of noise jamming in actual conditions, can further carried
The precision of high subsequent filtering, makes to be filtered more accurately, meets higher filtering requirements.
Also, the observation expression formula got through this embodiment can analog signal situation after transmission.
System data traffic is reduced particularly by communication constraint is added, ensure that filtering accuracy, and system can also be reduced
Energy consumption.
The application also provides a kind of linear stochaastic system filter, another filter and computer-readable storage
Medium has above-mentioned advantageous effect, and this will not be repeated here.
Description of the drawings
In order to illustrate the technical solutions in the embodiments of the present application or in the prior art more clearly, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of application for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
A kind of flow chart of the filtering method for linear stochaastic system that Fig. 1 is provided by the embodiment of the present application；
The flow chart of the observation value expression acquisition process for the filtering method that Fig. 2 is provided by the embodiment of the present application；
The flow chart of the filtering for the filtering method that Fig. 3 is provided by the embodiment of the present application；
A kind of structural schematic diagram of the filter for linear stochaastic system that Fig. 4 is provided by the embodiment of the present application；
A kind of process schematic of the filtering method for linear stochaastic system that Fig. 5 is provided by the embodiment of the present application.
Specific implementation mode
The core of the application is to provide a kind of filtering method of linear stochaastic system, filter, another filter
And computer readable storage medium, by linear stochaastic system add noise be superimposed item, make linear stochaastic system not only by
System noise to last moment is interfered, and is also interfered by the system noise at current time, therefore can more completely be fitted reality
The environment of noise jamming in the situation of border, can further improve the precision of subsequent filtering, make to be filtered more accurate.
To keep the purpose, technical scheme and advantage of the embodiment of the present application clearer, below in conjunction with the embodiment of the present application
In attached drawing, technical solutions in the embodiments of the present application is clearly and completely described, it is clear that described embodiment is
Some embodiments of the present application, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art
The every other embodiment obtained without making creative work, shall fall in the protection scope of this application.
Usual linear stochaastic system includes at least sensor network, shared transmission channel and long-range filter, and the application is real
Example is applied also to be unfolded in this structure.Therefore, it when the filtering accuracy of total system declines, can all be adjusted in three above part
The whole precision with raising filtering, but under complicated noise jamming environment, adjusting subsequent process anyway all can not be very
Good solution precision problem.Further, find in Complex Noise interference environment generally there is the noise of superposition to make through analysis
Ambient noise complicates, and is all in the prior art a noise item to modeling general, therefore can not simulate and make an uproar well
Acoustic environment, and then subsequent filtering accuracy is caused to decline.
Referring to FIG. 1, a kind of flow of the filtering method for linear stochaastic system that Fig. 1 is provided by the embodiment of the present application
Figure.
This implementation provides a kind of filtering method of linear stochaastic system, may include：
S101 is superimposed item to linear stochaastic system addition noise, obtains noise superposition linear stochaastic system；Wherein, linearly
Stochastic system carries out modeling and handles to obtain；
This step is intended to be superimposed item, the linear stochaastic system after being superimposed to linear stochaastic system addition noise.Wherein,
Linear stochaastic system namely carries out what modeling was handled.In general, being usually filtered head to linear stochaastic system
Modeling processing was carried out to system before this, accordingly, this step can also regard the modeling processing carried out to system as, be modeled usually
To linear stochaastic system on addition noise be superimposed item so that the system of fitting is not only dry by the system noise of last moment
It disturbs, is also interfered by the system noise at current time, more completely reacted the random noise in actual conditions.Therefore, may be used
The precision that filtering is improved with the noise superposition linear stochaastic system obtained by this step, obtains more accurate filter result.
In addition, in the prior art to the modeling of linear stochaastic system generally in order to which convenience of calculation is usually using the mark of modeling
Quasi- form, that is, only include the noise item of single point in time.The canonical form can facilitate and calculated when simulating calculating
Journey.But single noise item is the simple abstract to the random noise in actual conditions, and linear stochaastic system can not be made quasi-
Close complete fitting actual conditions.And it is in the prior art linear to be superimposed linear stochaastic system comparison by the noise in this step
Stochastic system has completely been fitted actual conditions, improves the precision of filtering, can obtain more accurate filter result.
Wherein, the noise superposition item added needs the difference of the specific formula expression-form according to linear stochaastic system,
And different set-up modes, but being makes stochastic linear system on the basis of last moment system noise, and addition is current
The system noise at moment.Either on the basis of receiving current time system noise, the system noise of subsequent time is added.Always
Be the noise superposition item that another moment is added on the basis of the noise to existing linear stochaastic system, make linear stochaastic system can
To be completely fitted the chance phenomenon in actual conditions.
Specifically, the noise superposition linear stochaastic system obtained in this step can be indicated by following formula：
Wherein, k=0,1 ... at the time of representing discrete time, m representative sensor quantity, x (k) represents n and maintains system shape
State, y1(k)…ym(k) measured value of m sensor is represented.W (k) represents system noise, v1(k)…vm(k) representative sensor is surveyed
Measure noise.A(k),B1(k),B2(k+1),c1(k),...,cm(k) the systematic parameter matrix of suitable dimension is represented.Wherein, B2(k+
1) w (k+1) is the noise superposition item added in this formula.
S102 is superimposed linear stochaastic system according to noise and is observed value expression Construction treatment, obtains observation expression
Formula；
On the basis of step S101, this step is intended to the noise obtained according to previous step superposition stochastic system and is seen
Measured value expression formula Construction treatment obtains observation value expression.
Wherein, observation value expression is the expression formula for the signal that filter receives, and is transmitted in the channel due to signal
Certain loss is had in the process, and receives the limitation of physics, and the signal that filter is received is not that system is directly sent out
The signal sent, it is therefore desirable to observation value expression be constructed according to selected situation, and then can according to the observation value expression
With the filtering situation of test filter in a practical situation.
Further, the observation value expression Construction treatment conducted in this step is mainly superimposed line according to according to the noise
Property stochastic system construct the expression formula after signal transmission, specifically, wherein to transmission signal be added to which type of constrain
Can be selected according to actual conditions, wherein include but not limited to communication channel constraint and or data packetloss constraint.
S103, according to optimal filter method, using noise superposition linear stochaastic system and observation value expression as parameter into
Row is filtered, and obtains filter result.
On the basis of step S102, this step is intended to, according to optimal filter method, the noise that above-mentioned steps obtain be folded
Add linear stochaastic system and observation value expression to be filtered, obtains filter result.
This step is mainly filtered, and the method filtered in the art there are many kinds of, but the present embodiment
In expression-form meet the requirement of optimal filter method, and optimal filter method can obtain good filter effect, because
This this step selects optimal filter method to be filtered, and obtains filter result.
To sum up, the present embodiment can be superimposed item by adding noise to linear stochaastic system, make linear stochaastic system not only
It is interfered, is also interfered by the system noise at current time, or make linear stochaastic system system by the system noise of last moment
System is not only interfered by the system noise at this moment, is also interfered by the system noise of subsequent time, therefore can be more complete
It is fitted the environment of noise jamming in actual conditions, the precision of subsequent filtering can be further improved, make to be filtered more
Add accurate.
Referring to FIG. 2, the stream for observing value expression acquisition process for the filtering method that Fig. 2 is provided by the embodiment of the present application
Cheng Tu.
Based on a upper embodiment, the present embodiment mainly for how to be constructed in a upper embodiment observation value expression do one
A to illustrate, other parts are substantially the same with a upper embodiment, and same section can refer to a upper embodiment, not do herein superfluous
It states.
The present embodiment may include：
S201 is superimposed linear stochaastic system according to noise and carries out initializer Construction treatment, obtains initial observation value table
Up to formula；
S202 adds communication bound term to initial observation value expression, obtains communication constraint observation value expression；
S203 constrains observation value expression to communication and adds packet loss coefficient entry, obtains observation value expression.
The present embodiment is mainly to add communication constraint and packet loss about to expression formula in the construction process of observation value expression
Beam, that is, the communication bound term of step S202 addition and step S203 additions packet loss coefficient entry, finally obtaining can be fitted
The observation value expression of actual transmissions situation.
Wherein, step S202 adds communication bound term to initial observation value expression, can obtain communication constraint observation
Expression formula.The reason is that since channel in a practical situation receives the limitation of physics, there can not be the so big traffic, therefore
It needs to add corresponding communication constraint in the expression formula of observation, to reduce traffic, reduces the energy consumption of data transmission.
Optionally, step S202 can add Markov to initial observation value expression and communicate bound term, be communicated
Constraint observation value expression.
Wherein, Markov communication constraint, specific features are that each moment, only there are one the measuring values of sensor to account for
With shared channel, and the selection rule of each moment sensor is according to a given Markov state transition probability square
Battle array.Therefore, when carrying out status predication in filtering, only know Markov state probability transfer matrix, in this case it is not apparent that tool
The measuring value of that sensor of body occupies shared channel.
Specifically, can be expressed according to following formula：
Wherein,For the metrical information received by long-range filter；γ is the weight coefficient between 0 to 1；{θ(k)}
For defined Markov Chain, value is 1 to the integer between m, indicate the k moment choose the observation data of which sensor into
Row transmission；ΓiIt is m dimension diagonal matrix, only i-th of element of diagonal line is 1, remaining element is 0；ImUnit matrix is tieed up for m.
Generally speaking, communication constraint can be Markov communication constraint, i.e., only there are one the surveys of sensor for synchronization
Amount information can be transferred to long-range filter, and Sensor scheduling sequence obeys Markov Chain { θ (k) }.
Wherein, step S203 mainly constrains observation value expression to communication and adds packet loss coefficient entry, obtains the observation table
Up to formula.Wherein, packet loss coefficient entry is the coefficient entry for simulating packet drop.
Specifically, being to add corresponding packet loss coefficient entry on the basis of above-mentioned formula, observation expression formula is obtained, is indicated such as
Under：
Wherein α (k) be used for describe the k moment whether packet loss.If packet loss occurs, α (k)=0, otherwise it is equal to 1.Y be γ and
ImProduct.
Wherein, the probability of data packetloss obeys Bernoulli Jacob's distribution.
Therefore, the observation expression formula got through this embodiment can analog signal situation after transmission.
System data traffic is reduced particularly by communication constraint is added, ensure that filtering accuracy, and system can also be reduced
Energy consumption.
Referring to FIG. 3, the flow chart of the filtering for the filtering method that Fig. 3 is provided by the embodiment of the present application.
Based on a upper embodiment, how the present embodiment is mainly for being filtered do one specifically in a upper embodiment
Bright, other parts are substantially the same with a upper embodiment, and same section can refer to a upper embodiment, and this will not be repeated here.
The present embodiment may include：
S301, according to optimal filter method, using noise superposition linear stochaastic system and observation value expression as parameter structure
Produce optimal filter equation；
This step is intended to the method according to optimal filter, and linear stochaastic system and observation value expression structure are superimposed according to noise
Produce optimal filter equation.
Specifically, the optimal filter equation can indicate as follows：
Wherein, g (k)=[gT(k,1),...,gT(k,m)]T
Wherein K (k), F (k) are two crucial gain parameters, all include P (k) in expression formula.
Wherein, K (k) is filtering gain, and F (k) is prediction gain, and e (k) is the new breath at k moment, for correcting filter value.For parameter matrix.
S302 determines error co-variance matrix according to primary condition and optimal filter equation, by optimal filter equation and really
Fixed error co-variance matrix carries out iteration calculating, obtains filter result.
On the basis of step S301, this step is intended to be iterated calculating according to primary condition and optimal filter equation,
Obtain filter result to the end.
Specifically, wherein determining error co-variance matrix can be expressed according to following formula：
Wherein,For parameter matrix, Re(k) the association side of e (k) is represented
Poor matrix can calculate P (k+1), to calculate the filter value at k+1 moment by the iterative equation by P (k).
The embodiment of the present application provides a kind of filtering method of linear stochaastic system, can be by adding to linear stochaastic system
Plus noise is superimposed item, so that linear stochaastic system is not only interfered by the system noise of last moment, is also by current time
System noise jamming, or linear stochaastic system system is made not only to be interfered by the system noise at this moment, also by subsequent time
System noise interference, therefore can more completely be fitted the environment of noise jamming in actual conditions, can further carry
The precision of high subsequent filtering makes to be filtered more accurate.
A kind of filter of linear stochaastic system provided by the embodiments of the present application is introduced below, it is described below
A kind of filter of linear stochaastic system can correspond ginseng with a kind of above-described filtering method of linear stochaastic system
According to.
Referring to FIG. 4, a kind of structure of the filter for linear stochaastic system that Fig. 4 is provided by the embodiment of the present application is shown
It is intended to.
The present embodiment provides a kind of filters of linear stochaastic system, may include：
Noise laminating module 100 obtains noise superposition linear random for being superimposed item to linear stochaastic system addition noise
System；Wherein, linear stochaastic system carries out modeling and handles to obtain；
Observation acquisition module 200 is observed for being superimposed linear stochaastic system according to noise at value expression construction
Reason obtains observation value expression；
Filter module 300, for according to optimal filter method, by noise superposition linear stochaastic system and observing value expression
It is filtered as parameter, obtains filter result.
Optionally, which may include：
Initial construction unit carries out initializer Construction treatment for being superimposed linear stochaastic system according to noise, obtains
Initial observation value expression；
Communication constraint unit obtains communication constraint observation for adding communication bound term to initial observation value expression
Expression formula；
Packet loss constraint element adds packet loss coefficient entry for constraining observation value expression to communication, obtains observation expression
Formula.
Optionally, communication constraint element can be also used for the addition Markov communication constraint of initial observation value expression
, obtain communication constraint observation value expression.
Optionally, filter module 300 may include：
Optimal filter equation acquiring unit, for according to optimal filter method, noise to be superimposed linear stochaastic system and sight
Measured value expression formula goes out optimal filter equation as parametric configuration；
Iterative filtering computing unit will for determining error co-variance matrix according to primary condition and optimal filter equation
Optimal filter equation and the error co-variance matrix determined carry out iteration calculating, obtain filter result.
The embodiment of the present application also provides a kind of filter of linear stochaastic system, may include
Memory, for storing computer program；
Processor, when for executing computer program the step of the realization such as filtering method of above-described embodiment.
The embodiment of the present application also provides a kind of computer readable storage medium, and meter is stored on computer readable storage medium
Calculation machine program, when computer program is executed by processor the step of the realization such as filtering method of above-described embodiment.
Based on all of above embodiment, there can also be following examples.System provided in this implementation is united for 3 levels,
Specific implementation step is as follows：
The first step, communication constraint analysis modeling：
For following 3 rank linear stochaastic system：
Wherein,
c1(k)=[0.5 1 0.6], c2(k)=[1 0.5 1.5], c3(k)=[0.8 1.3 1]；
Define v (k)=[v1(k),v2(k),v3(k)]T, then w (k) and v (k) is orthogonal white Gaussian noise, variance
Respectively Q (k)=0.1 and R (k)=0.1I3, wherein I3For 3 rank unit matrixs.The mean value of original state x (0) is distinguished with variance
For x0=[2 1 1.5]T, Px(0)=I3。
Referring to FIG. 5, a kind of process of the filtering method for linear stochaastic system that Fig. 5 is provided by the embodiment of the present application is shown
It is intended to.
As shown in fig. 5, it is assumed that there are 3 sensors to measure system signal, corresponding three measuring value y1(k)、y2(k)、
y3(k).Communication is constrained to Markov agreement, i.e., each moment, only there are one the measuring values of sensor can occupy shared letter
Road, and the selection rule of each moment sensor is according to a given Markov state transition probability matrix.Long-range filter
Wave device only knows Markov state probability transfer matrix when carrying out status predication, in this case it is not apparent that specific sensor
Measuring value occupies shared channel.
Set Markov state transition probability matrixInitial probability distribution is π1(0)=
0.1、π2(0)=0.2, π3(0)=0.7.
Under Markov communication constraint, the sensor network for constructing following form measures value expression：
Wherein, γ=0.8,θ (k) is to meet above-mentioned horse
The markovian state of Er Kefu state-transition matrixes, value range are { 1,2,3 }, and represent each moment selection is which
The measuring value of one sensor,Represent the signal that long-range filter can receive.
Second step：Data packetloss analysis modeling：
Data packetloss occurs at random in the remote transmission of shared channel, and the probability of data packetloss obeys Bernoulli Jacob's distribution,
I.e. the drop probabilities at k moment are 1-q (k), and q (k) represents k time datas and is bundled into the probability that work(is transferred to long-range filter, are implemented
Q (k)=0.9 is assumed in example, constructs the measurement value expression of following form：
Wherein, α (k)=1 represents data packet transmission success, and α (k)=0 represents data-bag lost.
Third walks, and seeks the optimal filter value of linear stochaastic system state：
Optimal filter value meets following equation：
Wherein,
Wherein, augmentation coefficient matrix is combined to obtain by original system coefficient matrix with probabilistic model.
4th step seeks the evaluated error covariance matrix under least mean-square error index：
Later according to primary condition, iteration carries out the calculating of third step and the 4th step, obtains filter result.
Each embodiment is described by the way of progressive in specification, the highlights of each of the examples are with other realities
Apply the difference of example, just to refer each other for identical similar portion between each embodiment.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related place is referring to method part illustration
.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, depends on the specific application and design constraint of technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think to exceed scope of the present application.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Above to a kind of filtering method of linear stochaastic system provided herein, filter, another filtering dress
It sets and computer readable storage medium is described in detail.Used herein principle and reality of the specific case to the application
The mode of applying is expounded, the description of the example is only used to help understand the method for the present application and its core ideas.It answers
It, for those skilled in the art, can also be to this under the premise of not departing from the application principle when pointing out
Some improvement and modification can also be carried out for application, these improvement and modification are also fallen into the application scope of the claims.
Claims (10)
1. a kind of filtering method of linear stochaastic system, which is characterized in that including：
Item is superimposed to linear stochaastic system addition noise, obtains noise superposition linear stochaastic system；Wherein, the linear random system
System carries out modeling and handles to obtain；
It is superimposed linear stochaastic system according to the noise and is observed value expression Construction treatment, obtains observation value expression；
According to optimal filter method, the noise is superimposed linear stochaastic system and the observation value expression and is carried out as parameter
It is filtered, obtains filter result.
2. filtering method according to claim 1, which is characterized in that be superimposed linear stochaastic system according to the noise and carry out
Value expression Construction treatment is observed, observation value expression is obtained, including：
It is superimposed linear stochaastic system according to the noise and carries out initializer Construction treatment, obtains initial observation value expression；
Communication bound term is added to the initial observation value expression, obtains communication constraint observation value expression；
Observation value expression is constrained to the communication and adds packet loss coefficient entry, obtains the observation value expression.
3. filtering method according to claim 2, which is characterized in that about to initial observation value expression addition communication
Shu Xiang obtains communication constraint observation value expression, including：
Bound term is communicated to initial observation value expression addition Markov, obtains the communication constraint observation expression
Formula.
4. filtering method according to claim 3, which is characterized in that according to optimal filter method, the noise is superimposed
Linear stochaastic system and the observation value expression are filtered as parameter, obtain filter result, including：
According to optimal filter method, the noise is superimposed linear stochaastic system and the observation value expression as parametric configuration
Go out optimal filter equation；
Error co-variance matrix is determined according to primary condition and the optimal filter equation, by the optimal filter equation and determination
Error co-variance matrix carry out iteration calculating, obtain filter result.
5. a kind of filter of linear stochaastic system, which is characterized in that including：
Noise laminating module obtains noise superposition linear stochaastic system for being superimposed item to linear stochaastic system addition noise；Its
In, the linear stochaastic system carries out modeling and handles to obtain；
Observation acquisition module is observed value expression Construction treatment for being superimposed linear stochaastic system according to the noise,
Obtain observation value expression；
Filter module, for according to optimal filter method, the noise being superimposed linear stochaastic system and the observation is expressed
Formula is filtered as parameter, obtains filter result.
6. filter according to claim 5, which is characterized in that the observation acquisition module includes：
Initial construction unit carries out initializer Construction treatment for being superimposed linear stochaastic system according to the noise, obtains
Initial observation value expression；
Communication constraint unit obtains communication constraint observation for adding communication bound term to the initial observation value expression
Expression formula；
Packet loss constraint element adds packet loss coefficient entry for constraining observation value expression to the communication, obtains the observation
Expression formula.
7. filter according to claim 6, which is characterized in that the communication constraint element is specifically used for described first
Begin observation value expression addition Markov communication bound term, obtains the communication constraint observation value expression.
8. filter according to claim 7, which is characterized in that the filter module includes：
Optimal filter equation acquiring unit, for according to optimal filter method, the noise to be superimposed linear stochaastic system and institute
It states observation value expression and goes out optimal filter equation as parametric configuration；
Iterative filtering computing unit will for determining error co-variance matrix according to primary condition and the optimal filter equation
The optimal filter equation and the error co-variance matrix determined carry out iteration calculating, obtain filter result.
9. a kind of filter of linear stochaastic system, which is characterized in that including：
Memory, for storing computer program；
Processor realizes the step such as Claims 1-4 any one of them filtering method when for executing the computer program
Suddenly.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer on the computer readable storage medium
Program realizes the step such as Claims 1-4 any one of them filtering method when the computer program is executed by processor
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810293455.5A CN108549221B (en) | 2018-03-30 | 2018-03-30 | Filtering method and related device of linear stochastic system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810293455.5A CN108549221B (en) | 2018-03-30 | 2018-03-30 | Filtering method and related device of linear stochastic system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108549221A true CN108549221A (en) | 2018-09-18 |
| CN108549221B CN108549221B (en) | 2020-11-10 |
Family
ID=63514114
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810293455.5A Active CN108549221B (en) | 2018-03-30 | 2018-03-30 | Filtering method and related device of linear stochastic system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108549221B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6158338A (en) * | 1984-08-29 | 1986-03-25 | Nec Corp | Spread spectrum phase encoding power communication system |
| CN103336525A (en) * | 2013-06-18 | 2013-10-02 | 哈尔滨工程大学 | Convenient UKF (Unscented Kalman Filter) filtering method for high weights of stochastic system |
| CN103888100A (en) * | 2014-03-29 | 2014-06-25 | 北京航空航天大学 | Method for filtering non-Gaussian linear stochastic system based on negentropy |
| CN103941692A (en) * | 2014-04-21 | 2014-07-23 | 南京师范大学 | Random switching system H8 filtering information processing method based on resident probabilistic information method |
-
2018
- 2018-03-30 CN CN201810293455.5A patent/CN108549221B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6158338A (en) * | 1984-08-29 | 1986-03-25 | Nec Corp | Spread spectrum phase encoding power communication system |
| CN103336525A (en) * | 2013-06-18 | 2013-10-02 | 哈尔滨工程大学 | Convenient UKF (Unscented Kalman Filter) filtering method for high weights of stochastic system |
| CN103888100A (en) * | 2014-03-29 | 2014-06-25 | 北京航空航天大学 | Method for filtering non-Gaussian linear stochastic system based on negentropy |
| CN103941692A (en) * | 2014-04-21 | 2014-07-23 | 南京师范大学 | Random switching system H8 filtering information processing method based on resident probabilistic information method |
Non-Patent Citations (4)
| Title |
|---|
| DONGYAN CHEN 等: "Optimal filtering for systems with finite-step autocorrelated process noises, random one-step sensor delay and missing measurements", 《COMMUN NONLINEAR SCI NUMER SIMULAT》 * |
| YONG XU 等: "Passive filter design for periodic stochastic systems with quantized measurements and randomly occurring nonlinearities", 《JOURNAL OF THE FRANKLIN INSTITUTE》 * |
| 褚东升 等: "噪声相关时多通道带乘性噪声系统最优滤波", 《青岛海洋大学学报》 * |
| 陈云 等: "时滞随机双线性系统控制与观测器设计", 《PROCEEDINGS OF THE 33RD CHINESE CONTROL CONFERENCE》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108549221B (en) | 2020-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kruse et al. | Benchmarking invertible architectures on inverse problems | |
| Ostwald | The fractal analysis of architecture: calibrating the box-counting method using scaling coefficient and grid disposition variables | |
| JP2010281826A5 (en) | ||
| CN105046366B (en) | model training method and device | |
| Abudu et al. | Modeling of daily pan evaporation using partial least squares regression | |
| EP2584520A1 (en) | Method for providing a study pattern analysis service on a network, and a server used therewith | |
| Yuan et al. | Calibration of stochastic computer models using stochastic approximation methods | |
| CN108460462A (en) | A kind of Interval neural networks learning method based on interval parameter optimization | |
| CN109409739A (en) | A kind of crowdsourcing platform method for allocating tasks based on part Observable markov decision process | |
| CN105043433A (en) | Neural-network-based rapid compensation method for photoelectric encoder | |
| CN105424043B (en) | It is a kind of based on judging motor-driven estimation method of motion state | |
| CN110390399A (en) | A kind of efficient heuristic approach of intensified learning | |
| CN107819810A (en) | Adaptive planning formulates system | |
| CN109657922A (en) | Marine ecology suitability evaluation assembly apparatus | |
| CN113095389B (en) | Method and system for calibrating monitoring network of gridding sensor | |
| Patterson et al. | Comparative performance of ADAPT and Laurec–Shepherd methods for estimating fish population parameters and in stock management | |
| CN110211670A (en) | Index prediction technique, device, electronic equipment and storage medium | |
| CN108549221A (en) | A kind of filtering method and relevant apparatus of linear stochaastic system | |
| CN108764523A (en) | Predictive Methods of Road Accidents based on unbiased nonhomogeneous gray model and geneva model | |
| CN113095513A (en) | Double-layer fair federal learning method, device and storage medium | |
| CN103925939B (en) | The calibration compensation method of a kind of photoelectric coding device gamut | |
| CN115687854B (en) | High-precision soil sample parameter measuring method and system thereof | |
| CN116933034A (en) | Method, system, electronic equipment and medium for determining carbon reserve change of ecosystem | |
| CN115239567B (en) | Automobile collision dummy model scaling method | |
| Stringer | Implementing approximate Bayesian inference using adaptive quadrature: The AGHQ package |
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 |