CN113361562A - Multi-sensor fusion method and device for power battery reaction control module - Google Patents

Multi-sensor fusion method and device for power battery reaction control module Download PDF

Info

Publication number
CN113361562A
CN113361562A CN202110423936.5A CN202110423936A CN113361562A CN 113361562 A CN113361562 A CN 113361562A CN 202110423936 A CN202110423936 A CN 202110423936A CN 113361562 A CN113361562 A CN 113361562A
Authority
CN
China
Prior art keywords
kalman
matrix
covariance matrix
value
error
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110423936.5A
Other languages
Chinese (zh)
Other versions
CN113361562B (en
Inventor
王卓尔
彭峰
邹家喻
郑丽萍
彭斌
周波
高鹏
肖文超
管世辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan Bus Manufacturing Co ltd
Wuhan University of Technology WUT
Original Assignee
Wuhan Bus Manufacturing Co ltd
Wuhan University of Technology WUT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan Bus Manufacturing Co ltd, Wuhan University of Technology WUT filed Critical Wuhan Bus Manufacturing Co ltd
Priority to CN202110423936.5A priority Critical patent/CN113361562B/en
Publication of CN113361562A publication Critical patent/CN113361562A/en
Application granted granted Critical
Publication of CN113361562B publication Critical patent/CN113361562B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/25—Fusion techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computing Systems (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a multi-sensor fusion method for a power battery reaction control module, which comprises the following steps: s1, defining a true value xtAnd the predicted value
Figure DDA0003029085220000011
Error e 'between'tAnd the true value xtAnd the estimated value
Figure DDA0003029085220000012
Error e betweentObtaining a prediction error covariance matrix P'tSum estimation error covariance matrix Pt(ii) a S2, predicting an error covariance matrix P'iSubstituting, expanding, deriving, finally defining and performing data fusion operation to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system; s3, by basing actual sensorsAnd (4) acquiring data, calculating and deducing actual theory, and verifying the effectiveness of the algorithm. The invention also discloses a multi-sensor fusion device for the reaction control module of the power battery, which comprises the following components: an error definition module; a data fusion module; and a validity verification module. The invention can ensure the accuracy and the system reliability of the sensor signal acquisition system and can be widely applied to the field of sensors.

Description

Multi-sensor fusion method and device for power battery reaction control module
Technical Field
The invention relates to the field of sensors, in particular to a multi-sensor fusion method and device for a power battery reaction control module.
Background
A sensor is a device or apparatus that senses a measurement and converts it to a usable signal according to a certain rule. As an important means for information acquisition, the method and the communication technology and the computer technology jointly form three major pillars of the information technology. The world is facing a technical revolution based on information technology, and with the development of modern science, the sensing technology is rapidly developed and applied to various fields as an emerging subject closely related to the modern science. With the rapid development of intelligent technology, the requirements of the sensing capability and the robustness of the multi-sensor system are more and more strict. In order to ensure the reliability of the multi-sensor system, the fault-tolerant function and the input signal precision of the system need to be effectively analyzed, so that the precision and the system reliability of the signals acquired by the sensors are effectively guaranteed after the signals are processed.
Disclosure of Invention
The invention aims to overcome the defects of the background technology and provides a multi-sensor fusion method and a multi-sensor fusion device for a power battery reaction control module, so that the accuracy and the system reliability of a sensor signal acquisition system can be guaranteed.
The invention provides a multi-sensor fusion method for a power battery reaction control module, which comprises the following steps: s1, defining a true value xtAnd the predicted value
Figure BDA0003029085200000011
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000021
Error e betweentTo obtain a prediction error covariance matrix Pt' sum estimation error covariance matrix Pt(ii) a S2, predicting the error covariance matrix PiSubstituting and expanding, deriving, finally defining and performing data fusion operation to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system; and S3, carrying out actual theoretical calculation derivation based on the actual sensor collected data, and verifying the effectiveness of the algorithm.
In the above technical solution, in the step S1, the true value xtAnd the predicted value
Figure BDA0003029085200000022
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000023
Error e betweentRespectively as follows:
Figure BDA0003029085200000024
Figure BDA0003029085200000025
Figure BDA0003029085200000026
wherein Q is a measurement matrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix; the prediction error covariance matrix Pt' sum estimation error covariance matrix PtRespectively as follows:
Figure BDA0003029085200000027
Figure BDA0003029085200000028
wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure BDA0003029085200000029
for prediction of true values, Pt' as prediction error covariance matrix, PtTo estimate an error covariance matrix.
In the above technical solution, the specific process of step S2 is as follows: s21, predicting the error covariance matrix Pt' substitution to give Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T(5) Wherein, K iskalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix; s22, and the nature of Kalman filtering is minimum mean square error estimation, equation (5) is expanded and trace-derived tr (P)t)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6) wherein KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix; s23 optimal estimation KkalmanLet tr (P)t) At a minimum, K can be obtained by applying a derivative on both sides of equation (6) equal to 0kalman＝Pt′QT(QPt′QT+C)-1(7) Wherein, K iskalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and values are taken to be debugged according to the system; by working out the derivation process
Figure BDA0003029085200000031
Pt′＝Pt-KkalmanHPt＝(I-KkalmanH)Pt(9) Wherein, in the step (A),
Figure BDA0003029085200000032
is a predicted value of the true value,
Figure BDA0003029085200000033
is an estimated value of the true value of the image,
Figure BDA0003029085200000034
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix; s24, the calculation process of the Kalman filtering is finally defined as
Figure BDA0003029085200000035
Wherein A is a state transition matrix, B is an input gain matrix,
Figure BDA0003029085200000036
predicted value of the true value of the previous state, ut-1 is the input of the previous state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix; the correction process of the Kalman filter is defined as
Figure BDA0003029085200000037
Wherein z istIn order to be able to take the value of the observation,
Figure BDA0003029085200000038
is a predicted value of the observed value, Q is a measurement matrix,
Figure BDA0003029085200000039
as a prediction of the true value, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix,
Figure BDA00030290852000000310
is an estimated value of the true value of the image,
Figure BDA00030290852000000311
measuring the allowance; and S25, performing data fusion on the data of various sensors to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system.
In the above technical solution, the specific process of step S25 is as follows: s251, assuming that n sensors which are mutually independent exist, wherein the state of the ith sensor is Xi(i 1.. n.) the error of the i-th sensor is MSEi(i 1.. n.) where the data of each sensor is fused by a weighted average fusion method, the data can be obtained
Figure BDA00030290852000000312
Wherein, ω isi(i 1.. n.) is the weight given to each sensor, XaverRepresenting the fused average sensor state; s252, introducing the total mean square error MSE
Figure BDA0003029085200000041
Wherein, ω isiAs a weight on the ith sensor, XaverFor the fused average sensor state, XiOr XjIs the state of the ith or jth sensor; s253, according to the limit theory, the weight corresponding to the minimum total Mean Square Error (MSE) can be obtained as
Figure BDA0003029085200000042
Therein, MSEiMean square error, MSE, for the ith sensorjIs the mean square error of the jth sensor.
The invention also provides a multi-sensor fusion device for the reaction control module of the power battery, which comprises the following parts: an error definition module: defining a true value xtAnd the predicted value
Figure BDA0003029085200000043
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000044
Error e betweentTo obtain a prediction error covariance matrix Pt' sum estimation error covariance matrix Pt(ii) a A data fusion module: will predict the error covariance matrix PiSubstituting and expanding, deriving, finally defining and performing data fusion operation to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system; a validity verification module: and the effectiveness of the algorithm is verified by carrying out actual theoretical calculation derivation based on actual sensor collected data.
In the above technical solution, the error definition module includes the following parts: an error definition unit: the true value xtAnd the predicted value
Figure BDA0003029085200000045
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000046
Error e betweentRespectively as follows:
Figure BDA0003029085200000047
Figure BDA0003029085200000048
Figure BDA0003029085200000049
wherein Q is a measurementMatrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix; error covariance matrix unit: the prediction error covariance matrix Pt' sum estimation error covariance matrix PtRespectively as follows:
Figure BDA00030290852000000410
Figure BDA00030290852000000411
wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure BDA0003029085200000051
for prediction of true values, Pt' as prediction error covariance matrix, PtTo estimate an error covariance matrix.
In the above technical solution, the data fusion module includes the following parts: a prediction error covariance matrix substitution unit: will predict the error covariance matrix Pt' substitution to give Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T(5) Wherein, K iskalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix; a Kalman filtering expansion unit: the nature of Kalman filtering is minimum mean square error estimation, equation (5) is expanded and trace-derived tr (P) is obtainedt)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6) wherein KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix; a derivation unit: optimal estimate KkalmanLet tr (P)t) At a minimum, K can be obtained by applying a derivative on both sides of equation (6) equal to 0kalman＝Pt′QT(QPt′QT+C)-1(7) Wherein, K iskalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and values are taken to be debugged according to the system; by working out the derivation process
Figure BDA0003029085200000052
Pt′＝Pt-KkalmanHPt＝(I-KkalmanH)Pt(9) Wherein, in the step (A),
Figure BDA0003029085200000053
is a predicted value of the true value,
Figure BDA0003029085200000054
is an estimated value of the true value of the image,
Figure BDA0003029085200000055
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix; kalman filter definition unit: the calculation process of the Kalman filtering is finally defined as
Figure BDA0003029085200000056
Wherein A is a state transition matrix, B is an input gain matrix,
Figure BDA0003029085200000057
predicted value of the true value of the previous state, ut-1For the input of the last-state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix; the correction process of the Kalman filter is defined as
Figure BDA0003029085200000061
Wherein z istIn order to be able to take the value of the observation,
Figure BDA0003029085200000062
is a predicted value of the observed value, Q is a measurement matrix,
Figure BDA0003029085200000063
as a prediction of the true value, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix,
Figure BDA0003029085200000064
is an estimated value of the true value of the image,
Figure BDA0003029085200000065
measuring the allowance; a data fusion unit: and data fusion is carried out on data of various sensors to obtain a data fusion value with the minimum mean square error, so that the precision and the reliability of the system are improved.
The invention discloses a multi-sensor fusion method and a device for a power battery reaction control module, which have the following beneficial effects: the invention ensures the reliability of the multi-sensor system by introducing an intelligent algorithm, and effectively ensures the accuracy and the system reliability after the sensor acquires signals by effectively analyzing the fault-tolerant function and the input signal accuracy of the system.
Drawings
FIG. 1 is a schematic flow diagram of a multi-sensor fusion method for a power cell reaction control module according to the present invention;
FIG. 2 is a schematic flow chart of step S2 of the multi-sensor fusion method for a power cell reaction control module according to the present invention;
FIG. 3 is a schematic structural diagram of a multi-sensor fusion device for a power cell reaction control module according to the present invention;
FIG. 4 is a schematic diagram of an error definition module of the multi-sensor fusion apparatus for a power battery reaction control module according to the present invention;
fig. 5 is a schematic structural diagram of a data fusion module in the multi-sensor fusion device for a power battery reaction control module according to the present invention.
Detailed Description
The invention is described in further detail below with reference to the following figures and examples, which should not be construed as limiting the invention.
Referring to fig. 1, the multi-sensor fusion method for a power battery reaction control module of the present invention includes the following steps:
step S1, defining a true value xtAnd the predicted value
Figure BDA0003029085200000071
Error e betweent', true value xtAnd the estimated value
Figure BDA0003029085200000072
Error e betweentAre respectively as
Figure BDA0003029085200000073
Figure BDA0003029085200000074
Wherein Q is a measurement matrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix;
obtaining a prediction error covariance matrix Pt', estimate the error covariance matrix PtAre respectively as
Figure BDA0003029085200000075
Figure BDA0003029085200000076
Wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure BDA0003029085200000077
for prediction of true values, Pt' as prediction error covariance matrix, PtAn estimation error covariance matrix;
referring to fig. 2, step S2, a prediction error covariance matrix P is formediThe data fusion value with the minimum mean square error is obtained through substitution, expansion, derivation, final definition and data fusion operation, so that the precision and the reliability of the system are improved, and the derivation process is as follows:
will predict the error covariance matrix Pt' substitution to give
Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T (5)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
the nature of Kalman filtering is minimum mean square error estimation, and the formula (5) is developed and traced to obtain
tr(Pt)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
optimal estimate KkalmanLet tr (P)t) At a minimum, therefore, deriving both sides of equation (6) to be equal to 0 can be obtained
Kkalman＝Pt′QT(QPt′QT+C)-1 (7)，
Wherein, KkalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and the value is taken and debugged according to the system;
obtained by organizing the derivation process
Figure BDA0003029085200000081
Figure BDA0003029085200000082
Wherein the content of the first and second substances,
Figure BDA0003029085200000083
is a predicted value of the true value,
Figure BDA0003029085200000084
is an estimated value of the true value of the image,
Figure BDA0003029085200000085
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix;
the calculation process of the Kalman filtering is finally defined as
Figure BDA0003029085200000086
Wherein A is a state transition matrix, B is an input gain matrix,
Figure BDA0003029085200000087
predicted value of the true value of the previous state, ut-1For the input of the last-state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix;
the correction process of the Kalman filter is defined as
Figure BDA0003029085200000091
Wherein z istIn order to be able to take the value of the observation,
Figure BDA0003029085200000092
is a predicted value of the observed value, Q is a measurement matrix,
Figure BDA0003029085200000093
as a prediction of the true value, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix,
Figure BDA0003029085200000094
is an estimated value of the true value of the image,
Figure BDA0003029085200000095
measuring the allowance;
and data fusion is needed to be carried out on data of various sensors, so that a data fusion value with the minimum mean square error is obtained, and the precision and the reliability of the system are improved. It is therefore assumed here that there are n sensors which are independent of one another, the state of the i-th sensor being Xi(i 1.. n.) the error of the i-th sensor is MSEi(i＝1,...,n)，
The data of each sensor are fused by a weighted average fusion method to obtain
Figure BDA0003029085200000096
Wherein, ω isi(i 1.. n.) is the weight given to each sensor, XaverThe average sensor state after fusion is represented, and the distribution of the weight has great influence on the improvement of the overall performance of the system;
the total mean square error MSE is introduced here
Figure BDA0003029085200000097
Wherein, ω isiAs a weight on the ith sensor, XaverFor the fused average sensor state, XiOr XjIs the state of the ith or jth sensor;
according to the limit theory, the weight corresponding to the minimum MSE can be obtained as
Figure BDA0003029085200000101
Therein, MSEiMean square error, MSE, for the ith sensorjMean square error for the jth sensor;
the numerical sensor data with any precision can be fused by the weighting factor obtained by the formula (14), so that the measurement precision of the system is effectively improved;
and step S3, verifying the effectiveness of the algorithm by carrying out actual theoretical calculation derivation based on the actual sensor collected data.
Referring to fig. 3, the multi-sensor fusion device for a power battery reaction control module of the present invention comprises the following parts:
an error definition module: defining a true value xtAnd the predicted value
Figure BDA0003029085200000102
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000103
Error e betweentTo obtain a prediction error covariance matrix Pt' sum estimation error covariance matrix PtAs shown in fig. 4, the following parts are included:
an error definition unit: the true value xtAnd the predicted value
Figure BDA0003029085200000104
Error e betweent', and the true value xtAnd the estimated value
Figure BDA0003029085200000105
Error e betweentRespectively as follows:
Figure BDA0003029085200000106
Figure BDA0003029085200000107
wherein Q is a measurement matrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix;
error covariance matrix unit: the prediction error covariance matrix Pt' sum estimation error covariance matrix PtRespectively as follows:
Figure BDA0003029085200000108
Figure BDA0003029085200000111
wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure BDA0003029085200000112
for prediction of true values, Pt' as prediction error covariance matrix, PtEstimating an error covariance matrix;
a data fusion module: will predict the error covariance matrix Pi' substituting and expanding, deriving, finally defining and data fusing operation to obtain the data fusion value with the minimum mean square error, and improving the precision and reliability of the system, as shown in fig. 5, the following parts are included:
a prediction error covariance matrix substitution unit: will predict the errorCovariance matrix Pt' substitution to give
Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T (5)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
a Kalman filtering expansion unit: the nature of Kalman filtering is minimum mean square error estimation, and the formula (5) is developed and traced to obtain
tr(Pt)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
a derivation unit: optimal estimate KkalmanLet tr (P)t) At a minimum, a derivation of 0 on both sides of equation (6) can be obtained
Kkalman＝Pt′QT(QPt′QT+C)-1 (7)，
Wherein, KkalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and values are taken to be debugged according to the system;
by working out the derivation process
Figure BDA0003029085200000121
Pt′＝Pt-KkalmanHPt＝(I-KkalmanH)Pt (9)，
Wherein the content of the first and second substances,
Figure BDA0003029085200000122
is a predicted value of the true value,
Figure BDA0003029085200000123
is an estimated value of the true value of the image,
Figure BDA0003029085200000124
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix;
kalman filter definition unit: the calculation process of the Kalman filtering is finally defined as
Figure BDA0003029085200000125
Wherein A is a state transition matrix, B is an input gain matrix,
Figure BDA0003029085200000126
predicted value of the true value of the previous state, ut-1For the input of the last-state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix;
the correction process of the Kalman filter is defined as
Figure BDA0003029085200000127
Wherein z istIn order to be able to take the value of the observation,
Figure BDA0003029085200000128
is a predicted value of the observed value, Q is a measurement matrix,
Figure BDA0003029085200000129
as a prediction of the true value, KkalmanIs a Kalman gain momentArray, Pt' is a prediction error covariance matrix,
Figure BDA00030290852000001210
is an estimated value of the true value of the image,
Figure BDA00030290852000001211
measuring the allowance;
a data fusion unit: carrying out data fusion on data of various sensors to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system, wherein the data fusion value comprises the following parts;
assume that there are n sensors independent of each other, wherein the i-th sensor has a state Xi(i 1.. n.) the error of the i-th sensor is MSEi(i＝1,...,n)，
The data of each sensor are fused by a weighted average fusion method to obtain
Figure BDA0003029085200000131
Wherein, ω isi(i 1.. n.) is the weight given to each sensor, XaverRepresenting the fused average sensor state;
introducing an overall mean square error MSE of
Figure BDA0003029085200000132
Wherein, ω isiAs a weight on the ith sensor, XaverFor the fused average sensor state, XiOr XjIs the state of the ith or jth sensor;
according to the limit theory, the weight corresponding to the minimum MSE can be obtained as
Figure BDA0003029085200000133
Therein, MSEiMean square error, MSE, for the ith sensorjMean square error for the jth sensor;
a validity verification module: and the effectiveness of the algorithm is verified by carrying out actual theoretical calculation derivation based on actual sensor collected data.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Those not described in detail in this specification are within the skill of the art.

Claims (7)

1. A multi-sensor fusion method for a power battery reaction control module is characterized in that: the method comprises the following steps:
s1, defining a true value xtAnd the predicted value
Figure FDA0003029085190000011
Error e 'between'tAnd the true value xtAnd the estimated value
Figure FDA0003029085190000012
Error e betweentTo obtain a prediction error covariance matrix Pt' sum estimation error covariance matrix Pt；
S2, predicting the error covariance matrix PiSubstituting and expanding, deriving, finally defining and performing data fusion operation to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system;
and S3, carrying out actual theoretical calculation derivation based on the actual sensor collected data, and verifying the effectiveness of the algorithm.
2. The plurality for power cell reaction control module of claim 1A sensor fusion method, characterized by: in the step S1, the true value xtAnd the predicted value
Figure FDA0003029085190000013
Error e 'between'tAnd the true value xtAnd the estimated value
Figure FDA0003029085190000014
Error e betweentRespectively as follows:
Figure FDA0003029085190000015
Figure FDA0003029085190000016
wherein Q is a measurement matrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix;
the prediction error covariance matrix Pt' sum estimation error covariance matrix PtRespectively as follows:
Figure FDA0003029085190000017
Figure FDA0003029085190000018
wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure FDA0003029085190000019
for prediction of true values, Pt' as prediction error covariance matrix, PtFor estimating error co-channelAnd (4) a variance matrix.
3. The multi-sensor fusion method for a power cell reaction control module according to claim 2, characterized in that: the specific process of step S2 is as follows:
s21, predicting the error covariance matrix Pt' substitution to give
Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T (5)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
s22, and the nature of Kalman filtering is minimum mean square error estimation, and the formula (5) is developed and traced to obtain
tr(Pt)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
s23 optimal estimation KkalmanLet tr (P)t) At a minimum, a derivation of 0 on both sides of equation (6) can be obtained
Kkalman＝Pt′QT(QPt′QT+C)-1 (7)，
Wherein, KkalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and values are taken to be debugged according to the system;
by working out the derivation process
Figure FDA0003029085190000021
Pt′＝Pt-KkalmanHPt＝(I-KkalmanH)Pt (9)，
Wherein the content of the first and second substances,
Figure FDA0003029085190000022
is a predicted value of the true value,
Figure FDA0003029085190000023
is an estimated value of the true value of the image,
Figure FDA0003029085190000024
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix;
s24, the calculation process of the Kalman filtering is finally defined as
Figure FDA0003029085190000031
Wherein A is a state transition matrix, B is an input gain matrix,
Figure FDA0003029085190000032
predicted value of the true value of the previous state, ut-1For the input of the last-state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix;
the correction process of the Kalman filter is defined as
Figure FDA0003029085190000033
Wherein z istIn order to be able to take the value of the observation,
Figure FDA0003029085190000034
is a predicted value of the observed value, Q is a measurement matrix,
Figure FDA0003029085190000035
as a prediction of the true value, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix,
Figure FDA0003029085190000036
is an estimated value of the true value of the image,
Figure FDA0003029085190000037
measuring the allowance;
and S25, performing data fusion on the data of various sensors to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system.
4. The multi-sensor fusion method for a power cell reaction control module according to claim 3, characterized in that: the specific process of step S25 is as follows:
s251, assuming that n sensors which are mutually independent exist, wherein the state of the ith sensor is Xi(i 1.. n.) the error of the i-th sensor is MSEi(i＝1,...,n)，
The data of each sensor are fused by a weighted average fusion method to obtain
Figure FDA0003029085190000038
Wherein, ω isi(i 1.. n.) is the weight given to each sensor, XaverRepresenting the fused average sensor state;
s252, introducing the total mean square error MSE
Figure FDA0003029085190000041
Wherein, ω isiAs a weight on the ith sensor, XaverFor the fused average sensor state, XiOr XjIs the state of the ith or jth sensor;
s253, according to the limit theory, the weight corresponding to the minimum total Mean Square Error (MSE) can be obtained as
Figure FDA0003029085190000042
Therein, MSEiMean square error, MSE, for the ith sensorjIs the mean square error of the jth sensor.
5. A multisensor fuses device for power battery reaction control module which characterized in that: the method comprises the following steps:
an error definition module: defining a true value xtAnd the predicted value
Figure FDA0003029085190000043
Error e 'between'tAnd the true value xtAnd the estimated value
Figure FDA0003029085190000044
Error e betweentTo obtain a prediction error covariance matrix Pt' sum estimation error covariance matrix Pt；
A data fusion module: will predict the error covariance matrix PiSubstituting and expanding, deriving, finally defining and performing data fusion operation to obtain a data fusion value with the minimum mean square error, and improving the precision and reliability of the system;
a validity verification module: and the effectiveness of the algorithm is verified by carrying out actual theoretical calculation derivation based on actual sensor collected data.
6. The multi-sensor fusion device for a power cell reaction control module of claim 5, wherein: the error definition module comprises the following parts:
an error definition unit: the true value xtAnd the predicted value
Figure FDA0003029085190000045
Error e 'between'tAnd the true value xtAnd the estimated value
Figure FDA0003029085190000051
Error e betweentRespectively as follows:
Figure FDA0003029085190000052
Figure FDA0003029085190000053
wherein Q is a measurement matrix, KkalmanIs a Kalman gain matrix, vtIs a process noise matrix;
error covariance matrix unit: the prediction error covariance matrix Pt' sum estimation error covariance matrix PtRespectively as follows:
Figure FDA0003029085190000054
Figure FDA0003029085190000055
wherein, E [ v ]tvt T]＝C，KkalmanIs a Kalman gain matrix, Q is a measurement matrix, xtIn order to be the true value of the value,
Figure FDA0003029085190000056
for prediction of true values, Pt' as prediction error covariance matrix, PtTo estimate an error covariance matrix.
7. The multi-sensor fusion device for a power cell reaction control module of claim 6, wherein: the data fusion module comprises the following parts:
a prediction error covariance matrix substitution unit: will predict the error covariance matrix Pt' substitution to give
Pt＝(I-KkalmanQ)Pt′(I-KkalmanQ)T+KkalmanCKkalman T (5)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
a Kalman filtering expansion unit: the nature of Kalman filtering is minimum mean square error estimation, and the formula (5) is developed and traced to obtain
tr(Pt)＝tr(Pt′)-2tr(KkalmanQPt′)+tr(Kkalman(QPt′QT+C)Kkalman T) (6)，
Wherein, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix, Q is a measurement matrix, PtEstimating an error covariance matrix;
a derivation unit: optimal estimate KkalmanLet tr (P)t) At a minimum, a derivation of 0 on both sides of equation (6) can be obtained
Kkalman＝Pt′QT(QPt′QT+C)-1 (7)，
Wherein, KkalmanIs a Kalman gain matrix, PtThe' is a prediction error covariance matrix, Q is a measurement matrix, and values are taken to be debugged according to the system;
by working out the derivation process
Figure FDA0003029085190000061
Pt′＝Pt-KkalmanHPt＝(I-KkalmanH)Pt (9)，
Wherein the content of the first and second substances,
Figure FDA0003029085190000062
is a predicted value of the true value,
Figure FDA0003029085190000063
is an estimated value of the true value of the image,
Figure FDA0003029085190000064
as a prediction of the observed value, KkalmanIs a Kalman gain matrix, H is a process excitation noise covariance matrix, PtTo estimate the error covariance matrix, Pt' is a prediction error covariance matrix;
kalman filter definition unit: the calculation process of the Kalman filtering is finally defined as
Figure FDA0003029085190000065
Wherein A is a state transition matrix, B is an input gain matrix,
Figure FDA0003029085190000066
predicted value of the true value of the previous state, ut-1For the input of the last-state prediction model, Pt' as prediction error covariance matrix, Pt-1Estimating an error covariance matrix for the previous state, and H is a process excitation noise covariance matrix;
the correction process of the Kalman filter is defined as
Figure FDA0003029085190000067
Wherein z istIn order to be able to take the value of the observation,
Figure FDA0003029085190000068
is a predicted value of the observed value, Q is a measurement matrix,
Figure FDA0003029085190000069
as a prediction of the true value, KkalmanIs a Kalman gain matrix, Pt' is a prediction error covariance matrix,
Figure FDA0003029085190000071
is an estimated value of the true value of the image,
Figure FDA0003029085190000072
measuring the allowance;
a data fusion unit: and data fusion is carried out on data of various sensors to obtain a data fusion value with the minimum mean square error, so that the precision and the reliability of the system are improved.
CN202110423936.5A 2021-04-20 2021-04-20 Multi-sensor fusion method and device for power battery reaction control module Active CN113361562B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110423936.5A CN113361562B (en) 2021-04-20 2021-04-20 Multi-sensor fusion method and device for power battery reaction control module

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110423936.5A CN113361562B (en) 2021-04-20 2021-04-20 Multi-sensor fusion method and device for power battery reaction control module

Publications (2)

Publication Number Publication Date
CN113361562A true CN113361562A (en) 2021-09-07
CN113361562B CN113361562B (en) 2024-03-15

Family

ID=77525340

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110423936.5A Active CN113361562B (en) 2021-04-20 2021-04-20 Multi-sensor fusion method and device for power battery reaction control module

Country Status (1)

Country Link
CN (1) CN113361562B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105825239A (en) * 2016-03-31 2016-08-03 中国科学院电子学研究所 Multi-sensor track fusion method based on sparse expression
CN107832575A (en) * 2017-10-10 2018-03-23 中国航空无线电电子研究所 Band feedback maneuvering target Asynchronous Track Fusion based on pseudo-measurement
CN110034559A (en) * 2019-04-18 2019-07-19 南京邮电大学 Electric system Fusion state estimation method based on switching system model
WO2020105812A1 (en) * 2018-11-22 2020-05-28 제주대학교 산학협력단 Prediction system and method on basis of parameter improvement through learning
CN111536967A (en) * 2020-04-09 2020-08-14 江苏大学 EKF-based multi-sensor fusion greenhouse inspection robot tracking method
CN111881955A (en) * 2020-07-15 2020-11-03 北京经纬恒润科技有限公司 Multi-source sensor information fusion method and device

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105825239A (en) * 2016-03-31 2016-08-03 中国科学院电子学研究所 Multi-sensor track fusion method based on sparse expression
CN107832575A (en) * 2017-10-10 2018-03-23 中国航空无线电电子研究所 Band feedback maneuvering target Asynchronous Track Fusion based on pseudo-measurement
WO2020105812A1 (en) * 2018-11-22 2020-05-28 제주대학교 산학협력단 Prediction system and method on basis of parameter improvement through learning
CN110034559A (en) * 2019-04-18 2019-07-19 南京邮电大学 Electric system Fusion state estimation method based on switching system model
CN111536967A (en) * 2020-04-09 2020-08-14 江苏大学 EKF-based multi-sensor fusion greenhouse inspection robot tracking method
CN111881955A (en) * 2020-07-15 2020-11-03 北京经纬恒润科技有限公司 Multi-source sensor information fusion method and device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
SHOHEI NIWA 等: "Kalman Filter with Time-variable Gain for a Multisensor Fusion System", 《IEEE XPLORE》, pages 56 - 61 *
张银: "基于EKF智能车辆多传感器融合定位算法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, no. 7, pages 1 - 54 *

Also Published As

Publication number Publication date
CN113361562B (en) 2024-03-15

Similar Documents

Publication Publication Date Title
CN110602647B (en) Indoor fusion positioning method based on extended Kalman filtering and particle filtering
WO2016155241A1 (en) Method, system and computer device for capacity prediction based on kalman filter
CN105579922A (en) Information processing device and analysis method
CN106878375A (en) A kind of cockpit pollutant monitoring method based on distribution combination sensor network
CN111829505B (en) Multi-sensor track quality extrapolation track fusion method
EP3767551A1 (en) Inspection system, image recognition system, recognition system, discriminator generation system, and learning data generation device
CN108319570B (en) Asynchronous multi-sensor space-time deviation joint estimation and compensation method and device
TWI584134B (en) Method for analyzing variation causes of manufacturing process and system for analyzing variation causes of manufacturing process
CN108009566B (en) Improved PCA (principal component analysis) damage detection method under space-time window
CN102257448B (en) Method and device for filtering signal using switching models
CN110889091A (en) Machine tool thermal error prediction method and system based on temperature sensitive interval segmentation
CN108846200B (en) Quasi-static bridge influence line identification method based on iteration method
CN111679657A (en) Attack detection method and system based on industrial control equipment signals
CN112560981A (en) Training method, apparatus, device, program and storage medium for generating countermeasure model
CN110298409A (en) Multi-source data fusion method towards electric power wearable device
CN111950115A (en) Residual force vector-based interval damage identification method
Chen et al. Two-stage automated operational modal analysis based on power spectrum density transmissibility and support-vector machines
Liu et al. A data‐driven combined deterministic‐stochastic subspace identification method for condition assessment of roof structures subjected to strong winds
CN113361562A (en) Multi-sensor fusion method and device for power battery reaction control module
Liu et al. Distributed state estimation for dynamic positioning systems with uncertain disturbances and transmission time delays
CN112765219B (en) Stream data abnormity detection method for skipping steady region
CN114666525A (en) Audio and video switching verification system based on ASIC structure
CN114004138A (en) Building monitoring method and system based on big data artificial intelligence and storage medium
CN115728383B (en) Bridge structure damage positioning method, device, computer equipment and medium
CN117723782B (en) Sensor fault identification positioning method and system for bridge structure health monitoring

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