CN109061068B - The fault-tolerant measurement estimation method of cabin pollutant concentration - Google Patents
The fault-tolerant measurement estimation method of cabin pollutant concentration Download PDFInfo
- Publication number
- CN109061068B CN109061068B CN201810989793.2A CN201810989793A CN109061068B CN 109061068 B CN109061068 B CN 109061068B CN 201810989793 A CN201810989793 A CN 201810989793A CN 109061068 B CN109061068 B CN 109061068B
- Authority
- CN
- China
- Prior art keywords
- node
- sensor
- consistency
- matrix
- moment
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
Landscapes
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Physics & Mathematics (AREA)
- Combustion & Propulsion (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Arrangements For Transmission Of Measured Signals (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The present invention provides a kind of fault-tolerant measurement estimation methods of cabin pollutant concentration, it include: that sensor for measuring pollutant concentration is set in the predeterminated position of cockpit, and the sensor is constituted into wireless sensor network, the sensor includes the master reference and auxiliary sensor of the measuring node of composition, the predeterminated position of the cockpit includes window, cockpit top, bottom, one or more positions in seat, and the measuring node as composed by master reference and auxiliary sensor is for carrying out estimation processing to metrical information；The measurement result for obtaining the sensor in wireless sensor network, including obtaining the measurement result of master reference and the measurement result of auxiliary sensor；The optimal estimation of practical pollutant concentration is obtained using fault-tolerant measurement estimation method.The fault-tolerant measurement estimation method of cabin pollutant concentration provided by the invention can have preferable fault-tolerant estimated capacity in the case where there is interference, path loss.
Description
Technical field
The present invention relates to the processing method of the monitoring data of pollutant in cabin, in particular to a kind of cabin pollutant concentration
Fault-tolerant measurement estimation method.
Background technique
The fast development of civil aviaton's industry makes the cabin ambient of safety and comfort as aircraft designers, the important pass of operator
One of note point, and passenger and crew member etc. are also to the environment of cabin that more stringent requirements are proposed.In narrow closed cabin
Once situations such as interior generation mechanical breakdown, pipeline breaking or seal failure will cause in cabin, air is contaminated, makes passenger and unit
Personnel generate undesirable somatic reaction, such as dizziness, headache, ear disease, xerophthalmia and have sore throat, serious meeting
Lead to neurological dysfunction, these seriously threaten flight safety.In addition, world energy sources day is becoming tight, energy-saving and emission-reduction
It is one of the significant consideration of modern aircraft development.Therefore, the design of following passenger plane will will use higher air recycling
Rate, and this is also that Cabin air quality guarantee brings bigger challenge.
Only with smog fire condition in single-sensor monitoring cabin in the cockpit of service aircraft, but due to oversensitiveness
Or easily there is false alarm under the interference of extraneous finely ground particles, it makes preparation for dropping so as to cause aircraft, causes the damage of a large amount of manpower financial capacities
It loses.If simply increasing alarm threshold value, safe hidden danger will increase.It mainly lands for the monitoring of Cabin contamination object at present
Detection afterwards, it is difficult to the monitoring Cabin contamination object concentration of real-time online.And the monitoring Cabin contamination object concentration of real-time online is not only
Certain energy consumption is needed, and tends not to detect accurate pollutant in time under observation packet loss and path loss
Burst size.
Summary of the invention
In order to solve problem above, the present invention provides the fault-tolerant measurement estimation method of cabin pollutant concentration, guaranteeing
The method that can be accurately estimated pollutant state under the premise of warning sensitivity provides support for accurate alarm.This
The fault-tolerant measurement estimation method of the cabin pollutant concentration provided is provided, comprising:
The sensor for measuring pollutant concentration is set in the predeterminated position of cockpit, and the sensor is constituted wirelessly
Sensing network, the sensor include the master reference and auxiliary sensor of the measuring node of composition, the predeterminated position of the cockpit
Including one or more positions in window, cockpit top, bottom, seat；It is surveyed as composed by master reference and auxiliary sensor
Amount node is for carrying out estimation processing to metrical information；
The measurement result for obtaining the sensor in wireless sensor network, the measurement result z including obtaining master referencePiWith
The measurement result z of auxiliary sensorSi；
The optimal estimation z of practical pollutant concentration is obtained using fault-tolerant measurement estimation methodOi。
Preferably, described to obtain the optimal estimation z of practical pollutant concentration using fault-tolerant measurement estimation methodOi, comprising:
By the measurement result z of master referencePiWith the measurement result z of auxiliary sensorSiDifference ziBy consistency, Kalman is filtered
Wave device obtains the optimal estimation of the error of master reference, auxiliary sensor
With the optimal estimation of the errorThe measured value for correcting master reference, obtains the optimal of practical pollutant concentration and estimates
Count zOi；
Wherein, the consistency Kalman filter, filtering algorithm are as follows:
Preferably,
The fault-tolerant measurement estimation method is the fault-tolerant measurement estimation method of low transmission energy consumption；
The fault-tolerant measurement estimation method of the low transmission energy consumption, its consistency Kalman filter, to be touched by event
The triggering of hair mechanism；Wherein, event trigger mechanism triggering event generating function used are as follows:
Wherein, whereinIndicate the last estimated value propagated, giIt is a preset positive scalar；
The consistency Kalman filter triggered by event trigger mechanism, estimation error equation are as follows:
Wherein γijPath loss rate between node i and node j,For the estimated value that the last time propagates, αkIt is 0
Or 1 bi-distribution, meet P { αk=1 }=μ, Ci,kFor consistency gain matrix；
The consistency Kalman filter triggered by event trigger mechanism, optimum gain equation are as follows:
Preferably,
The consistency Kalman filter is the expansible filter of low complex degree, filtering algorithm are as follows:
Wherein, Ci,kFor consistency gain matrix.
Preferably,
The consistency Kalman filter, information filter form are as follows:
Wherein,
Some beneficial effects of the invention may include:
The fault-tolerant measurement estimation method of cabin pollutant concentration provided by the invention, under the premise of guaranteeing warning sensitivity
The method that can be accurately estimated pollutant state provides support for accurate alarm.Moreover, cabin provided by the invention
The preferred embodiment of the fault-tolerant measurement estimation method of pollutant concentration, additionally it is possible in transmission failure, the appearance of operative sensor
Still there is preferable fault-tolerant ability in the case where interference, path loss etc., be capable of the dense of normal measuring machine Cabin contamination object
Degree, and the data traffic of entire testing scheme is also reduced, communication burden is greatly reduced, while also ing save data biography
Energy consumption needed for defeated, processing.
Other features and advantages of the present invention will be illustrated in the following description, also, partly becomes from specification
It obtains it is clear that understand through the implementation of the invention.The objectives and other advantages of the invention can be by written explanation
Specifically noted structure is achieved and obtained in book, claims and attached drawing.
Below by drawings and examples, technical scheme of the present invention will be described in further detail.
Detailed description of the invention
Attached drawing is used to provide further understanding of the present invention, and constitutes part of specification, with reality of the invention
It applies example to be used to explain the present invention together, not be construed as limiting the invention.In the accompanying drawings:
Fig. 1 is a kind of flow chart of the fault-tolerant measurement estimation method of cabin pollutant concentration in the embodiment of the present invention；
Fig. 2 is the schematic diagram of output calibration data fusion in the embodiment of the present invention；
Fig. 3 is a certain sensor node Digital Simulation schematic diagram in the embodiment of the present invention；
Fig. 4 is the cockpit pollutant monitoring network topological diagram in the embodiment of the present invention with 10 nodes；
Fig. 5 is activation threshold value s=0.05 in the embodiment of the present invention, the state averaged power spectrum error change figure under no packet loss；
Fig. 6 is activation threshold value s=0.05 in the embodiment of the present invention, the state averaged power spectrum error change under packet loss 0.8
Figure；
Fig. 7 is activation threshold value s=0.05 in the embodiment of the present invention, the lower state average homogeneity error change of packet loss 0.8
Figure；
Fig. 8 is that activation threshold value is 0.05 lower averagely triggering times comparison diagram in the embodiment of the present invention；
Fig. 9 is averaged power spectrum error comparison diagram under activation threshold values different in the embodiment of the present invention；
Figure 10 is average triggering times comparison diagram under activation threshold values different in the embodiment of the present invention；
Figure 11 is averaged power spectrum error comparison diagram under different packet loss arrival rate in the embodiment of the present invention；
Figure 12 is averaged power spectrum error comparison diagram under path loss rates different in the embodiment of the present invention.
Specific embodiment
Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings, it should be understood that preferred reality described herein
Apply example only for the purpose of illustrating and explaining the present invention and is not intended to limit the present invention.
The fault-tolerant measurement estimation method of cabin pollutant concentration provided by the invention, comprising:
Step S1, the sensor for measuring pollutant concentration is set in the predeterminated position of cockpit, and by the sensor
Wireless sensor network is constituted, the sensor includes the master reference and auxiliary sensor of the measuring node of composition, the cockpit
Predeterminated position includes window, cockpit top, bottom, one or more positions in seat；By master reference and auxiliary sensor institute
The measuring node of composition is for carrying out estimation processing to metrical information；
Step S2, the measurement result for obtaining the sensor in wireless sensor network, the measurement knot including obtaining master reference
Fruit zPiWith the measurement result z of auxiliary sensorSi；
Step S3, the optimal estimation z of practical pollutant concentration is obtained using fault-tolerant measurement estimation methodOi。
Method provided by the invention, while fuel oil service efficiency is improved, reduce the discharge amount of carbon dioxide and not to winged
Row weight band carrys out bigger burden.Wireless sensor network (Wireless SensorNetworks, WSN) is used to monitor in main cabin
The variation of pollutant has very big advantage, and the use of wireless sensor network can increase to meet enhancing system redundancy characteristic
Add the requirement of number of sensors, improve robustness, reduces direct cost, maintenance cost etc..In addition, its answering in aircraft system
With being also development trend that future aircraft manufactures.
There is no each pollutant is included in examination range, aircraft handling for civil aviaton's design standard and air worthiness regulation now
Cabin and main cabin can only " Indoor Air Quality standards " (GB/T in reference chamber as the current detection method in workplace and public place
18883-2002) and " public places sanitary standard detecting method " (GB/T 18204-2000).China is directed to aircarrier aircraft at present
Cabin air quality standard only has " public transport sanitary standard " (GB 9673-1996), and a ministerial standard is prevailing for the time being in force, but portion
Divide limit value and detection method not yet in line with international standards.
The present invention places sensor in each node region to monitor different types of pollutant, for same pollutant,
One group of different sensor of measuring principle is chosen, and its measurement result is subjected to data fusion, is estimated with obtaining more accurate parameter
Count result.
Feedback compensation more can accurately reflect the true dynamic process of systematic error state, but consider aviation field equipment because
It is excessively high to reliability requirement therefore most appropriate using output calibration for there are the fastidious of personal safety.By main and auxiliary sensor
Measurement result zPiAnd zSiDifference ziThe optimal estimation of major-minor sensor error is obtained by consistency Kalman filterIt
The measured value that output calibration removes amendment master reference is done with this error estimate again afterwards, finally obtains practical pollutant concentration most
Excellent estimation zOi, schematic diagram is as shown in Fig. 2, the schematic diagram merged for output calibration data in the embodiment of the present invention.
In one embodiment of the invention, the present invention provides the universal model of pollutant monitoring algorithm, does not consider specific
The pollutant monitoring dynamic model of classification.In the schematic diagram of output calibration data fusion shown in Fig. 2, local Kalman filtering
Device carries out data to the measurement error of master reference (Primary Sensor), auxiliary sensor (Secondary Sensor) and melts
It closes, therefore firstly the need of the error dynamics equation for establishing combination sensor.According to system actual performance statistical analysis and sensor
Error analysis, the error model of available each sensor.
Master reference dynamical equation is shown below:
Auxiliary sensor dynamical equation is shown below:
Wherein, xp∈Rmp,xs∈RmsFor the state vector of main and auxiliary sensor；zpi∈Rp,zsi∈RsFor main and auxiliary sensor
Output vector；(Ap,Bp,Hpi,Fpi) and (As,Bs,Hsi,Fsi) be appropriate dimension parameter matrix. wp,vpi,ws,vsiIt is one
Mutually independent white Gaussian noise signal is tieed up, is met:
E[w(k)w(l)T]=Q (k) δkl
E[vi(k)vj(l)T]=Rij(k)δkl
Wherein, δklFor unit impulse function, the i.e. δ as k=lklValue is 1, is otherwise 0；Q and RijFor corresponding noise
Covariance matrix.It is convenient for label, later by RiiIt is abbreviated as Ri。
According to the actual performance and error analysis of system, the global error equation of combination sensor can be obtained by above formula, with
This system equation as local Kalman filter.Since the sampling instant of main and auxiliary sensor cannot be guaranteed completely the same, it is
The systematic observation equation at simplified alignment moment, can ignore the measurement noise of auxiliary sensor, and by the various mistakes of master reference
Difference is classified as the measurement noise of combination sensor.The error equation for obtaining combination sensor is as follows:
zi=zpi-zsi
For convenience of hereafter analyzing, above formula is expressed as Unified Form:
By above equation, when available k, is inscribed, and combined error state vector is xk=col (xp,xs), combined error system
System noise is wk=col (wp,ws)；Assuming that combined error observation noise takes master reference to observe noise, i.e. vi=vpi.Combination misses
Poor system equation parameter is Ak=diag (Ap,AS)；Bk=diag (Bp,BS)；HiFor combined error observation matrix；FiFor failure
Matrix can be actually obtained according to sensor measurement errors principle and engineering.
In practical applications, due to the constraint by the conditions such as itself sensing capability or perception environment, it is sown in difference
All there is certain uncertainty and errors for local message acquired in the same kind sensor node of position.A kind of tradition
Solution be to be weighted and averaged the measurement result of each sensor, but need to be arranged in a data fusion in this way
The heart, and if data fusion center failure, will lead to whole system failure, and program data traffic is big, communication burden
Seriously.In addition, the monitoring of cockpit pollutant is different from general pollutant monitoring, because there is the fastidious of the safety factors such as fire alarm to deposit
, it is therefore desirable to high real-time and accuracy, this just needs stringent monitoring system.In one embodiment of the present of invention
In, using one group of distributed sensor to obtain localized target information, and data are carried out using distributed kalman filter algorithm
Fusion.Consistency policy is a kind of important method of distributed kalman filter algorithm, it can make the estimated value of each sensor
Consistently approach true value.The consistency Kalman filter, filtering algorithm are as follows:
Each sensor node need not carry out information exchange with all nodes in algorithm, it is only necessary to adjacent several sensors
Node intercourses information, so that it may reach the Uniform estimates of global data.By intermediate variable MiIt substitutes intoAnd it willIt substitutes intoThe filtering algorithm of available following more compact form.
In one embodiment of the invention, described to obtain practical pollutant concentration most using fault-tolerant measurement estimation method
Excellent estimation zOi, comprising:
By the measurement result z of master referencePiWith the measurement result z of auxiliary sensorSiDifference ziBy consistency, Kalman is filtered
Wave device obtains the optimal estimation of the error of master reference, auxiliary sensor
With the optimal estimation of the errorThe measured value for correcting master reference, obtains the optimal of practical pollutant concentration and estimates
Count zOi；
Wherein, the consistency Kalman filter, filtering algorithm are as follows:
In one embodiment of the invention, the fault-tolerant measurement estimation method is the fault-tolerant survey of low transmission energy consumption
Amount estimation method；
The fault-tolerant measurement estimation method of the low transmission energy consumption, its consistency Kalman filter, to be touched by event
The triggering of hair mechanism；Wherein, event trigger mechanism triggering event generating function used are as follows:
Wherein, whereinIndicate the last estimated value propagated, giIt is a preset positive scalar；
The consistency Kalman filter triggered by event trigger mechanism, estimation error equation are as follows:
Wherein γijPath loss rate between node i and node j,For the estimated value that the last time propagates, αkIt is 0
Or 1 bi-distribution, meet P { αk=1 }=μ, Ci,kFor consistency gain matrix；
The consistency Kalman filter triggered by event trigger mechanism, optimum gain equation are as follows:
It since consistency Kalman filtering is time trigger, needs to interact at each moment with neighbours, it is therefore desirable to big
The communication cost of amount.In order to reduce communication cost, event trigger mechanism is designed to replace original time trigger mechanism, for this purpose,
It is necessary to define an event generating function:
WhereinIndicate the last estimated value propagated, giIt is a preset positive scalar.Work as satisfactionWhen, thing
Part triggering, transmits current estimated value, whenWhen, illustrate estimated value at this time and last moment at a distance of smaller, to save energy
It need not transmit, neighbor node continues to use the estimated value of the last propagation in estimation.
In addition, wireless signal has the path loss in transmission under cabin atmosphere, and passed in actual network data
During defeated, due to enchancement factors such as network delay, network congestion or sensor faults, often there is packet loss phenomenon.Packet loss is existing
The precision that the presence of elephant can reduce algorithm estimation even results in estimated value diverging.Packet loss is broadly divided into two kinds: 1) network internal section
The packet loss of communication value between point, the i.e. packet loss of node in network and neighbor node interaction data；2) each nodal test number in network
According to packet loss.Packet loss refers to the probability that packet drop occurs in detection, it is the important of the communication quality of one network of measurement
Index.In order to make the present invention have more practicability, the present invention will take into account under path loss and observation packet loss based on consistency
Distributed filtering algorithm, the algorithm on the basis of consistency Kalman filtering, using event trigger mechanism, it is not necessary in institute
Have and transmit data constantly, therefore reduces transmission power consumption.The energy in addition, algorithm remains unchanged under observation packet loss and path loss
Reach the Uniform estimates of global data, it can be with the burst size of the accurate judgement pollutant whether in limit value.
The consistency Kalman filter of the event triggering proposed in the present invention is as follows:
Wherein γijPath loss rate between node i and node j,For the estimated value that the last time propagates, αkIt is 0
Or 1 bi-distribution, meet P { αk=1 }=μ, Ci,kFor consistency gain matrix.
It is an object of the present invention to minimize square evaluated error:
In order to further analyze, node i is defined in the evaluated error of moment k
With corresponding evaluated error covariance matrix
It is convenient for label, as i=j, enable Pii,k=Pi,k
It is obtained by above formula:
It is then possible to obtain covariance matrix:
Therefore, optimum gain matrix Ki,kIt is the solution of following equation
According to matrix calculus theory, for any two matrix X and Y, following formula is all set up
Therefore:
That is:
Therefore optimum gain are as follows:
It can be seen that consistency item and event triggering item are contained in the optimum gain of filtering algorithm proposed by the invention,
Particularly, when trigger condition threshold value is 0, and path loss and packet loss are 0, optimum gain and consistency Kalman filtering
Optimum gain is identical.
In view of the complexity of calculating and the scalability of algorithm, to the filter of each design of node suboptimum.This hair
Consistency Kalman's sub-optimal filters of bright event triggering, are the expansible filter of low complex degree, filtering algorithm are as follows:
Wherein, Ci,kFor consistency gain matrix.
According to Ci,kThe difference of selection mode, if intermediate variable can be arranged in filter in formula:Wherein Fi,k=I-Ki,kHi,k, then filter of the invention can also be write as information filter
The form of device.The consistency Kalman filter, information filter form are as follows:
Wherein,
In order to illustrate stability of the invention, it was demonstrated that as follows:
The topology diagram of wireless sensor network used in the present invention is defined as G=(U, E, A), wherein U={ v1,
v2,...vnIt is sensor node set；E=U × U is that sensor node information exchange connects line set.The neighbour of i-th of sensor
Occupy node set Ni={ vj∈U|(vi,vj) ∈ E indicate, and the number of its neighbor node is known as the degree of the sensor node
And it is denoted as di=| Ni|.The degree matrix D of topology diagram G is defined as using each sensor node degree as the diagonal matrix of diagonal element,
That is D=diag { d1,...dN}.A=[aij] it is the adjacency matrix for indicating each sensor correspondence, when i-th of sensor and the
When j sensor communicates with each other, aijValue is 1, is otherwise 0.The LaPlacian matrix definition of topology diagram G is L=D-
A illustrates that undirected topology diagram G is connection if matrix L contains a nonzero eigenvalue.
Assuming that 1: sytem matrix is nonsingular always in distributed system.
Assuming that 2: there are positive real numbersSo that algorithm parameter matrix has such as lower boundary:
Wherein matrix I is unit battle array.
It should be noted that this two hypotheses are not harsh.For the discrete distribution sampled by continuous system
Formula system, it is assumed that 1 sets up always.If systems compliant is considerable, Pi,kWith bound.
Lemma 1: for a random process Vk(ξk), if there is real numberAnd 0≤α≤1, so that:
E{Vk(ξk)|ξk-1}≤(1-α)Vk-1(ξk-1)+l
It sets up simultaneously, then the square Bounded Index of the random process, it may be assumed that
And the random process is according to 1 bounded of probability.
In next analysis, lemma 1 be used to judge the stabilization of consistency Kalman filtering algorithm state estimation
Property.For each measuring node, parametric boundary condition is as described in lemma 2 and lemma 3.
Lemma 2: assuming that under the premise of 1 and 2, for each node in distributed discrete system, there are 0 < κ of real numberi<
1, so that:
Wherein:
Lemma 3: assuming that under the premise of 1 and 2, for each node in distributed discrete system, there are real number oi> 0,
(i=1 ..., N) make:
Consider that the Main Conclusions of algorithm stability analysis when noise is as follows:
Theorem 1: considering the wireless sensor network being made of N number of node, touches to distributed discrete time-varying system application affairs
Send out suboptimal Kalman filtering algorithm.Assuming that under the premise of 1 and 2, if initial prediction error bounded, the prediction error of algorithm
Square Bounded Index and according to 1 bounded of probability.
It proves:
Definition prediction error vector firstAnd corresponding block diagonal battle array P=diag { P1,...PN}。
Following liapunov function is constructed as the random process in lemma 1.
According in hypothesis 2It is available
Wherein,Andp=maxp 1,...,p N}.Inequality meets first item in lemma 1
Part, in order to prove the square Bounded Index of random process e, it is necessary to consider the mathematical expectation E of subsequent time liapunov function
{Vk+1(ek+1)}.It is available as follows about e according to the definition of evaluated errori,k+1Dynamical equation.
It is substituted into the definition of liapunov function, is then had
It enables
And it analyzes for convenience, consistency gain is defined as form:
Then have:
Therefore:
Here:
Wherein
E is adjacency matrix, and D is the degree matrix of topology diagram G,
For first item
For Section 2
For Section 3
For Section 4
For Section 5
For Section 6
For Section 7
For Section 8
For Section 9
Therefore:
Here
From lemma 2:
Therefore:
Here:
It carries it into lemma 1, it is known that
In order to meet the 2nd condition in lemma 1, it is necessary to assure as lower inequality is set up:
1:0 < α < 1
2:
For first inequality, it may be assumed that
Above formula can regard the One- place 2-th Order inequality about σ as, and work as σ < σ*When the inequality set up, wherein
For the 2nd inequalityThe inequality is set up known to lemma 3.Therefore, evaluated error e is according to probability 1
Square Bounded Index.
In order to verify algorithm provided by the invention, in one embodiment of the invention, as shown in figure 3, being passed for wherein one
Sensor node Digital Simulation schematic diagram.Main and auxiliary sensor exports the measured value z of pollutant concentration respectivelyPiAnd zSi, the difference between the two
As measurement error zi, distributed consistent filter while also the measurement error z of reception neighbor nodej, triggered through event consistent
Output estimation value after property Kalman filteringTo correct master reference measured value zPiTo obtain final measured value zOi.Emulation is real
It tests and a large amount of independent repeated trials is carried out using Monte Carlo method, using the parametric statistics mean value at each moment to the property of monitoring network
It can be carried out analysis.The following performance indicator of definition.
State averaged power spectrum error (Mean Estimation Error, MEE)
State average homogeneity error (Mean Consensus Error, MCE)
Wherein, subscript k indicates the simulation run moment, and N is node total number.
The dynamic error equation of main and auxiliary sensor is derived from respective governing equation, can also use actual measurement data
System Discrimination is carried out to obtain, it will be closer to reality with the measurement data under consideration noise situations.It is however noted that being
The limitation of identification precision and the embedded device performance etc. of uniting, final main and auxiliary Sensor's Dynamic Error equation can only be to true
The approximation of sensor combinations.
As shown in figure 4, to have the cockpit pollutant monitoring network topological diagram of 10 nodes in the embodiment of the present invention, as
The sensor with 10 nodes of one simulation example of inventive algorithm, deployment detects network.
Consider the Measuring error model of combination sensor are as follows:
Error original state takes x0=(8,12)T, system noise wiWith measurement noise viTaking covariance respectively is 10 Hes
The mutually indepedent white Gaussian noise of 100i interferes, and wherein i is node ID, embodies the different measurement error of each node with this.Appoint
Path loss rate γ between node i of anticipating and jij0.05 is taken, system communication cycle 5ms, Kalman filter initial predicted is missed
Poor matrix takes P0=10I2。
Figure 5-8 is the result of this simulation example, wherein Fig. 5 is activation threshold value s=0.05 in the embodiment of the present invention,
Without the state averaged power spectrum error change figure under packet loss；Fig. 6 is activation threshold value s=0.05 in the embodiment of the present invention, packet loss 0.8
Under state averaged power spectrum error change figure；Fig. 7 is activation threshold value s=0.05 in the embodiment of the present invention, the lower state of packet loss 0.8
Average homogeneity error change figure；Fig. 8 is that activation threshold value is 0.05 lower averagely triggering times comparison diagram in the embodiment of the present invention.
The case where can be seen that under reasonable activation threshold value from Fig. 5,6, do not consider packet loss, ET-KCF can achieve and KCF
Almost consistent estimation effect, and not only serious forgiveness is low for local Kalman filtering, has not also made full use of estimating for surroundings nodes
Evaluation, therefore estimation effect is poor；In the case where considering packet loss, KCF there is the possibility of diverging, and ET-KCF due to
Observation correction term at this time has been abandoned in estimator when packet loss occurs, to reduce evaluated error, and because has had touching
The limitation of clockwork spring part, therefore evaluated error can control within threshold value, therefore no matter whether there is or not under packet drop, ET-KCF is
The stable convergence of energy.
In addition, can visually see from Fig. 7, the ET-KCF and KCF mono- that neighbor node data are estimated have been merged
Cause property error is significantly less than local Kalman filtering, this is because local Kalman filtering does not account for the information of surroundings nodes,
This also explains why single-sensor measurement is easy to happen false alarm, and wireless sensor network data fusion can then reduce wrong report
Alert rate.
Fig. 8 shows that ET-KCF compares local Kalman filtering or this kind of time trigger of consistency Kalman filtering is calculated
Method can considerably reduce the number of transmissions, that is, the consumption and communication for reducing transmission energy are born.It is noted that by
The mechanism of event triggering is used in inventive algorithm, so that can guarantee when measurement packet loss occur causes estimated bias excessive
Error makes algorithm have stronger fault-tolerance within threshold value.
For different activation threshold values, different packet loss and path loss rate, to indicate parameters to the shadow of algorithm
It rings.The result of emulation is as shown in Fig. 9-Figure 12, wherein Fig. 9 is that averaged power spectrum misses under different activation threshold values in the embodiment of the present invention
Poor comparison diagram；Figure 10 is average triggering times comparison diagram under activation threshold values different in the embodiment of the present invention；Figure 11 is that the present invention is real
Apply in example averaged power spectrum error comparison diagram under different packet loss arrival rate；Figure 12 is under path loss rate different in the embodiment of the present invention
Averaged power spectrum error comparison diagram.
Fig. 9, Figure 10 are averaged power spectrum error comparison diagram of the ET-KCF under different activation threshold values and average triggering times pair
Than figure, it can be seen that when activation threshold value is 0, ET-KCF becomes KCF, that is, becomes time trigger；When activation threshold value is larger,
Triggering times are greatly reduced, but evaluated error is also got higher simultaneously；It, then can reduction touching by a relatively large margin in reasonable activation threshold value
Number is sent out, while evaluated error also remains and degree of precision similar in KCF.
Figure 11, Figure 12 are averaged power spectrum error comparison of the ET-KCF under different path loss rates and under different arrival rate
Figure.It can be seen from figure 11 that the presence of packet loss has no effect on convergence, but to make mentioned algorithm meet practical need
It asks and converges on a lesser value, then packet loss need to control in reasonable numerical value, if packet loss is excessive, evaluated error circle
It is worth excessive.Influence of the path loss rate to consistency algorithm as can be seen from Figure 12, because consistency algorithm will integrate neighbours' section
The information of point, excessively high path loss rate will lead to the received neighbor information of node and actual information error is excessive, lead to algorithm
Diverging, therefore in practical applications, it needs through rational deployment sensor node with the control path proportion of goods damageds.
In practical flight, pollutant is distributed widely in aircraft cockpit, only establishes perfect pollutant monitoring network,
It is capable of the physiological health of effective guarantee passenger and crew.The present invention is directed to the monitoring of airline carriers of passengers Cabin contamination object concentration,
Combination sensor monitoring model based on wireless sensor network proposes a kind of consistency Kalman filtering algorithm of event triggering.
The algorithm makes up the big problem of single-sensor measurement error using multiple Combined structure of sensor of output calibration, and in this base
On plinth, the case where measuring packet loss and transmission path loss is taken into account.Compared with traditional local Kalman filter, what is mentioned is secondary
Excellent event triggering consistency Kalman filter has higher monitoring accuracy, and the conformity error between each sensor substantially drops
It is low, efficiently solve the problems, such as false alarm；Compared with consistency Kalman filter, can achieve under suitable activation threshold value with
The almost the same precision of consistency Kalman filter, and in the case where there is packet loss, consistency Kalman filter can dissipate,
The energy stable convergence of this algorithm；Compared with time trigger class algorithm, this algorithm significantly reduces the consumption of transmission energy, reduces logical
News burden.
It should be understood by those skilled in the art that, the embodiment of the present invention can provide as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, it wherein includes the computer of computer usable program code that the present invention, which can be used in one or more,
The shape for the computer program product implemented in usable storage medium (including but not limited to magnetic disk storage and optical memory etc.)
Formula.
The present invention be referring to according to the method for the embodiment of the present invention, the process of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that can be realized by computer program instructions each in flowchart and/or the block diagram
The combination of process and/or box in process and/or box and flowchart and/or the block diagram.It can provide these computers
Processor of the program instruction to general purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices
To generate a machine, so that generating use by the instruction that computer or the processor of other programmable data processing devices execute
In the dress for realizing the function of specifying in one or more flows of the flowchart and/or one or more blocks of the block diagram
It sets.
These computer program instructions, which may also be stored in, is able to guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works, so that it includes referring to that instruction stored in the computer readable memory, which generates,
Enable the manufacture of device, the command device realize in one box of one or more flows of the flowchart and/or block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device, so that counting
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, thus in computer or
The instruction executed on other programmable devices is provided for realizing in one or more flows of the flowchart and/or block diagram one
The step of function of being specified in a box or multiple boxes.Obviously, those skilled in the art can carry out the present invention various
Modification and variation is without departing from the spirit and scope of the present invention.In this way, if these modifications and changes of the present invention belongs to this hair
Within the scope of bright claim and its equivalent technologies, then the present invention is also intended to include these modifications and variations.
Claims (3)
1. a kind of fault-tolerant measurement estimation method of cabin pollutant concentration characterized by comprising
The sensor for measuring pollutant concentration is set in the predeterminated position of cockpit, and the sensor is constituted into wireless sensing
Network, the sensor include master reference and auxiliary sensor, and the predeterminated position of the cockpit includes window, cockpit top, bottom
One or more positions in portion, seat；
The measurement result for obtaining the sensor in wireless sensor network, the measurement result z including obtaining master referencePiIt is auxiliary with obtaining
The measurement result z of sensorSi；
By the measurement result z of master referencePiWith the measurement result z for obtaining auxiliary sensorSiDifference ziBy consistency, Kalman is filtered
Wave device obtains the optimal estimation of the error of master reference, auxiliary sensor
With the optimal estimation of the errorThe measured value for correcting master reference, obtains the optimal estimation of practical pollutant concentration
zOi；
The consistency Kalman filter, filtering algorithm are as follows:
The consistency Kalman filter, is triggered by event trigger mechanism；Wherein, event trigger mechanism triggering event used
Generator function are as follows:
Wherein, whereinIndicate the last estimated value propagated, giIt is a preset positive scalar；
The consistency Kalman filter triggered by event trigger mechanism, estimation error equation are as follows:
Wherein γijPath loss rate between node i and node j,For the estimated value that the last time propagates, αkIt is 0 or 1
Bi-distribution meets P { αk=1 }=μ, Ci,kFor consistency gain matrix；
The consistency Kalman filter triggered by event trigger mechanism, optimum gain equation are as follows:
AkFor sytem matrix,Estimated value for node i at the k moment,
N is connection matrix, is
NiFor the neighbor node set of node i,It is node j in k moment estimated value, Pi,kEstimation for node i at the k moment misses
Poor covariance matrix, Ri,kObservation error covariance matrix for node i at the k moment,For node i
In the evaluated error covariance matrix at k+1 moment, Qi,kSystem noise covariance matrix for node i at the k moment,Zi,kObservation for node i at the k moment, Hi,kFor observing matrix, BkAnd Fi,kFor the matrix of suitable parameter, use
In the various forms of noise vectors of characterization, vkAnd wkFor mutually independent white Gaussian noise signal；
It is the state estimation that recently blazes abroad of the node j at the k moment
γirPath loss rate between node i and node r, wherein r ∈ Ni s∈Nj
R belongs to the node of i neighbours, and s belongs to the neighbours of j
Wherein
When for r and j, as
When for i and i, asBy arranging, it is abbreviated as When for r
When with i, as Pri,k
It is the status predication value that recently blazes abroad of the node j at the k moment.
2. the method as described in claim 1, which is characterized in that
The consistency Kalman filter, filtering algorithm are as follows:
Wherein, Ci,kFor consistency gain matrix.
3. method according to claim 2, which is characterized in that
The consistency Kalman filter, information filter form are as follows:
Wherein,
For predicted value of the state x at the k moment of node i,
For predicting covariance matrix
Weighted measures
Information matrix
For consistency matrix parameter
zi,kFor node i state x the k moment observation.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810930323 | 2018-08-15 | ||
| CN2018109303239 | 2018-08-15 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109061068A CN109061068A (en) | 2018-12-21 |
| CN109061068B true CN109061068B (en) | 2019-05-21 |
Family
ID=64756350
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810989793.2A Active CN109061068B (en) | 2018-08-15 | 2018-08-28 | The fault-tolerant measurement estimation method of cabin pollutant concentration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109061068B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106841518A (en) * | 2016-12-29 | 2017-06-13 | 东南大学 | A kind of flue gas NOx concentration measuring method based on Kalman filtering |
| CN106878375A (en) * | 2016-12-22 | 2017-06-20 | 中国民航大学 | A kind of cockpit pollutant monitoring method based on distribution combination sensor network |
| CN107249167A (en) * | 2017-04-10 | 2017-10-13 | 沈磊 | Indoor comprehensive locating platform and localization method |
-
2018
- 2018-08-28 CN CN201810989793.2A patent/CN109061068B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106878375A (en) * | 2016-12-22 | 2017-06-20 | 中国民航大学 | A kind of cockpit pollutant monitoring method based on distribution combination sensor network |
| CN106841518A (en) * | 2016-12-29 | 2017-06-13 | 东南大学 | A kind of flue gas NOx concentration measuring method based on Kalman filtering |
| CN107249167A (en) * | 2017-04-10 | 2017-10-13 | 沈磊 | Indoor comprehensive locating platform and localization method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109061068A (en) | 2018-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106878375B (en) | A kind of cockpit pollutant monitoring method based on distributed combination sensor network | |
| CN108306756B (en) | Holographic evaluation system based on power data network and fault positioning method thereof | |
| Bhuiyan et al. | Dependable structural health monitoring using wireless sensor networks | |
| CN109829468B (en) | Bayesian network-based civil aircraft complex system fault diagnosis method | |
| CN105808366B (en) | A kind of System Safety Analysis method based on four variate models | |
| CN106156913A (en) | Health control method for aircraft department enclosure | |
| CN103281779B (en) | Based on the radio frequency tomography method base of Background learning | |
| CN111601269A (en) | Event trigger Kalman consistency filtering method based on information freshness judgment | |
| US20120203517A1 (en) | Method of determining the influence of a variable in a phenomenon | |
| Wang et al. | Online fault-tolerant dynamic event region detection in sensor networks via trust model | |
| CN106875613B (en) | Fire alarm situation analysis method | |
| CN109815149A (en) | It is a kind of to be distributed the software reliability prediction for introducing failure based on Weibull | |
| CN108508458B (en) | Unmanned aerial vehicle GPS positioning fault detection and reconstruction method based on inter-aircraft ranging | |
| CN109061068B (en) | The fault-tolerant measurement estimation method of cabin pollutant concentration | |
| CN107505126B (en) | Airborne product flight test testability evaluation method | |
| Moroni et al. | Performance improvement for optimization of the non-linear geometric fitting problem in manufacturing metrology | |
| Wang et al. | Multisensor-weighted fusion algorithm based on improved AHP for aircraft fire detection | |
| CN109861859A (en) | The Agent system fault detection method of comprehensive judgement is tested based on frontier inspection | |
| Li et al. | Abnormal data detection in sensor networks based on DNN algorithm and cluster analysis | |
| Bourdenas et al. | Towards self-healing in wireless sensor networks | |
| Wang et al. | Track fusion based on threshold factor classification algorithm in wireless sensor networks | |
| CN104080108B (en) | A kind of variable thresholding abnormal point detecting method for radio sensing network data | |
| Boracchi et al. | A cognitive monitoring system for contaminant detection in intelligent buildings | |
| Do | Statistical detection and isolation of cyber-physical attacks on SCADA systems | |
| Dehabadi et al. | Reliability modeling of anomaly detection algorithms for Wireless Body Area Networks |
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 |