CN108960334A - A kind of multi-sensor data Weighted Fusion method - Google Patents
A kind of multi-sensor data Weighted Fusion method Download PDFInfo
- Publication number
- CN108960334A CN108960334A CN201810765485.1A CN201810765485A CN108960334A CN 108960334 A CN108960334 A CN 108960334A CN 201810765485 A CN201810765485 A CN 201810765485A CN 108960334 A CN108960334 A CN 108960334A
- Authority
- CN
- China
- Prior art keywords
- data
- sensor
- fusion
- group
- amendment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Indication And Recording Devices For Special Purposes And Tariff Metering Devices (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a kind of multi-sensor data Weighted Fusion methods for distributed multi-sensor detection system, this method is filtered amendment to the actual measurement data of each sensor first, improve the confidence level of single-sensor measurement data, secondly based on the data fusion weight between data acquisition sensor after amendment, improve the reasonability of data fusion weight distribution, the Weighted Fusion of multi-sensor data is finally carried out using amendment data and based on data fusion weight between the sensor for correcting data acquisition, therefore, the distributed multi-sensor fusion results obtained using data weighting fusion method of the invention are closer to actual conditions, confidence level is higher.
Description
Technical field
The invention belongs to information to merge field, more particularly to a kind of data fusion of sensor.
Background technique
In distributed multi-sensor detection system, multiplicating survey is carried out to same parameters commonly using multiple sensors
Amount, the detection accuracy of system is improved by the method for data fusion.Method about data fusion is very more, compared to pattra leaves
Subjective random bigger and modeling limitation, weighting are melted in the integration technologies practice such as this decision and neural network
Conjunction method is with it without prior information and the relatively high advantage of fusion accuracy obtains extensive concern.However it senses in the prior art more
Device data weighting fusion method is all the data fusion carried out between sensor using sensor raw measurement data, and there is no consider
Sensor raw measurement data is filtered, causes fused data reliability not high.Therefore, how further to mention
Confidence level after high data fusion is that those skilled in the art are badly in need of technical problems to be solved.
Summary of the invention
The invention mainly solves the technical problem of providing a kind of multi-sensor data Weighted Fusion methods, solve existing skill
The not high technical problem of data fusion method confidence level in existing distributed multi-sensor detection system in art.
In order to solve the above technical problems, one technical scheme adopted by the invention is that: a kind of weighting of multi-sensor data is melted
Conjunction method, including step S1 are filtered amendment to each sensing data in the multiple sensor, obtain the multiple
Data after the amendment of sensor；Step S2 is calculated in the multiple sensor between different sensors between the sensor of data fusion
Merge weight；Step S3 carries out secondary weighted fusion based on weight is merged after the amendment between data and the sensor, to obtain
Take final fusion results.
In another embodiment of multi-sensor data Weighted Fusion method of the present invention, to each of the multiple sensor
Sensing data is filtered amendment, and data include: after obtaining the amendment of the multiple sensor
Step S21, setting sensor dataFor the measurement number of i-th of sensor jth time in the multiple sensor
According to wherein 1≤i≤N, N are the number of the multiple sensor, 1≤j≤M, M are each sensor in the multiple sensor
Overall measurement number；Obtain all master data groups of each sensor in the multiple sensorWherein basic group of j-th of i-th of sensor of B (i, j) expression；
Step S22 calculates the mean value EB (i, j) and variance DB (i, j) of each master data group B (i, j)；
Step S23 calculates each element in the master data group B (i, j) and goes out in the master data group B (i, j)
The existing interior probability P B (i, j) of primary group=[pb (i, 1), pb (i, 2) ..., pb (i, j)]；
Step S24, according to the master data group B (i, j) and it is primary group described in probability P B (i, j) to i-th sense
The measurement data of device jth timeIt is modified, obtains data X=[X after the amendment of all the sensors1,X2,...,XN], wherein
Xi=[xi(1),xi(2),…,xiIt (M)] is data after the amendment of i-th of sensor.
It is described to calculate in the multiple sensor in another embodiment of multi-sensor data Weighted Fusion method of the present invention
Merging weight between different sensors between the sensor of data fusion includes:
Step S31 calculates the mean value EX of data after the amendment of each sensor in the multiple sensoriAnd variance
DXi；
Step S32 calculates fusion weight W=[w between the sensor of different sensors data fusion1,w2,…,wN], whereinIndicate the data fusion weight of i-th of sensor.
In another embodiment of multi-sensor data Weighted Fusion method of the present invention, it is described based on data after the amendment and
Weight is merged between the sensor and carries out secondary weighted fusion, includes: to obtain final fusion results
Step S41 will merge weight and carry out one between data and the sensor after the amendment of the multiple sensor
Secondary fusion obtains Single cell fusion data group Z=[z1,z2,…,zM]=WXT；
Step S42 obtains the mean value EZ and variance DZ of the Single cell fusion data group Z；
Step S43 calculates each element in the Single cell fusion data group Z and occurs in the Single cell fusion data group Z
Secondary group in probability P Z=[pz1,pz2,…,pzM]；
Step S44, by the Single cell fusion data group Z and it is secondary group described in probability P Z carry out secondary data fusion, obtain
Obtain final fusion results Y=(PZ) Z of data fusionT。
The beneficial effects of the present invention are: the invention discloses a kind of more sensings for distributed multi-sensor detection system
Device data weighting fusion method, this method are filtered amendment to the actual measurement data of each sensor first, improve patrilineal line of descent with only one son in each generation
The confidence level of sensor measurement data improves data and melts secondly based on the data fusion weight between data acquisition sensor after amendment
Close weight distribution reasonability, finally using amendment data and based on amendment data acquisition sensor between data fusion weight into
The Weighted Fusion of row multi-sensor data, therefore, the distributed multi-sensor obtained using data weighting fusion method of the invention
For device fusion results closer to actual conditions, confidence level is higher.
Detailed description of the invention
Fig. 1 is the flow chart of an embodiment of multi-sensor data Weighted Fusion method of the present invention；
Fig. 2 is the flow chart for the embodiment that the present invention is filtered modified method to sensing data；
Fig. 3 is that the present invention calculates the side for merging weight in multiple sensors between different sensors between the sensor of data fusion
The flow chart of one embodiment of method；
Fig. 4 is the flow chart of an embodiment of the method that the present invention carries out secondary weighted fusion to multi-sensor data；
Fig. 5 is the flow chart of another embodiment of multi-sensor data Weighted Fusion method of the present invention.
Specific embodiment
To facilitate the understanding of the present invention, in the following with reference to the drawings and specific embodiments, the present invention will be described in more detail.
A better embodiment of the invention is given in the attached drawing.But the invention can be realized in many different forms, and unlimited
In this specification described embodiment.On the contrary, purpose of providing these embodiments is makes to the disclosure
Understand more thorough and comprehensive.
It should be noted that unless otherwise defined, all technical and scientific terms used in this specification with belong to
The normally understood meaning of those skilled in the art of the invention is identical.Used term in the description of the invention
It is the purpose in order to describe specific embodiment, is not intended to the limitation present invention.
Fig. 1 is the flow chart of an embodiment of multi-sensor data Weighted Fusion method of the present invention.In Fig. 1, the side
Method includes:
Step S1 is filtered amendment to each sensing data in the multiple sensor, obtains the multiple biography
Data after the amendment of sensor.In this step, based on the measurement data of some sensor, respectively to all surveys of the sensor
Amount data are modified, and obtain data after the amendment of all measurements of the sensor, and so on, obtain the amendment of all the sensors
Data afterwards.
Step S2 is calculated in the multiple sensor and is merged weight between the sensor of data fusion between different sensors.?
In the step, data after the amendment of all the sensors obtained based on step 1 calculate the power of data fusion between different sensors
Weight.
Step S3 carries out secondary weighted fusion based on weight is merged after the amendment between data and the sensor, to obtain
Take final fusion results.In this step, data and step S2 are obtained after the amendment based on the obtained all the sensors of step S1
Different sensors between data fusion sensor between merge weight, carry out the secondary fusion of data, obtain final fusion results.
From the foregoing it can be that data weighting fusion method of the invention, the first actual measurement to each sensor
Data are filtered amendment, the confidence level of single-sensor measurement data are improved, secondly based between data acquisition sensor after amendment
Data fusion weight, improve data fusion weight distribution reasonability, finally using amendment data and based on amendment data obtain
Therefore the Weighted Fusion that data fusion weight carries out multi-sensor data between the sensor taken utilizes data weighting of the invention
For the Multi-sensor Fusion result that fusion method obtains closer to actual conditions, confidence level is higher.
Fig. 2 is the flow chart for the embodiment that the present invention is filtered modified method to sensing data.In Fig. 2,
The described method includes:
Step S21, setting sensor dataFor the measurement number of i-th of sensor jth time in the multiple sensor
According to wherein 1≤i≤N, N are the number of the multiple sensor, 1≤j≤M, M are each sensor in the multiple sensor
Overall measurement number；Obtain all master data groups of each sensor in the multiple sensorWherein basic group of j-th of i-th of sensor of B (i, j) expression；
Step S22 calculates the mean value EB (i, j) and variance DB (i, j) of each master data group B (i, j)；
Step S23 calculates each element in the master data group B (i, j) and goes out in the master data group B (i, j)
The existing interior probability P B (i, j) of primary group=[pb (i, 1), pb (i, 2) ..., pb (i, j)]；
Step S24, according to the master data group B (i, j) and it is primary group described in probability P B (i, j) to i-th sense
The measurement data of device jth timeIt is modified, obtains data X=[X after the amendment of all the sensors1,X2,...,XN], wherein
Xi=[xi(1),xi(2),…,xiIt (M)] is data after the amendment of i-th of sensor.
Specifically, in the present invention, if there is the mutually independent sensor in the portion N to be measured right to certain in certain distributed detection system
As parameter measures, measurement equation are as follows:
In formula (1),For the measured value of parameters of i-th sensor kth time, xiFor the true value of parameter, δiIt (k) is the
The measurement noise of the portion i sensor kth time, obedience mean value is zero, variance isGaussian Profile, i.e.,Its
Middle k >=1.
And by formula (1) it is found that it is x that the measurement data of the sensor, which obeys mean value,i, variance beGaussian Profile, i.e.,
For the measurement data of i-th sensor kth timeIt will be before kth time and k of the kth time sensor
Measurement dataIt is for statistical analysis as one group of data.
Concept according to probability density function is it is found that i-th sensor jth time measurement dataIt is true with parameter to be measured
Value xiAnd measurement noise criteria difference σiThere are following relationships:
Before i-th sensor kth time can be obtained according to Maximum-likelihood estimation theory and k of the kth time sensor surveys
Measure dataMean value and standard deviation Maximum-likelihood estimation are as follows:
In view of k is a limited numerical value, by σiMaximum-likelihood estimation modification are as follows:
Then, it calculates before i-th sensor kth time and k measurement data of the kth time sensorIn
The probability that each measurement data occurs in all measurement data in this group.
According to normpdf, as k > 1, i-th secondary measured value of sensor jth (1≤j≤k)
The probability of appearance are as follows:
In formula,Before respectively i-th sensor kth time and k measuring value of the kth time sensorMean value and standard deviation.
Then i-th sensor jth time measurement dataThe probability occurred in all measurement data in this group are as follows:
Finally, being modified to the measurement data of i-th sensor kth time, and repeat the above steps, until by i-th
All measurement data of sensor and the measurement data of all the sensors are all modified.
Fig. 3 is that the present invention calculates the side for merging weight in multiple sensors between different sensors between the sensor of data fusion
The flow chart of one embodiment of method.In Fig. 3, this method comprises:
Step S31 calculates the mean value EX of data after the amendment of each sensor in the multiple sensoriAnd variance
DXi；
Step S32 calculates fusion weight W=[w between the sensor of different sensors data fusion1,w2,…,wN], whereinIndicate the data fusion weight of i-th of sensor.
Specifically, it is assumed that i-th sensor obtains M measurement data altogether at the appointed time(M > 1), warp
It crosses after above-mentioned steps, correction value isThe then mean value and variance of i-th sensor M amendment data are as follows:
According to Weighted Fusion thought, if the data of i-th sensor are Xi, weight ωi, then all portions N sensor
The result of Weighted Fusion are as follows:
It enablesThen ZiStandardized normal distribution is obeyed, fusion results may be expressed as: in formula (9)
From formula (10) it can be seen that fusion results Y obeys mean valueVarianceNormal state point
Cloth.It can be obtained using least square method, whenWhen, the variance of fusion results Y is minimum
Fig. 4 is the flow chart of an embodiment of the method that the present invention carries out secondary weighted fusion to multi-sensor data, should
Method includes:
Step S41 will merge weight and carry out one between data and the sensor after the amendment of the multiple sensor
Secondary fusion obtains Single cell fusion data group Z=[z1,z2,…,zM]=WXT；
Step S42 obtains the mean value EZ and variance DZ of the Single cell fusion data group Z；
Step S43 calculates what each element in the Single cell fusion data group Z occurred in the Single cell fusion data Z
Probability P Z=[pz in secondary group1,pz2,…,pzM]；
Step S44, by the Single cell fusion data group Z and it is secondary group described in probability P Z carry out secondary data fusion, obtain
Obtain final fusion results Y=(PZ) Z of data fusionT。
Specifically, firstly, correcting the weight of data fusion between data and different sensors according to formula using sensor
(11) it is weighted fusion.
Then, using fused data as one group of composition Single cell fusion data group, mean value, the variance of the data group are calculated, with
And each element carries out probability in secondary group of data fusion in the group, is then weighted the most termination that fusion obtains data fusion
Fruit.Wherein, each element carries out the method for probability in secondary group of data fusion and calculates some sensor in described calculating group
The method that each element probability of occurrence in organizing is calculated when the M times measurement data amendment is identical.
Specifically, the invention also discloses use the multi-sensor data Weighted Fusion method to carry out insulating box parameter
One embodiment of data fusion.Assuming that 3 thermocouple sensors carry out 6 to insulating box in certain distributed multi-sensor detection system
Secondary detection, wherein table 1 is multiple measured values of each sensor to insulating box.
6 measured values of each sensor of table 1 to insulating box
(1) different sensors are currently measured and historical measurement data is as a master data group, calculate the sensor number
According to the mean value and variance of group.
For sensor S1:
There is no measurement data before measuring due to first time, the mean value of sensor S1 the first master data group is surveyed for it
Magnitude 899.5, variance 0；
Have before being measured due to second and measures for the first time, it will be for the first time with second of measured value as the second master data
Group can be obtained by formula (3) and formula (4), and the mean value of sensor S1 the second master data group is 902.4, variance 16.82；
Similarly, the mean value and variance of the master data group of sensor S1 third, the four, the 5th and the 6th can be calculated.
Table 2 is the mean value and variance of six master data groups of sensor S1.
The mean value and variance of table 2 sensor S1, six master data groups
According to above step, the mean value of the first to the 6th master data group of sensor S2 and sensor S3 can be calculated
And variance.Table 3 is the mean value and variance of six master data groups of sensor S2 and sensor S3.
The mean value and variance of table six master data groups of 3 sensor S2 and sensor S3
(2) on the basis of step (1), each measurement data is calculated in the corresponding master data group of the sensor in group
The probability occurred in all measurement data.
For sensor S1:
There is no measurement data before measuring due to first time, the variance of the first master data group is 0, therefore, is measured for the first time
The probability that data occur in all measurement data in the first master data group is 1；
Have before being measured due to second and measure for the first time, can be obtained according to formula (5) and formula (6), the second base of sensor S1
The probability that first time and second of measurement data occur in all measurement data in the second master data group in notebook data group
For
Similarly, each measurement data in the master data group of sensor S1 third, the four, the 5th and the 6th can be calculated
The probability occurred in all measurement data in respective sets, table 4 are each measurement data in six master data groups of sensor S1
The probability occurred in all data in respective sets.
Each measurement data owns in respective sets in table 4 sensor S1, six master data groups
The probability occurred in data
According to above step, can calculate each in the first to the 6th master data group of sensor S2 and sensor S3
The probability that measurement data occurs in all data in respective sets, wherein table 5 is each in six master data groups of sensor S2
The probability that measurement data occurs in all data in respective sets, table 6 are each measurements in six master data groups of sensor S3
The probability that data occur in all data in respective sets.
The probability that each measurement data occurs in all data in respective sets in table 5 sensor S2, six master data groups
The probability that each measurement data occurs in all data in respective sets in table 6 sensor S3, six master data groups
(3) on the basis of step (2), the current measurement data in the sensor master data group is modified,
And steps 1 and 2 are repeated, until being modified to all measurement data of all the sensors.
For sensor S1:
It can be obtained according to formula (7), measuring revised data for the first time is 899.5；Revised data are measured for the second time
For 899.5 × 0.5+905.3 × 0.5=902.4.
Similarly, sensor S1 third time can be calculated, measure revised number the 4th time, the 5th time and the 6th time
Be according to, table 7 all measurement data of sensor S1 amendment after data.
Data after the amendment of 7 all measurement data of sensor S1 of table
According to above step, the amendment data that sensor S2 and sensor S3 are measured every time can be calculated, table 8 senses
Data after the amendment of device S2 and all measurement data of sensor S3.
Data after the amendment of 8 sensor S2 of table and all measurement data of sensor S3
(4) on the basis of step (3), the weight of data fusion between different sensors is calculated.
According to formula (8), the mean value and variance of all amendment data of different sensors are calculated, table 9 is different sensors institute
There are the mean value and variance of data after correcting.
The mean value and variance of data after all amendments of 9 different sensors of table
| Sensor | S1 | S2 | S3 |
| Mean value | 901.0857 | 890.8015 | 899.6151 |
| Variance | 1.1856 | 20.4217 | 3.4289 |
Then fusion weight is between the sensor of sensor S1 participation data fusionPower is merged between the sensor of sensor S2 participation data fusion
Weight isMelt between the sensor of sensor S3 participation data fusion
Closing weight is
(5) on the basis of step (4), after being corrected using sensor between data and different sensors data fusion biography
Weight is merged between sensor is weighted fusion.
Fused data group can be obtained according to formula (11) are as follows:
(6) on the basis of step (5), each member in the mean value, variance and data group Y of fused data group Y is calculated
Element carries out probability in secondary group of data fusion, is then weighted the final result that fusion obtains data fusion.
According to formula (3) and formula (4) can obtain the mean value of fused data group Y be 900.2982, variance 1.0241.
It is to melt that each data probability of occurrence in all data, table 10 in data Y, which can be obtained, according to formula (5) and formula (6)
Probability in each data occur in all data in organizing in conjunction data group Y secondary group, wherein in calculating fused data group Y
The method of probability and the 6th master data in sensor S1, S2, S3 is calculated in each data occur in all data secondary group
The method of probability is consistent in the primary group that each data occur in group in group.
Probability in each data occur in all data in organizing in 10 fused data group of table secondary group
| Data | 898.7607 | 901.6072 | 900.9032 | 900.7955 | 900.0160 | 899.7063 |
| Probability | 0.0738 | 0.1013 | 0.1956 | 0.2073 | 0.2250 | 0.1971 |
The final result that data fusion can be obtained according to formula (9) is 900.3587.
Fig. 5 is the flow chart of another embodiment of multi-sensor data Weighted Fusion method of the present invention.It is described in Fig. 5
Method includes:
All history before each sensor current measurement data and the current measurement data are measured number by step S51
According to as one basic group, the mean value and variance for organizing measurement data described in the sensor substantially are calculated.
Specifically, it is assumed that the mutually independent sensor in the portion N measures certain image parameter to be measured, measurement equation are as follows:
In formula (12),For the measured value of parameters that i-th sensor kth time obtains, xiFor the true value of parameter, δi
It (k) is the measurement noise of i-th sensor kth time, obedience mean value is zero, variance isGaussian Profile, i.e.,Wherein k >=1.
By formula (12) it is found that it is x that the measured value of the sensor, which obeys mean value,i, variance beGaussian Profile, i.e.,
For the measured value of i-th sensor kth timeIt will be before kth time and k of the kth time sensor surveys
MagnitudeIt is for statistical analysis as a master data group.
Concept according to probability density function can obtain, i-th sensor jth time measurement dataIt is true with parameter to be measured
Value xiAnd measurement noise criteria difference σiThere are following relationships:
Before i-th sensor kth time can be obtained according to Maximum-likelihood estimation theory and k of the kth time sensor surveys
Measure dataMean value and standard deviation Maximum-likelihood estimation are as follows:
In view of k is a limited numerical value, by σiMaximum-likelihood estimation modification are as follows:
Step S52 calculates each measurement data all measurement numbers in described basic group in organizing substantially described in the sensor
According to the probability of middle appearance.
According to normpdf, as k > 1, i-th secondary measured value of sensor jth (1≤j≤k)
The probability of appearance are as follows:
In formula,Before respectively i-th sensor kth time and k measuring value of the kth time sensorMean value and variance.
Then i-th sensor jth time measurement dataThe probability occurred in all measurement data in this group are as follows:
Step S53 is modified the interior current measurement data of this group of sensor, obtains the sensor and currently measure
Revised data.
On the basis of above-mentioned steps, the measurement data of i-th sensor kth time is modified.
Step S54, calculate different sensors between it is described it is current measurement amendment after data sensor between merge weight.
By the k correction value that i-th sensor kth is before secondary and kth is secondaryAs one group, according to formula
(14) and formula (15) can obtain before i-th sensor kth time and the mean value and variance of k correction value of kth time are as follows:
According to Weighted Fusion thought, if the data of i-th sensor are Xi, weight ωi, then all portions N sensor
The result of Weighted Fusion are as follows:
It enablesThen ZiStandardized normal distribution is obeyed, fusion results may be expressed as: in formula (20)
From formula (21) it can be seen that fusion results Y obeys mean valueVarianceNormal state point
Cloth.It can be obtained using least square method, whenWhen, the variance of fusion results Y is minimum
Step S55, described currently measure based on different sensors merge power between revised data and the sensor
It is weighted fusion again, is currently measured revised fused data.
Using the weight of kth time data fusion carries out between data and different sensors after the amendment of sensor kth time measurement
The Weighted Fusion of kth time.
Step S56, judges whether all measurement data of sensor have all been corrected, if it is not, then will be under sensor
One is not corrected measurement data as current measurement, executes step S51；It is no to then follow the steps S57.
The fused data group of data after all measurement amendments is combined into Single cell fusion data group, calculates described one by step S57
Probability in each data occur in all data in the Single cell fusion data group in secondary fused data group secondary group, is based on
The Single cell fusion data group and it is secondary group described in probability, carry out secondary weighted fusion, obtain the final result of data fusion.
From the foregoing it can be that multi-sensor data Weighted Fusion method of the invention, first to each sensor
Actual measurement data is filtered amendment, the confidence level of single-sensor measurement data is improved, secondly based on data acquisition after amendment
Data fusion weight between sensor improves the reasonability of data fusion weight distribution, finally using amendment data and based on repairing
Data fusion weight carries out the Weighted Fusion of multi-sensor data between the sensor that correction data obtains, therefore, using of the invention
For the Multi-sensor Fusion result that data weighting fusion method obtains closer to actual conditions, confidence level is higher.
Specifically, the invention also discloses use the multi-sensor data Weighted Fusion method to carry out insulating box parameter
One embodiment of data fusion.Assuming that 3 thermocouple sensors carry out 6 to insulating box in certain distributed multi-sensor detection system
Secondary detection, the table 1 are 6 measured values of each sensor to insulating box.Then use multi-sensor data Weighted Fusion of the present invention
Method includes:
(1) using different sensors current measurement data and historical measurement data as a master data group, sensing is calculated
The mean value and variance of the device master data group.
First time is measured:
There is no measurement data before measuring due to first time, the mean value of sensor S1 the first master data group is surveyed for it
Magnitude 899.5, variance 0；The mean value of sensor S2 the first master data group is its measured value 898.3000, variance 0；Sensing
The mean value of device S3 the first master data group is its measured value 896.7000, variance 0；At this point, first can be obtained according to formula (22)
The fusion value of secondary measurement is 898.1667.
Second is measured:
Have before being measured due to second and measures for the first time, it will be for the first time with second of measurement data as the second master data
Group can be obtained according to formula (14) and formula (15), and the mean value of sensor S1 the second master data group is 902.4, variance is
The mean value of 16.82, sensor S2 the second master data group is 887.1000, variance 250.8800, and sensor S3 second is basic
The mean value of data group is 901.7500, variance 51.0050.
(2) it on the basis of step (1), calculates and organizes interior a certain measurement data described in the sensor substantially described basic
The probability occurred in all measurement data in group.
Sensor S1 can be obtained for the first time according to formula (16) and formula (17) and second of measured value is in the second basic number
It is respectively according to the probability occurred in all data in organizingWith
The probability that sensor S2 first time and second of measured value occur in all data in the second master data group is 0.5 respectively
With 0.5, sensor S3 is for the first time and the probability point that occurs in all data in the second master data group of second of measured value
It is not 0.5 and 0.5.
(3) on the basis of step (2), the current measurement data in the basic group described in the sensor is modified.
It is 899.5 × 0.5+905.3 × 0.5 that the correction value of second of measurement data of sensor S1, which can be obtained, according to formula (18)
The correction value of second of measurement data of=902.4, sensor S2 is 887.1000, the amendment of second of measurement data of sensor S3
Value is 901.7500.
(4) on the basis of step (3), the fusion of data is weighed after the current measurement amendment between calculating different sensors
Weight.
The fusion weight that data after second of measurement of sensor S1 is corrected can be obtained according to formula (20) is 0.7159, sensor S2
The fusion weight of data is 0.0480 after second of measurement amendment, the fusion weight of data after second of measurement amendment of sensor S3
It is 0.2361.
(5) on the basis of step (4), described based on different sensors currently measures revised data and described
Fusion weight is weighted fusion, is currently measured revised Single cell fusion data.
Second of the three sensors fusion results for measuring data after amendment can be obtained according to formula (22) are as follows:
Y (2)=0.7159 × 902.4+0.048 × 887.1+0.2361 × 901.75=901.5122
(6) steps 1 and 2,3,4,5 are repeated, until completing the fusion of all observation data of different sensors.
According to the calculating process of above-mentioned steps 1,2,3,4,5, can obtain three sensors third times, the 4th time, the 5th time with
And after the 6th measurement amendment data fusion results, table 11 is to measure revised fused data 6 times between all the sensors.
Revised fused data is measured 6 times between 11 all the sensors of table
| Pendulous frequency | 1 | 2 | 3 | 4 | 5 | 6 |
| Fused data | 898.1667 | 901.5122 | 900.8111 | 900.8085 | 900.0013 | 899.7063 |
(7) on the basis of step (6), using the revised fused datas of all measurements as Single cell fusion data group,
Probability in secondary group that each data in the Single cell fusion data group occur in all data in organizing is calculated, is based on described one
Probability in each data occur in all data in organizing in secondary fused data group and described group secondary group carries out secondary weighted
Fusion, obtains the final result of data fusion.
The mean value that a fused data group after all measurements are corrected can be obtained according to formula (14) and formula (15) is
900.1677, variance 1.3754.
Each data all data in organizing in the Single cell fusion data group can be obtained according to formula (16) and formula (17)
Probability in secondary group of middle appearance, table 12 occur in organizing secondary for each data in Single cell fusion data group in all data
Probability in group.Wherein in calculating the Single cell fusion data group each data in described group the calculation method of probability of occurrence and
Calculate the method one for the probability that each data occur in all data in organizing in the 6th master data group in sensor S1, S2, S3
It causes.
Probability in each data occur in all data in organizing in 12 Single cell fusion data group of table secondary group
| Data | 898.1667 | 901.5122 | 900.8111 | 900.8085 | 900.0013 | 899.7063 |
| Probability | 0.0531 | 0.1181 | 0.1960 | 0.1963 | 0.2256 | 0.2109 |
It is 900.3372 that final fused data, which can be obtained, according to formula (18).
In summary, the invention discloses a kind of multi-sensor data Weighted Fusion methods, first to each sensor
Actual measurement data be filtered amendment, improve the confidence level of single-sensor measurement data, obtained secondly based on data after amendment
It takes the data fusion weight between sensor, improves the reasonability of data fusion weight distribution, finally using amendment data and be based on
Therefore the Weighted Fusion that data fusion weight carries out multi-sensor data between the sensor of amendment data acquisition utilizes the present invention
The obtained Multi-sensor Fusion result of data weighting fusion method closer to actual conditions, confidence level is higher.
The above is only the embodiment of the present invention, are not intended to limit the scope of the invention, all to be said using the present invention
Equivalent structure transformation made by bright book and accompanying drawing content, is applied directly or indirectly in other relevant technical fields, and includes
In scope of patent protection of the invention.
Claims (4)
1. a kind of multi-sensor data Weighted Fusion method, comprising:
Step S1 is filtered amendment to each sensing data in the multiple sensor, obtains the multiple sensor
Amendment after data；
Step S2 is calculated in the multiple sensor and is merged weight between the sensor of data fusion between different sensors；
Step S3 carries out secondary weighted fusion based on weight is merged after the amendment between data and the sensor, to obtain most
Whole fusion results.
2. multi-sensor data Weighted Fusion method according to claim 1, which is characterized in that the multiple sensor
In each sensing data be filtered amendment, data include: after obtaining the amendment of the multiple sensor
Step S21, setting sensor dataFor the measurement data of i-th of sensor jth time in the multiple sensor,
In 1≤i≤N, N be the multiple sensor number, 1≤j≤M, M be the multiple sensor in each sensor total survey
Measure number；Obtain all master data groups of each sensor in the multiple sensorWherein basic group of j-th of i-th of sensor of B (i, j) expression；
Step S22 calculates the mean value EB (i, j) and variance DB (i, j) of each master data group B (i, j)；
Step S23 calculates what each element in the master data group B (i, j) occurred in the master data group B (i, j)
The primary interior probability P B (i, j) of group=[pb (i, 1), pb (i, 2) ..., pb (i, j)]；
Step S24, according to the master data group B (i, j) and it is primary group described in probability P B (i, j) to i-th of sensor jth
Secondary measurement dataIt is modified, obtains data X=[X after the amendment of all measurements of all the sensors1,X2,...,XN],
Wherein Xi=[xi(1),xi(2),…,xiIt (M)] is data after the corresponding amendment of all measurements of i-th of sensor.
3. multi-sensor data Weighted Fusion method according to claim 2, which is characterized in that the calculating is the multiple
Merging weight in sensor between different sensors between the sensor of data fusion includes:
Step S31 calculates the mean value EX of data after the amendment of each sensor in the multiple sensoriWith variance DXi；
Step S32 calculates fusion weight W=[w between the sensor of different sensors data fusion1,w2,…,wN], whereinIndicate the data fusion weight of i-th of sensor.
4. multi-sensor data Weighted Fusion method according to claim 3, which is characterized in that described to be based on the amendment
Weight is merged between data and the sensor afterwards and carries out secondary weighted fusion, includes: to obtain final fusion results
Step S41 will merge weight and once be melted between data and the sensor after the amendment of the multiple sensor
It closes, obtains Single cell fusion data group Z=[z1,z2,…,zM]=WXT；
Step S42 obtains the mean value EZ and variance DZ of the Single cell fusion data group Z；
Step S43 calculates all fused datas in the Single cell fusion data group Z and occurs in the Single cell fusion data group Z
Secondary group in probability P Z=[pz1,pz2,…,pzM]；
Step S44, by the Single cell fusion data group Z and it is secondary group described in probability P Z carry out secondary data fusion, obtain number
According to final fusion results Y=(PZ) Z of fusionT。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810765485.1A CN108960334B (en) | 2018-07-12 | 2018-07-12 | Multi-sensor data weighting fusion method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810765485.1A CN108960334B (en) | 2018-07-12 | 2018-07-12 | Multi-sensor data weighting fusion method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108960334A true CN108960334A (en) | 2018-12-07 |
| CN108960334B CN108960334B (en) | 2021-09-14 |
Family
ID=64483166
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810765485.1A Active CN108960334B (en) | 2018-07-12 | 2018-07-12 | Multi-sensor data weighting fusion method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108960334B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298409A (en) * | 2019-07-03 | 2019-10-01 | 广东电网有限责任公司 | Multi-source data fusion method towards electric power wearable device |
| CN110362779A (en) * | 2019-06-11 | 2019-10-22 | 南京江岛环境科技研究院有限公司 | A kind of multiple dimensioned environmental data fusion method |
| CN111123406A (en) * | 2019-12-25 | 2020-05-08 | 哈尔滨工大雷信科技有限公司 | Handheld meteorological instrument temperature data fitting method |
| CN112003891A (en) * | 2020-07-16 | 2020-11-27 | 山东省网联智能车辆产业技术研究院有限公司 | Multi-sensing data fusion method for intelligent networked vehicle controller |
| CN113011479A (en) * | 2021-03-05 | 2021-06-22 | 北京中医药大学 | Multi-source information fusion method for intelligent manufacturing |
| CN113038411A (en) * | 2021-03-09 | 2021-06-25 | 电子科技大学 | Multi-sensor joint detection method based on decision-level and signal-level data fusion |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030191610A1 (en) * | 2002-03-26 | 2003-10-09 | Hai-Wen Chen | Method and system for multi-sensor data fusion using a modified dempster-shafer theory |
| CN101441718A (en) * | 2008-12-19 | 2009-05-27 | 福建三元达通讯股份有限公司 | Sensor information fuse device and method |
| CN102056192A (en) * | 2010-11-09 | 2011-05-11 | 广东石油化工学院 | WSN (wireless sensor network) intra-network data fusion method based on kernel density estimation and non-parameter belief propagation |
| CN102156476A (en) * | 2011-04-14 | 2011-08-17 | 山东大学 | Intelligent space and nurse robot multi-sensor system and information fusion method of intelligent space and nurse robot multi-sensor system |
| CN103776654A (en) * | 2014-02-21 | 2014-05-07 | 黑龙江省科学院自动化研究所 | Method for diagnosing faults of multi-sensor information fusion |
| CN106815591A (en) * | 2015-11-29 | 2017-06-09 | 西南技术物理研究所 | The raising method of multisensor syste data fusion precision |
| CN106840179A (en) * | 2017-03-07 | 2017-06-13 | 中国科学院合肥物质科学研究院 | A kind of intelligent vehicle localization method based on multi-sensor information fusion |
| CN107273692A (en) * | 2017-06-21 | 2017-10-20 | 电子科技大学 | A kind of theoretical distributed fusion method of the random set of sensor senses limited ability |
-
2018
- 2018-07-12 CN CN201810765485.1A patent/CN108960334B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030191610A1 (en) * | 2002-03-26 | 2003-10-09 | Hai-Wen Chen | Method and system for multi-sensor data fusion using a modified dempster-shafer theory |
| CN101441718A (en) * | 2008-12-19 | 2009-05-27 | 福建三元达通讯股份有限公司 | Sensor information fuse device and method |
| CN102056192A (en) * | 2010-11-09 | 2011-05-11 | 广东石油化工学院 | WSN (wireless sensor network) intra-network data fusion method based on kernel density estimation and non-parameter belief propagation |
| CN102156476A (en) * | 2011-04-14 | 2011-08-17 | 山东大学 | Intelligent space and nurse robot multi-sensor system and information fusion method of intelligent space and nurse robot multi-sensor system |
| CN103776654A (en) * | 2014-02-21 | 2014-05-07 | 黑龙江省科学院自动化研究所 | Method for diagnosing faults of multi-sensor information fusion |
| CN106815591A (en) * | 2015-11-29 | 2017-06-09 | 西南技术物理研究所 | The raising method of multisensor syste data fusion precision |
| CN106840179A (en) * | 2017-03-07 | 2017-06-13 | 中国科学院合肥物质科学研究院 | A kind of intelligent vehicle localization method based on multi-sensor information fusion |
| CN107273692A (en) * | 2017-06-21 | 2017-10-20 | 电子科技大学 | A kind of theoretical distributed fusion method of the random set of sensor senses limited ability |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110362779A (en) * | 2019-06-11 | 2019-10-22 | 南京江岛环境科技研究院有限公司 | A kind of multiple dimensioned environmental data fusion method |
| CN110298409A (en) * | 2019-07-03 | 2019-10-01 | 广东电网有限责任公司 | Multi-source data fusion method towards electric power wearable device |
| CN111123406A (en) * | 2019-12-25 | 2020-05-08 | 哈尔滨工大雷信科技有限公司 | Handheld meteorological instrument temperature data fitting method |
| CN112003891A (en) * | 2020-07-16 | 2020-11-27 | 山东省网联智能车辆产业技术研究院有限公司 | Multi-sensing data fusion method for intelligent networked vehicle controller |
| CN113011479A (en) * | 2021-03-05 | 2021-06-22 | 北京中医药大学 | Multi-source information fusion method for intelligent manufacturing |
| CN113038411A (en) * | 2021-03-09 | 2021-06-25 | 电子科技大学 | Multi-sensor joint detection method based on decision-level and signal-level data fusion |
| CN113038411B (en) * | 2021-03-09 | 2022-04-12 | 电子科技大学 | Multi-sensor joint detection method based on decision-level and signal-level data fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108960334B (en) | 2021-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960334A (en) | A kind of multi-sensor data Weighted Fusion method | |
| CN105633956B (en) | A kind of power distribution network three-phase state method of estimation that modeling is measured based on Spiking neutral nets puppet | |
| CN110263866A (en) | A kind of power consumer load setting prediction technique based on deep learning | |
| CN111415010B (en) | Bayesian neural network-based wind turbine generator parameter identification method | |
| CN117436700B (en) | BIM-based new energy engineering data management system and method | |
| CN107658881A (en) | Voltage stability critical point determination methods based on Thevenin's equivalence method | |
| CN108985373B (en) | Multi-sensor data weighting fusion method | |
| CN104102836B (en) | A kind of quick robust state estimation method of electric system | |
| CN110060342B (en) | Three-dimensional curved surface fitting method | |
| CN112803402B (en) | Radiation network forward-push back substitution robust state estimation method containing bad data preprocessing | |
| CN104776827B (en) | The Detection of Gross Errors method of GPS height anomaly data | |
| Kong et al. | A remote estimation method of smart meter errors based on neural network filter and generalized damping recursive least square | |
| CN104407273A (en) | Electric energy quality disturbance source positioning method considering monitoring reliability | |
| CN103886193A (en) | Fuzzy self-adaptation robust estimation method of electric power system | |
| CN110147827A (en) | A kind of failure prediction method based on IAALO-SVM and similarity measurement | |
| CN117351166B (en) | Land management drawing and measuring system based on big data | |
| CN104393588B (en) | A kind of optimum PMU method for arranging based on least mean-square error | |
| CN111999691A (en) | Error calibration method and error calibration device for metering sensor device | |
| CN114492984A (en) | Method, device, equipment and storage medium for predicting time-space distribution of dust concentration | |
| CN1782672A (en) | Method and apparatus for improved fault detection in power generation equipment | |
| CN104899464B (en) | A kind of sampling study machine remote sensing quantitative inversion method under adaptation noise conditions | |
| CN109193639B (en) | Robust estimation method for power system | |
| Wu et al. | Comparing Deep Learning with Statistical Control Methods for Anomaly Detection:(072) | |
| CN114221334B (en) | Rapid state estimation method based on graph neural network | |
| CN108710745A (en) | Electromechanical product reliability assessment method combined with manufacturing process data |
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 |