CN116972827A - Multi-source information fusion speed measurement method for high-speed train - Google Patents
Multi-source information fusion speed measurement method for high-speed train Download PDFInfo
- Publication number
- CN116972827A CN116972827A CN202310759295.XA CN202310759295A CN116972827A CN 116972827 A CN116972827 A CN 116972827A CN 202310759295 A CN202310759295 A CN 202310759295A CN 116972827 A CN116972827 A CN 116972827A
- Authority
- CN
- China
- Prior art keywords
- speed
- fusion
- model
- state
- measurement
- 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.)
- Pending
Links
- 230000004927 fusion Effects 0.000 title claims abstract description 82
- 238000000691 measurement method Methods 0.000 title abstract description 7
- 238000005259 measurement Methods 0.000 claims abstract description 128
- 238000000034 method Methods 0.000 claims abstract description 76
- 238000001914 filtration Methods 0.000 claims abstract description 71
- 230000007704 transition Effects 0.000 claims abstract description 57
- 238000001514 detection method Methods 0.000 claims abstract description 45
- 230000005856 abnormality Effects 0.000 claims abstract description 32
- 230000002159 abnormal effect Effects 0.000 claims abstract description 28
- 230000001133 acceleration Effects 0.000 claims abstract description 28
- 239000011159 matrix material Substances 0.000 claims description 21
- 230000003044 adaptive effect Effects 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 15
- 238000013178 mathematical model Methods 0.000 claims description 7
- 239000013078 crystal Substances 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 3
- 230000001360 synchronised effect Effects 0.000 claims description 3
- 238000007499 fusion processing Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 6
- 238000012546 transfer Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000546 chi-square test Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/10—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration
- G01C21/12—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning
- G01C21/16—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation
- G01C21/165—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by using measurements of speed or acceleration executed aboard the object being navigated; Dead reckoning by integrating acceleration or speed, i.e. inertial navigation combined with non-inertial navigation instruments
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
- G01S13/50—Systems of measurement based on relative movement of target
- G01S13/58—Velocity or trajectory determination systems; Sense-of-movement determination systems
- G01S13/589—Velocity or trajectory determination systems; Sense-of-movement determination systems measuring the velocity vector
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/93—Radar or analogous systems specially adapted for specific applications for anti-collision purposes
- G01S13/931—Radar or analogous systems specially adapted for specific applications for anti-collision purposes of land vehicles
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/418—Theoretical aspects
Abstract
The invention discloses a multi-source information fusion speed measurement method of a high-speed train, which comprises the steps of obtaining a speed measurement value of a wheel axle speed sensor, a Doppler radar speed measurement value and an accelerometer acceleration measurement value according to calculation of output signals of the wheel axle speed sensor, the Doppler radar and the accelerometer which are synchronously collected; establishing a speed measurement model and a state transfer model; detecting the abnormality of the speed measurement model and the state transition model, judging whether the abnormality exists or not, and distinguishing whether the abnormality comes from the speed measurement model or the state transition model; based on the speed measurement model, the state transition model and the abnormality detection result, the train fusion speed is obtained by an improved Kalman filtering method. The invention can effectively improve the speed measurement precision of the train and has stronger fault tolerance for abnormal measurement data.
Description
Technical Field
The invention relates to the technical field of speed measuring methods of high-speed trains, in particular to the technical field of multi-source information fusion speed measuring methods of high-speed trains.
Background
The high-speed train needs to detect the train speed in real time in the running process, so that on one hand, the train running interval is controlled according to the train speed generated speed-distance curve, and on the other hand, the train speed is monitored in real time to prevent the train from running overspeed. Therefore, the train speed measurement technology plays an important role in the running process of the high-speed train, and the accuracy and reliability of train speed measurement directly influence the running safety and efficiency of the high-speed train.
Currently, the devices used for measuring the speed of the high-speed train mainly comprise a wheel axle speed sensor, a Doppler radar, an accelerometer and the like. The wheel axle speed sensor generates serious deviation in measurement results when the wheel idles or slides; the Doppler radar is sensitive to the application environment and the external weather conditions, so that the speed measurement accuracy is reduced; the accelerometer is easily affected by factors such as train impact and vibration, and the measured value is suddenly changed. The accuracy and the reliability of the single speed measurement mode are limited, and in order to ensure the driving safety, the high-speed train generally adopts various speed measurement devices to measure the speed of the train at the same time, and the accuracy and the reliability of the speed measurement of the train are improved through multi-source information fusion. Therefore, a reasonable multisource information fusion speed measurement method is one of key research problems of a high-speed train.
The prior research mainly realizes information fusion by a weighted average method or a Kalman filtering technology. However, the working principles, error sources and precision characteristics of different speed measuring devices are different, the influences of the factors cannot be fully considered in the existing method, and effective measures cannot be taken for the running scene under the complex condition, so that the speed of the fused train still deviates from the actual condition. In addition, the existing method does not have an abnormality detection function and fault tolerance, and when some equipment has abnormal conditions, if an original fusion strategy is still adopted, the speed measurement precision is seriously reduced, and even the reference value is lost.
Disclosure of Invention
The invention aims to provide a multi-source information fusion speed measurement method for a high-speed train, which is used for solving the problems of precision reduction and poor fault tolerance in the fusion speed measurement process in the prior art.
The technical scheme of the invention is as follows:
a multi-source information fusion speed measurement method for a high-speed train comprises the following steps:
a1, synchronously acquiring output signals of an axle speed sensor, a Doppler radar and an accelerometer which are arranged on a high-speed train, and resolving acquired information to obtain an axle speed sensor speed measurement value, a Doppler radar speed measurement value and an accelerometer acceleration measurement value;
a2, establishing a system mathematical model capable of describing a train motion system by using a state space method, wherein the system mathematical model comprises a speed measurement model consisting of all speed measurement information in the system and a state transition model for reflecting the change process of the train motion state along with time; wherein the speed measurement model is established based on the axle speed sensor speed measurement and the doppler radar speed measurement, and the state transition model is established based on the accelerometer acceleration measurement;
a3, carrying out abnormality detection on the speed measurement model and the state transition model, judging whether an abnormal condition exists, and specifically distinguishing whether the abnormal condition comes from the speed measurement model or the state transition model;
a4, according to the abnormal detection result, different Kalman filtering methods are selected to fuse the results of the speed measurement model and the state transition model, and the fusion speed of the train is obtained;
wherein the different kalman filtering methods include: kalman filtering; a first modified Kalman filtering method and a second modified Kalman filtering method; the first improved Kalman filtering method is to perform robust filtering on the basis of Kalman filtering, and the second improved Kalman filtering method is to perform adaptive filtering on the basis of Kalman filtering.
According to some embodiments of the invention, the synchronous acquisition comprises: and corresponding the output signals of the wheel axle speed sensor, the output signals of the Doppler radar and the time axis of the output signals of the accelerometer, and acquiring the output signals by using the same acquisition period.
According to some embodiments of the invention, the output signal comprises: generating pulse frequency of pulse signals according to wheel rotation obtained by the wheel axle speed sensor; according to the frequency difference between the radar transmitting wave and the receiving wave obtained by the Doppler radar, namely Doppler frequency shift; the voltage generated by the pressure deformation of the piezoelectric crystal inside the accelerometer obtained from the accelerometer, namely the output voltage thereof.
According to some embodiments of the invention, the method for obtaining the wheel axle speed sensor speed measurement, the Doppler radar speed measurement and the accelerometer acceleration measurement comprises the following steps:
wherein ,vodo -providing a speed measurement for said axle speed sensor; d is the diameter of the wheel; f is the pulse frequency output by the axle speed sensor; m is the number of teeth of the wheel axle speed sensor;
wherein ,vrad -providing said doppler radar velocity measurement; c is the speed of light; f (f) d Doppler shift for Doppler radar output; f is the frequency of Doppler radar emission waves; θ is the included angle between the Doppler radar emission wave and the rail surface;
wherein ,aacc -providing an accelerometer acceleration measurement; u is the output voltage of the accelerometer;is the piezoelectric coefficient of the piezoelectric crystal inside the accelerometer.
According to some embodiments of the invention, the state transition model is built as follows:
X(k)＝ΦX(k-1)+BΔa acc (k)+W(k)
wherein k is the current time; x (k) is a state vector, X (k) = [ v (k) a (k)] T V is the speed of the train during movement, and a is the acceleration of the train during movement; phi is a state transition matrix; b is an input control matrix; Δa acc The difference between the acceleration measurements of the accelerometer at times k-1 to k; t is the time step from k-1 to k; w (k) = [ W v (k)w a (k)] T To process noise vector, w v and wa Process noise affecting speed and acceleration, respectively, the covariance matrix of which is denoted as Q; t represents a transpose;
the speed measurement model is built as follows:
Z(k)＝HX(k)+V(k)
where Z (k) is a velocity measurement vector, Z (k) = [ v ] odo (k) v rad (k)] T ，v odo and vrad Respectively representing a wheel axle speed sensor speed measurement value and a Doppler radar speed measurement value; h is a measurement matrix; v (k) = [ delta ] odo (k) δ rad (k)] T Representing the measured noise vector, delta odo and δrad The measurement noise of the wheel axle speed sensor and the Doppler radar are respectively represented by a covariance matrix of the wheel axle speed sensor and the Doppler radar as R.
According to some embodiments of the invention, the anomaly detection comprises: detecting whether the speed measurement model and the state transition model have abnormality; and a fine detection for judging whether an abnormality is from the speed measurement model or the state transition model in the case where an abnormality exists, wherein the whole detection uses a card method inspection method.
According to some embodiments of the invention, the refinement detection comprises:
the detection amount of the speed detection model and the detection amount of the state transition model are constructed as follows:
wherein ,representing the difference between the current time speed measured value and the previous time fusion state value;representing the difference between the predicted state value at the current moment and the fusion state value at the previous moment; t is t 1 Detecting the quantity for a speed measurement model; t is t 2 Detecting the quantity for the state transition model; h is the observation matrix; />The mean square error generated by fusion is the fusion state value which is the estimated value of the train motion state at k moment obtained by fusion processing of the system mathematical modelP(k)；/>P (k-1) is the fusion mean square error of the previous moment; r is a covariance matrix of the measured noise vector; />For passing the fusion state value of the previous moment +.>Predicting the obtained state value at the current moment; />The mean square error generated for the prediction is the predicted mean square error;
comparison t 1 and t2 The numerical value of (1), if t 1 >t 2 If t, determining that the abnormal condition is from the speed measurement model 1 <t 2 It is determined that the abnormal condition is from the state transition model.
According to some embodiments of the invention, the fusion is based on a fusion variance minimization principle to fuse the speed measurement model and the state transition model.
According to some embodiments of the invention, the selecting different kalman filtering methods to fuse the velocity measurement model and the state transition model includes:
when no abnormal condition occurs at the moment k, a Kalman filtering method is selected;
when abnormality occurs in the speed measurement model at the moment k, selecting the first improved Kalman filtering method;
and when the state transition model at the moment k is abnormal, selecting a second improved Kalman filtering method.
According to some embodiments of the invention, the obtaining of the fusion speed of the train comprises:
b1, predicting the state at the current moment by utilizing the fusion state at the previous moment according to the state transition model, and calculating the mean square error of a state prediction result;
b2, selecting a filtering type according to an abnormality detection result, and if the filtering type is selected as a first improved Kalman filtering type, calculating an robust factor; if the second modified Kalman filtering is selected, calculating an adaptive factor; if the Kalman filtering method is selected, setting the robust factor and the adaptive factor to be 1, and enabling the robust factor and the adaptive factor to be in an invalid state;
b3, calculating a filtering gain according to the obtained robust factor and the adaptive factor, and calculating a mean square error of information fusion according to the filtering gain;
and B4, calculating the fusion state of the current moment according to the speed measurement model, the prediction state and the obtained filter gain at the current moment to obtain the train fusion speed.
The invention has the following beneficial effects:
(1) The invention has the function of abnormality detection, and by establishing an abnormality detection function, the information to be fused is subjected to abnormality detection, and the sources of the abnormal conditions are specifically distinguished.
(2) The invention adopts an improved Kalman filtering technology to realize information fusion, adds robust filtering and adaptive filtering on the basis of Kalman filtering, selects corresponding filtering methods aiming at different abnormal conditions, and improves the fault tolerance of the information fusion in a complex operation environment.
Drawings
Fig. 1 is a flowchart of a method for measuring speed by fusion of multisource information of a high-speed train in an embodiment of the invention.
Fig. 2 is a flowchart of the improved kalman filter technique information fusion in an embodiment of the invention.
Detailed Description
The present invention will be described in detail with reference to the following examples and drawings, but it should be understood that the examples and drawings are only for illustrative purposes and are not intended to limit the scope of the present invention in any way. All reasonable variations and combinations that are included within the scope of the inventive concept fall within the scope of the present invention.
Referring to fig. 1, according to the technical scheme of the invention, some specific embodiments of the high-speed train multisource information fusion speed measurement method comprise the following steps:
a1 synchronously acquiring output signals of a wheel axle speed sensor, a Doppler radar and an accelerometer which are arranged on a high-speed train, and resolving acquired information to obtain a wheel axle speed sensor speed measurement value, a Doppler radar speed measurement value and an accelerometer acceleration measurement value.
The output signal refers to an original signal which is detected by the speed measuring equipment and can reflect the motion state of the train, for example, the output signal of the wheel axle speed sensor is the counting frequency of a pulse signal generated by the rotation of wheels; the output signal of the Doppler radar is the frequency difference between the radar transmitting wave and the receiving wave, namely Doppler frequency shift; the output signal of the accelerometer is the voltage generated by the pressure deformation of the internal piezoelectric crystal.
Further, in some embodiments, the synchronous acquisition includes: and corresponding the output signals of the wheel axle speed sensor, the output signals of the Doppler radar and the time axis of the output signals of the accelerometer, and synchronously acquiring the output signals by using the same acquisition period so as to ensure the synchronism among the signals.
Further, in some embodiments, the resolving includes resolving the collected information of the axle speed sensor as follows:
wherein ,vodo -providing a speed measurement for said axle speed sensor; d is the diameter of the wheel; f is the pulse frequency output by the axle speed sensor; m is the number of teeth of the wheel axle speed sensor.
Further, in some embodiments, the resolving includes resolving acquired information of the doppler radar as follows:
wherein ,vrad -providing said doppler radar velocity measurement; c is the speed of light; f (f) d Doppler shift for Doppler radar output; f is the frequency of Doppler radar emission waves; θ is the angle between the Doppler radar emission wave and the rail surface.
Further, in some embodiments, the resolving includes resolving the acquired information of the accelerometer as follows:
wherein, a is acc -providing an accelerometer acceleration measurement; u is the output voltage of the accelerometer;is the piezoelectric coefficient of the piezoelectric crystal inside the accelerometer.
A2, establishing a system mathematical model which can fuse the speed measurement value of the wheel shaft speed sensor, the Doppler radar speed measurement value and the accelerometer acceleration measurement value, wherein the system mathematical model comprises a speed measurement model and a state transfer model.
The state transition model is used for reflecting the change process of the train motion state along with time, namely the relation between the motion state at the previous moment of the train and the motion state at the current moment.
Further, in some embodiments, the establishing the state transition model includes:
taking the speed v and the acceleration a of the train during movement as state parameters, taking the acceleration measured value of an accelerometer as input control quantity, and constructing a state transition model by adopting a state space method, wherein the state transition model comprises the following steps of:
wherein k is the current time; k-1 is the previous time; t is the time step from k-1 to k; Δa acc The difference between acceleration measurements of the accelerometer at times k-1 to k is used to describe the external force effects at times k-1 to k; w (w) v and wa The process noise affecting the speed and the acceleration respectively can be Gaussian white noise with zero mean value, and the noise variance can be set according to specific situations.
Further, in some embodiments, to facilitate subsequent operational representations, the state transition model is represented briefly as:
X(k)＝ΦX(k-1)+BΔa acc (k)+W(k)
wherein X (k) is a state vector, X (k) = [ v (k) a (k)] T V is the speed of the train during movement, and a is the acceleration of the train during movement; phi is a state transition matrix; b is an input control matrix; Δa acc The difference between the acceleration measurements of the accelerometer at times k-1 to k; w (k) = [ W v (k) w a (k)] T To process noise vector, w v and wa Process noise affecting speed and acceleration, respectively, the covariance matrix of which is denoted as Q; t represents the transpose.
Further, in some embodiments, the speed measurement model is derived from axle speed sensor speed measurements v odo And Doppler radar velocity measurement v rad The composition is expressed as:
Z(k)＝[v odo (k) v rad (k)] T
where Z (k) is the velocity measurement vector.
According to the following relation between the speed measurement model and the state transition model:
wherein ,δodo and δrad The measuring noise of the wheel axle speed sensor and the Doppler radar respectively can be zero average in the practical implementationThe value of gaussian white noise, the noise variance can be set according to the device accuracy parameters.
Further, a short expression of the velocity measurement model can be obtained as follows:
Z(k)＝HX(k)+V(k)
wherein H is an observation matrix; v (k) = [ delta ] odo (k) δ rad (k)] T The measurement noise vector is represented, and the covariance matrix of the two is represented as R.
A3, carrying out abnormality detection on the speed measurement model and the state transition model, judging whether an abnormal condition exists, and specifically distinguishing whether the abnormal condition comes from the speed measurement model or the state transition model.
Further, in some embodiments, the anomaly detection is divided into two stages, where the first stage anomaly detection uses an anomaly detection function to perform overall detection on the speed measurement model and the state transition model to determine whether an anomaly exists, and if so, the anomaly detection is performed in a second stage, which is a refined detection on the speed measurement model and the state transition model, and determines whether the anomaly is from the speed measurement model or the state transition model according to a result of the refined detection.
Further, in some embodiments, the overall detection includes:
setting the fusion state value asThe mean square error generated by the fusion is fusion mean square error P (k);
according to the state transition model, the state value is fused at the previous momentPredicting the state value of the current moment +.>I.e. < ->The mean square error generated by the prediction, i.e. the prediction mean square error
When there is an abnormality in accelerometer measurement information, i.e. Δa acc (k) Abnormal state prediction value of state transition modelDeviations will occur;
when the speed of the wheel axle speed sensor is measured v odo (k) Or Doppler radar speed measurement v rad (k) When an abnormality exists, the speed measurement model Z (k) is deviated;
because the probability of deviation of two models at the same time is extremely low, namelyAnd Z (k) are normal, so that the overall test, i.e. +.>More specifically, since the noise of the two models is the process noise and the measurement noise respectively, and is the Gaussian white noise with zero mean value, epsilon (k) follows the Gaussian distribution with zero mean value and the variance is +.>Whether an anomaly has occurred can be determined by examining the epsilon (k) mean.
Further, in some embodiments, epsilon (k) based anomaly detection is performed using a chi-square test, which includes:
the epsilon (k) obeying the gaussian distribution is converted into an anomaly detection function lambda (k) obeying the chi-square distribution as follows:
wherein λ (k) obeys a chi-square distribution with a degree of freedom of m, which is the dimension of Z (k), e.g., in some embodiments taking the value of m as 2;
when Z (k) orWhen an abnormality occurs, lambda (k) is no longer subject to the distribution, so that a detection threshold can be set according to the upper-part point property of chi-square distribution and compared with lambda (k), if lambda (k) is larger than the threshold, the occurrence of the abnormality is judged, otherwise, the abnormality is not considered.
Further, in one embodiment, the upper 10 is distributed in a chi-square -3 If the quantile 13.82 is used as the detection threshold, Z (k) or Z (k) is considered as long as lambda (k) at time k is greater than 13.82There is an abnormality.
Further, the refinement detection includes:
by the difference epsilon between the speed measured value at the current moment and the fusion state value at the previous moment 1 As errors in the velocity measurement model, i.e.
By the difference epsilon between the predicted state value at the current moment and the fused state value at the previous moment 2 As errors in state transition models, i.e.
Based on the two errors, a speed detection model detection quantity and a state transition model detection quantity are constructed as follows
wherein ,t1 Detecting the quantity for a speed measurement model; t is t 2 Detecting the quantity for the state transition model;
comparison t 1 and t2 The numerical value of (1), if t 1 >t 2 If t, determining that the abnormal condition is from the speed measurement model 1 <t 2 It is determined that the abnormal condition is from the state transition model.
And A4, according to the abnormal detection result, selecting different Kalman filtering methods to fuse the results of the speed measurement model and the state transition model, and obtaining the fusion speed of the train.
Wherein, further, the abnormality detection result may be represented by a ternary function f (k), when f (k) =0 indicates that no abnormality occurs at time k; f (k) =1 indicates that abnormality occurs in the velocity measurement model at time k; f (k) =2 indicates that abnormality occurs in the state transition model at time k.
Further, the selecting different kalman filtering methods to fuse the speed measurement model and the state transition model includes:
when f (k) =0, a kalman filter method is selected;
when f (k) =1, selecting a first modified kalman filter method;
when f (k) =2, selecting a second modified kalman filter method;
the first improved Kalman filtering method is to perform robust filtering on the basis of Kalman filtering, and the second improved Kalman filtering method is to perform adaptive filtering on the basis of Kalman filtering.
According to the filtering selection of the specific embodiment, the fault tolerance of information fusion to factors such as speed measurement errors, acceleration measurement errors, abnormal model interference and the like is enhanced, the operand is not greatly increased, and the instantaneity of information fusion is ensured.
According to a specific embodiment of the present invention, further, referring to fig. 2, the obtaining of the fusion speed of the train includes:
b1, utilizing the fusion state of the previous moment according to the state transition modelPredicting state at current timeI.e. < ->And calculates the mean square error of the state prediction result>The calculation formula is as follows:
b2, selecting a filtering type according to the value of the abnormality detection result f (k), and if the filtering type is selected as the first improved Kalman filtering type, calculating an robust factor r (k) is needed; if a second modified Kalman filter is selected, then an adaptive factor alpha (k) needs to be calculated; if the Kalman filtering method is selected, the robust factor r (k) and the adaptive factor alpha (k) are set to be 1, so that the robust factor r (k) and the adaptive factor alpha (k) are in an invalid state.
The robust filtering can reduce the weight of the speed measurement model in information fusion by an robust factor r (k), reduce the influence of abnormal speed measurement information on fusion results, and the expression of r (k) is as follows:
wherein ,is a normalization of ε (k), i.e., ε (k) divided by its standard deviation; k (k) 0 and k1 Is a constant threshold, k 0 Has a value in the range of 1.0 to 2.5, k 1 The value range of (2) is 3.5 to 8.0, and the specific value can be manually selected in the value range.
The weight of the state transfer model in information fusion is reduced through the self-adaptive factor alpha (k), the influence of model inaccuracy factors or abnormal acceleration measurement information on fusion results is reduced, and the expression of alpha (k) is as follows:
wherein e is a natural constant; c is a constant threshold value, the value range is 1.0 to 3.0, and a proper value can be manually selected in the value range; delta epsilon (k) is a statistic constructed from epsilon (k) as follows:
further, in order to avoid interference between filtering methods, when robust filtering is used, α (k) is set to 1, so that the adaptive factor is in an invalid state; when the adaptive filtering is used, r (k) is set to be 1, so that the robust factor is in an invalid state; when Kalman filtering is used, r (k) and alpha (k) are set to 1, so that the robust factor and the adaptive factor are simultaneously in an invalid state.
And B3, calculating a filter gain K (K) according to the robust factor and the adaptive factor, and calculating a mean square error P (K) of information fusion according to the filter gain.
Wherein, the calculation formula of K (K) is:
the calculation formula of P (k) is:
wherein, I is an identity matrix, and the matrix order is the same as the number of the state vector parameters.
B4, predicting the state according to the speed measurement model Z (k) at the current momentAnd the obtained filtering gain, calculating the fusion state +.>And further obtaining the train fusion speed.
wherein ,the calculation formula is as follows:
wherein the fusion stateThe first component of (2) is the train fusion speed.
In the above embodiment, as can be seen from steps B1 to B4, the present invention is a continuous recursive process in selecting different kalman filtering methods, and besides the fusion result at the previous moment needs to be stored, the fusion result and measurement information at the other moments need not to be stored, and the overall calculation amount is smaller, so that the real-time performance of information fusion is ensured.
Example 1
In order to verify the speed information fusion method, a laboratory simulation test platform is utilized to generate train operation data (speed data and acceleration data), the simulation data is used as the actual operation state of the train, and the speed measuring equipment measurement data is simulated on the basis. The measurement data generated by simulation is fused by adopting the fusion speed measuring method, and the experimental results at part of moments are shown in table 1.
TABLE 1 actual speed, measured speed and fusion speed vs. conditions
As can be seen from Table 1, the fusion speed error is obviously smaller than the measurement errors of the two speed measuring devices, and the absolute speed error is controlled within 0.1m/s, so that the validity of the invention in the aspect of information fusion is verified. In addition, at time 103 and time 107, a large error (abnormal condition) occurs in the wheel axle speed sensor and the Doppler radar respectively, but the fusion speed error is not affected, and still high precision is maintained, so that the effectiveness of the invention in fault tolerance is verified.
The above examples are only preferred embodiments of the present invention, and the scope of the present invention is not limited to the above examples. All technical schemes belonging to the concept of the invention belong to the protection scope of the invention. It should be noted that modifications and adaptations to the present invention may occur to one skilled in the art without departing from the principles of the present invention and are intended to be within the scope of the present invention.
Claims (10)
1. The method for measuring the speed of the high-speed train by multi-source information fusion is characterized by comprising the following steps of:
a1, synchronously acquiring output signals of an axle speed sensor, a Doppler radar and an accelerometer which are arranged on a high-speed train, and resolving acquired information to obtain an axle speed sensor speed measurement value, a Doppler radar speed measurement value and an accelerometer acceleration measurement value;
a2, establishing a system mathematical model capable of describing a train motion system by using a state space method, wherein the system mathematical model comprises a speed measurement model consisting of all speed measurement information in the system and a state transition model for reflecting the change process of the train motion state along with time; wherein the speed measurement model is established based on the axle speed sensor speed measurement and the doppler radar speed measurement, and the state transition model is established based on the accelerometer acceleration measurement;
a3, carrying out abnormality detection on the speed measurement model and the state transition model, judging whether an abnormal condition exists, and specifically distinguishing whether the abnormal condition comes from the speed measurement model or the state transition model;
a4, according to the abnormal detection result, different Kalman filtering methods are selected to fuse the results of the speed measurement model and the state transition model, and the fusion speed of the train is obtained;
wherein the different kalman filtering methods include: kalman filtering; a first modified Kalman filtering method and a second modified Kalman filtering method; the first improved Kalman filtering method is to perform robust filtering on the basis of Kalman filtering, and the second improved Kalman filtering method is to perform adaptive filtering on the basis of Kalman filtering.
2. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the synchronous acquisition comprises: and corresponding the output signals of the wheel axle speed sensor, the output signals of the Doppler radar and the time axis of the output signals of the accelerometer, and acquiring the output signals by using the same acquisition period.
3. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the output signals comprise: generating pulse frequency of pulse signals according to wheel rotation obtained by the wheel axle speed sensor; according to the frequency difference between the radar transmitting wave and the receiving wave obtained by the Doppler radar, namely Doppler frequency shift; the voltage generated by the pressure deformation of the piezoelectric crystal inside the accelerometer obtained from the accelerometer, namely the output voltage thereof.
4. The method for measuring speed by multi-source information fusion of high-speed trains according to claim 3, wherein the method for calculating the speed measurement value of the wheel axle speed sensor, the Doppler radar speed measurement value and the accelerometer acceleration measurement value is as follows:
wherein ,vodo -providing a speed measurement for said axle speed sensor; d is the diameter of the wheel; f is the pulse frequency output by the axle speed sensor; m is the number of teeth of the wheel axle speed sensor;
wherein ,vrad -providing said doppler radar velocity measurement; c is the speed of light; f (f) d Doppler shift for Doppler radar output; f is the frequency of Doppler radar emission waves; θ is the included angle between the Doppler radar emission wave and the rail surface;
wherein ,aacc -providing an accelerometer acceleration measurement; u is the output voltage of the accelerometer;is the piezoelectric coefficient of the piezoelectric crystal inside the accelerometer.
5. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the state transition model is established as follows:
X(k)＝ΦX(k-1)+BΔa acc (k)+W(k)
wherein k is the current time; x (k) is a state vector, X (k) = [ v (k) a (k)] T V is the speed of the train during movement, and a is the acceleration of the train during movement; phi is a state transition matrix; b is an input control matrix; Δa acc The difference between the acceleration measurements of the accelerometer at times k-1 to k; t is k-1 toA time step at time k; w (k) = [ W v (k)w a (k)] T To process noise vector, w v and wa Process noise affecting speed and acceleration, respectively, the covariance matrix of which is denoted as Q; t represents a transpose;
the speed measurement model is built as follows:
Z(k)＝HX(k)+V(k)
where Z (k) is a velocity measurement vector, Z (k) = [ v ] odo (k)v rad (k)] T ，v odo and vrad Respectively representing a wheel axle speed sensor speed measurement value and a Doppler radar speed measurement value; h is a measurement matrix;
V(k)＝[δ odo (k)δ rad (k)] T representing the measured noise vector, delta odo and δrad The measurement noise of the wheel axle speed sensor and the Doppler radar are respectively represented by a covariance matrix of the wheel axle speed sensor and the Doppler radar as R.
6. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the anomaly detection comprises: detecting whether the speed measurement model and the state transition model have abnormality; and a fine detection for judging whether an abnormality is from the speed measurement model or the state transition model in the case where an abnormality exists, wherein the whole detection uses a card method inspection method.
7. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 6, wherein the fine detection comprises:
the detection amount of the speed detection model and the detection amount of the state transition model are constructed as follows:
wherein ,representing the difference between the current time speed measured value and the previous time fusion state value;representing the difference between the predicted state value at the current moment and the fusion state value at the previous moment; t is t 1 Detecting the quantity for a speed measurement model; t is t 2 Detecting the quantity for the state transition model; h is the observation matrix; />The motion state estimated value of the train at the k moment, namely the fusion state value, is obtained after the fusion processing is carried out on the mathematical model of the system, and the mean square error generated by fusion, namely the fusion mean square error, is P (k); />P (k-1) is the fusion mean square error of the previous moment; r is a covariance matrix of the measured noise vector; />For passing the fusion state value of the previous moment +.>Predicting the obtained state value at the current moment; />Mean square error generated for prediction, i.e. predictionSquare error;
comparison t 1 and t2 The numerical value of (1), if t 1 >t 2 If t, determining that the abnormal condition is from the speed measurement model 1 <t 2 It is determined that the abnormal condition is from the state transition model.
8. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the fusion is based on a fusion variance minimum principle to fuse a speed measurement model and a state transition model.
9. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 1, wherein the selecting different kalman filtering methods to fuse the speed measurement model and the state transition model comprises:
when no abnormal condition occurs at the moment k, a Kalman filtering method is selected;
when abnormality occurs in the speed measurement model at the moment k, selecting the first improved Kalman filtering method;
and when the state transition model at the moment k is abnormal, selecting a second improved Kalman filtering method.
10. The method for measuring speed by fusion of multi-source information of high-speed trains according to claim 9, wherein the obtaining of the fusion speed of the trains comprises:
b1, predicting the state at the current moment by utilizing the fusion state at the previous moment according to the state transition model, and calculating the mean square error of a state prediction result;
b2, selecting a filtering type according to an abnormality detection result, and if the filtering type is selected as a first improved Kalman filtering type, calculating an robust factor; if the second modified Kalman filtering is selected, calculating an adaptive factor; if the Kalman filtering method is selected, setting the robust factor and the adaptive factor to be 1, and enabling the robust factor and the adaptive factor to be in an invalid state;
b3, calculating a filtering gain according to the obtained robust factor and the adaptive factor, and calculating a mean square error of information fusion according to the filtering gain;
and B4, calculating the fusion state of the current moment according to the speed measurement model, the prediction state and the obtained filter gain at the current moment to obtain the train fusion speed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310759295.XA CN116972827A (en) | 2023-06-26 | 2023-06-26 | Multi-source information fusion speed measurement method for high-speed train |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310759295.XA CN116972827A (en) | 2023-06-26 | 2023-06-26 | Multi-source information fusion speed measurement method for high-speed train |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116972827A true CN116972827A (en) | 2023-10-31 |
Family
ID=88470351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310759295.XA Pending CN116972827A (en) | 2023-06-26 | 2023-06-26 | Multi-source information fusion speed measurement method for high-speed train |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116972827A (en) |
-
2023
- 2023-06-26 CN CN202310759295.XA patent/CN116972827A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7593796B2 (en) | Torque estimator for internal combustion engine | |
| CN109471143B (en) | Self-adaptive fault-tolerant train combined positioning method | |
| CN110161543B (en) | Partial gross error tolerance self-adaptive filtering method based on chi-square test | |
| CN104596514B (en) | The Real-time Noisy Reducer and method of accelerometer and gyroscope | |
| RU2686654C2 (en) | Method and device for aircraft engine impeller monitoring by measuring equilibrium position | |
| CN110058222B (en) | Double-layer particle filter tracking-before-detection method based on sensor selection | |
| CN109716157A (en) | Axle offset estimation device | |
| CN113901379B (en) | Real-time data dynamic online quick processing method for edge terminal | |
| CN112146648A (en) | Multi-target tracking method based on multi-sensor data fusion | |
| CN112046761B (en) | Airplane icing on-line detection method based on statistical test and filtering | |
| CN116972827A (en) | Multi-source information fusion speed measurement method for high-speed train | |
| CN113071509A (en) | Fault detection method for wheel angle sensor of intelligent vehicle | |
| CN116660965A (en) | Beidou inertial navigation positioning method, device and storage medium | |
| CN111319787B (en) | Helicopter moving part vibration monitoring data validity evaluation method | |
| CN112651078B (en) | Arrow body structure state abnormality detection method based on symbol dynamics | |
| Byington et al. | False alarm mitigation of vibration diagnostic systems | |
| CN106324590A (en) | Method and device of improving measuring precision of train velocity measuring system | |
| CN109877647A (en) | A kind of lathe axis servomotor performance degradation assessment system based on built-in encoder | |
| CN116560219B (en) | Self-adaptive monitoring control method and system based on transmission tower wind speed joint analysis | |
| KR20140139955A (en) | System and method of detecting failure of engine | |
| CN113524260B (en) | Multi-joint robot collision detection method | |
| CN117092636B (en) | System and method for recognizing false track of millimeter wave radar multi-target tracking | |
| CN117668755A (en) | Multi-sensor information fusion speed measurement method for train | |
| CN112198504B (en) | Fusion filtering method for interleaving of active and passive observation features | |
| JP2003175896A (en) | Failure and damage detecting apparatus for rudder surface |
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 |