Urban rail locomotive position estimation method considering electrical measurement information

Abstract

The present invention proposes a method for estimating the position of a rail locomotive taking into account electrical measurement information. When establishing a model for estimating the position of a rail locomotive taking into account electrical measurement information, the method first determines the number of locomotives in operation between adjacent traction stations according to the positioning result of the locomotive section, and constructs a circuit equivalent model of a traction power supply system under different numbers of locomotives; then, based on the circuit theorem, the multi-equation relationship between the locomotive position and each electrical measurement value in the traction power supply system is analyzed, and the locomotive position calculation value is obtained as a supplement to the position measurement; then, considering the influence of the measurement error and the modeling error, the locomotive position is optimally estimated by using the method for estimating the position of a rail locomotive based on a robust adaptive Kalman filter in combination with the locomotive operation process; the present invention can make up for the defect of the existing lack of locomotive positioning from the perspective of electrical measurement, and obtain an effective and practical method for positioning a rail locomotive.

Description

Technical Field

[0001] The invention relates to the technical field of power system dispatching automation. The power system may include power supply from renewable energy, in particular to a method for estimating the position of a rail vehicle taking into account electrical measurement information. Background Art

[0002] At present, the urban rail transit locomotive positioning method used in the project is mainly based on the section positioning method of the trackside equipment. However, due to the influence of the sampling frequency, the section identification result has a strong lag, and it is impossible to determine the specific location of the locomotive with clear rules, and the positioning accuracy is low. The research on locomotive positioning mostly adopts a combined positioning method based on multi-sensor data fusion. However, since urban rail locomotives basically run in special environments such as underground and tunnels, sensor signals are easily interfered and blocked, resulting in signal loss. Such algorithms have great difficulties in practical applications. At the same time, there is currently a lack of relevant research on locomotive positioning from the perspective of electrical measurement. In the circuit equivalent model of the traction power supply system, the locomotive position mainly affects the resistance value of each branch. Therefore, by constructing a circuit equivalent model of the traction power supply system and using the electrical quantity measurement in the system, such as node voltage, feeder current, etc., to solve the resistance value of each branch, the locomotive position can be effectively obtained, and the multi-equation relationship can be further derived according to the circuit theory. However, due to the presence of noise in the actual measurement, the solution result is not completely accurate, and its processing method needs to be further considered. Summary of the invention

[0003] The present invention proposes a method for estimating the position of a rail locomotive taking into account electrical measurement information. Based on circuit constraints and the locomotive operation process, the multi-coupling relationship between electrical measurement and locomotive position is fully analyzed, and measurement and modeling errors are further considered to seek the optimal estimation method for the position of the rail locomotive. The method can make up for the defect of the existing lack of locomotive positioning from the perspective of electrical measurement, and obtain an effective and practical method for positioning the rail locomotive.

[0004] The present invention adopts the following technical solutions.

[0005] The invention discloses a method for estimating the position of a rail locomotive taking into account electrical measurement information. When establishing a model for estimating the position of a rail locomotive taking into account electrical measurement information, the method first determines the number of locomotives in operation between adjacent traction depots according to the positioning result of the locomotive section, and constructs a circuit equivalent model of a traction power supply system under different numbers of locomotives; then, based on the circuit theorem, the multi-equation relationship between the locomotive position and each electrical measurement value in the traction power supply system is analyzed, and the locomotive position calculation value is obtained as a supplement to the position measurement; then, considering the influence of measurement error and modeling error, the locomotive position is optimally estimated by using a method for estimating the position of a rail locomotive based on a robust adaptive Kalman filter in combination with the locomotive operation process.

[0006] The method comprises the following steps:

[0007] Step S1, based on the locomotive section positioning result of the trackside equipment based on the axle counter and the transponder, obtaining the number of locomotives between adjacent traction depots;

[0008] Step S2, dividing the section by adjacent traction depots, and constructing a circuit equivalent model of the traction power supply system with different numbers of locomotives in the section;

[0009] Step S3, analyzing the correlation between the electrical measurement values ​​in the traction power supply system and the locomotive position in the section based on circuit theory, and obtaining a locomotive position calculation formula;

[0010] Step S4, constructing a rail locomotive position estimation model based on Kalman filtering, and adding the locomotive position calculation value obtained in step S3 to the measured value;

[0011] Step S5: Considering the influence of measurement errors, based on the robust estimation concept, identify and remove bad data in the measurement values, while reducing the influence of low-quality data;

[0012] Step S6: Considering the influence of modeling error, construct an adaptive adjustment factor to act on the estimation model to achieve optimal estimation of the locomotive position and output the estimation result.

[0013] Step S1 includes the following steps:

[0014] Step S11, obtaining the section location information of the running locomotive in the traction power supply system in real time from the locomotive automatic monitoring system;

[0015] Step S12: Acquire the physical location information of each section and traction depot, and obtain the number of locomotives in operation between adjacent traction depots according to the section location of the locomotive.

[0016] Step S2 specifically includes the following steps:

[0017] Step S21: The traction station and the locomotive are regarded as power supply nodes, and the overhead contact network and rails between the nodes are regarded as branches. The resistance of each branch is:

[0018] Formula 1;

[0019] In the formula, R c and R r are the branch resistances of the overhead contact network and rails respectively; r c and r r are the resistance per unit length of the contact network and the rail respectively; l is the branch length, i.e. the distance between adjacent nodes; i Indicates the branch number;

[0020] Step S22: Ignore the longitudinal resistance and stray current between the rail and the ground, and merge the R c and R r , and further construct the chain circuit model of the traction power supply system under different numbers of locomotives.

[0021] Step S3 specifically includes the following steps:

[0022] Step S31: Using the traction voltage U tss , locomotive voltage U train and feeder current I According to the circuit theorem, a multi-coupling relationship between the locomotive position in the section and the measured values ​​of each electrical quantity is established, and its formula form is:

[0023] Formula 2:

[0024] In the formula, m The distance from the locomotive to the traction station at the head end of the section; L The distance between adjacent traction stations;

[0025] Step S32: When there is a locomotive on both the up and down routes in the section, the position of the up locomotive relative to the front traction station is:

[0026] Formula 3;

[0027] In the formula, , U 1 , U 2 , U 3 , U 4 They are the voltage of the first and last traction stations and the voltage of the up and down locomotives respectively. I 1 , I 2 , I 3 , I 4 The four feeder currents drawn from the head and end traction;

[0028] Step S33: The situation where the interval contains the remaining number of locomotives is similar to the above, and the locomotive voltage value needs to be additionally added or deleted, and the feeder current remains unchanged.

[0029] Step S4 specifically includes the following steps:

[0030] Step S41: Get state variables x and quantity measurement z They are:

[0031]

[0032] In the formula, m is the position vector of all locomotives; v is the velocity vector of all locomotives; z m,mea is the position sensor measurement value vector of all locomotives; z v,mea is the speed measurement vector of all locomotives;

[0033] Step S42: The state transfer equation and the measurement equation are respectively expressed as:

[0034]

[0035] In the formula, I and 0 are the identity matrix and the zero matrix respectively; is the time interval; a is the acceleration measurement vector of all locomotives; F , B They are the state transfer matrix and the control matrix respectively; u is the control vector; H is the measurement matrix. Only when the measurement quantity describes the same locomotive status information as the state quantity, the corresponding element in the matrix is ​​1, and the rest are 0. It is a sparse matrix. w and v They represent process noise and measurement noise respectively, which satisfy the zero mean and covariance. Q and R Gaussian white noise; subscript t Indicates the moment;

[0036] Step S43: The position estimation model based on Kalman filtering includes two steps: prediction and update, which can be expressed as:

[0037]

[0038] In the formula, Represent the predicted value vector, the prediction error covariance matrix and the state transition covariance matrix respectively; and P Represent the estimated value vector and the estimated error covariance matrix respectively; R represents the measurement error covariance matrix; K represents the Kalman gain matrix.

[0039] Step S5 specifically includes the following steps:

[0040] Step S51: The residual vector measured at any time is V , the weight matrix of the quantity measurement is B , the equivalent weight matrix is , each measurement is independent of each other, and there are:

[0041]

[0042] In the formula, n Indicates the number of measured values; ω i Represents the equivalent weight factor, which can be determined by the equivalent weight function; the IGGⅢ equivalent weight function is used:

[0043]

[0044] In the formula, Respectively represent i The measured residuals and their standardized residuals, omitting the time subscripts t ; c 0 and c 1 is the harmonic coefficient;

[0045] Step S52: At any time, the predicted value is used as the initial value of the robust estimation iteration, and the initial residual value is obtained as follows:

[0046] Formula XVII;

[0047] Step S53: The measurement values ​​of different locomotive positions are processed separately. For all measurement residuals of the same locomotive position, the absolute values ​​are taken and normalized at the measurement residuals corresponding to the predicted values ​​to further unify the error measurement standard:

[0048] Formula XVIII;

[0049] In the formula, the superscript p Indicates the locomotive number, omitting the iteration number superscript ( k ) and time subscript t ; In the first iteration, ;

[0050] Step S54: Measure the minimum residual value of the normalized sensor As harmonic coefficient c 0 and c 1 Selection basis:

[0051] Formula 19;

[0052] Step S55: Obtaining the equivalent weight factor of each measurement value ω i And remove ω i <0.1 measurement value;

[0053] Step S56: The robust estimation result is solved by iterative method:

[0054] Formula 20;

[0055] In the formula, the superscript k Represents the number of iterations, and the following is similar; when Stop the iteration when , and get the robust solution at the current moment , is a minimum value;

[0056] Step S57: Maintain the initial weight, i.e. , the operation is iterated until the convergence condition is met, and the retained measurement value has a new equivalent weight factor ;

[0057] Step S58: Calculate the latest equivalent weight of each measurement using the formula As the measurement value weight of the subsequent filtering process, the calculation formula obtains the robust solution at the current moment .

[0058] Step S6 specifically includes the following steps:

[0059] Step S61: The innovation vector is the actual measurement value The error vector between the measured predicted value is:

[0060] Formula 21;

[0061] The theoretical covariance matrix of the innovation and the actual covariance matrix They are:

[0062]

[0063] In the formula, N is the sliding window length;

[0064] Step S62: Constructing adaptive factors α 1t for:

[0065]

[0066] In the formula, represents the trace of a matrix;

[0067] Step S63: Modify the formula and obtain:

[0068]

[0069] In the formula, is the multiplication factor, Q t The value is determined;

[0070] Step S64: Status mismatch statistics by Reflection as a benchmark The degree of deviation:

[0071] Formula XXVII;

[0072] Step S65: Constructing adaptive factors α 2t :

[0073]

[0074] In the formula, m and n is a constant, Approaching 0; , m Take the middle and late Maximum experience value;

[0075] Step S66: further modify the formula and obtain:

[0076]

[0077] In the formula, Depend on The value is determined and satisfies ;

[0078] Step S67: Estimating the locomotive position and saving the estimation result.

[0079] The locomotive is an urban rail locomotive driven by electricity and using a direct current traction power supply system.

[0080] In step S1 , sensor measurement data is selected to determine the segment positioning result.

[0081] The present invention establishes a model for estimating the position of a rail locomotive taking into account electrical measurement information. First, the number of locomotives in operation between adjacent traction stations is determined based on the results of locomotive section positioning, and a circuit equivalent model of the traction power supply system under different numbers of locomotives is constructed; secondly, the multi-equation relationship between the locomotive position and each electrical measurement value in the traction power supply system is analyzed based on the circuit theorem, and the locomotive position calculation value is obtained as a supplement to the position measurement; then, considering the influence of measurement error and modeling error, combined with the locomotive operation process, an urban rail locomotive position estimation method based on robust adaptive Kalman filtering is designed to achieve the optimal estimation of the locomotive position; the method described in the present invention is conducive to improving the accuracy of urban rail locomotive positioning under limited measurement conditions, thereby ensuring the safe and efficient operation of urban rail transportation.

[0082] In the urban rail locomotive position estimation method proposed in the present invention, the method of calculating the locomotive position by using electrical measurement values ​​in combination with circuit characteristic constraints can effectively solve the problem of insufficient real-time measurement of the locomotive position, and the reliability of the robust estimation result is high. At the same time, in the case of inaccurate measurement, this method can effectively improve the accuracy of locomotive positioning, reducing the error from 100 meters to 10 meters.

[0083] The main processes implemented by the method of the present invention include a correlation analysis between various electrical measurement information in the traction power supply system and the locomotive position and a method for establishing a calculation formula, as well as a method for establishing and solving a urban rail locomotive position estimation model based on a robust adaptive Kalman filter. At the same time, relevant research on urban rail locomotive position estimation is carried out based on circuit constraints and locomotive operation processes, and an urban rail locomotive position estimation method taking electrical measurement information into account is proposed, which is beneficial to improving the accuracy of urban rail locomotive positioning under limited measurement conditions, thereby ensuring the safe and efficient operation of urban rail transportation.

[0084] The urban rail locomotive position estimation method proposed in the present invention fully analyzes the locomotive position information contained in the electrical measurement values ​​of the traction power supply system through system circuit topology and circuit theory, establishes multiple calculation equations for each locomotive position, obtains the locomotive position calculation results and serves as a measurement supplement. A position estimation model based on robust adaptive Kalman filtering is established, which can effectively eliminate possible bad data in the measurement values, and can adaptively adjust the most critical prediction and measurement error covariance parameters in the filter model according to the locomotive running state, thereby achieving the optimal estimation of the locomotive position, improving the applicability and robustness of the algorithm, and further improving the accuracy of urban rail locomotive positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0086] Attached Figure 1 It is a schematic diagram of the overall process of an embodiment of the present invention;

[0087] Attached Figure 2 It is a schematic diagram of an equivalent model of a traction power supply system circuit when there is a locomotive both in the uplink and downlink directions according to an embodiment of the present invention;

[0088] Attached Figure 3 It is a schematic diagram of a locomotive position estimation error result obtained based on the locomotive position estimation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0089] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail as follows:

[0090] like Figure 1 As shown, a method for estimating the position of a rail locomotive taking into account electrical measurement information is provided. When establishing a model for estimating the position of a rail locomotive taking into account electrical measurement information, the method first determines the number of locomotives in operation between adjacent traction depots according to the positioning result of the locomotive section, and constructs a circuit equivalent model of the traction power supply system under different numbers of locomotives; then, based on the circuit theorem, the multi-equation relationship between the locomotive position and each electrical measurement value in the traction power supply system is analyzed to obtain the calculated value of the locomotive position as a supplement to the position measurement; then, considering the influence of measurement error and modeling error, combined with the locomotive operation process, the locomotive position is optimally estimated by using an urban rail locomotive position estimation method based on robust adaptive Kalman filtering.

[0091] The method comprises the following steps:

[0092] Step S1, based on the locomotive section positioning result of the trackside equipment based on the axle counter and the transponder, obtaining the number of locomotives between adjacent traction depots;

[0093] Step S2, dividing the section by adjacent traction depots, and constructing a circuit equivalent model of the traction power supply system with different numbers of locomotives in the section;

[0094] Step S3, analyzing the correlation between the electrical measurement values ​​in the traction power supply system and the locomotive position in the section based on circuit theory, and obtaining a locomotive position calculation formula;

[0095] Step S4, constructing a rail locomotive position estimation model based on Kalman filtering, and adding the locomotive position calculation value obtained in step S3 to the measured value;

[0096] Step S5: Considering the influence of measurement errors, based on the robust estimation concept, identify and remove bad data in the measurement values, while reducing the influence of low-quality data;

[0097] Step S6: Considering the influence of modeling error, construct an adaptive adjustment factor to act on the estimation model to achieve optimal estimation of the locomotive position and output the estimation result.

[0098] Step S1 includes the following steps:

[0099] Step S11, obtaining the section location information of the running locomotive in the traction power supply system in real time from the locomotive automatic monitoring system;

[0100] Step S12: Acquire the physical location information of each section and traction depot, and obtain the number of locomotives in operation between adjacent traction depots according to the section location of the locomotive.

[0101] Step S2 specifically includes the following steps:

[0102] Step S21: The traction station and the locomotive are regarded as power supply nodes, and the overhead contact network and rails between the nodes are regarded as branches. The resistance of each branch is:

[0103] Formula 1;

[0104] In the formula, R c and R r are the branch resistances of the overhead contact network and rails respectively; r c and r r are the resistance per unit length of the contact network and the rail respectively; l is the branch length, i.e. the distance between adjacent nodes; i Indicates the branch number;

[0105] Step S22: Ignore the longitudinal resistance and stray current between the rail and the ground, and merge the R c and R r , and further construct the chain circuit model of the traction power supply system under different numbers of locomotives.

[0106] Step S3 specifically includes the following steps:

[0107] Step S31: Using the traction voltage U tss , locomotive voltage U train and feeder current I According to the circuit theorem, a multi-coupling relationship between the locomotive position in the section and the measured values ​​of each electrical quantity is established, and its formula form is:

[0108] Formula 2:

[0109] In the formula, m The distance from the locomotive to the traction station at the head end of the section; L The distance between adjacent traction stations;

[0110] Step S32: When there is a locomotive on both the up and down routes in the section, the position of the up locomotive relative to the front traction station is:

[0111] Formula 3;

[0112] In the formula, , U 1 , U 2 , U 3 , U 4 They are the voltage of the first and last traction stations and the voltage of the up and down locomotives respectively. I 1 , I 2 , I 3 , I 4 The four feeder currents drawn from the head and end traction;

[0113] Step S33: The situation where the interval contains the remaining number of locomotives is similar to the above, and the locomotive voltage value needs to be additionally added or deleted, and the feeder current remains unchanged.

[0114] Step S4 specifically includes the following steps:

[0115] Step S41: Get state variables x and quantity measurement z They are:

[0116]

[0117] In the formula, m is the position vector of all locomotives; v is the velocity vector of all locomotives; z m,mea is the position sensor measurement value vector of all locomotives; z v,mea is the speed measurement vector of all locomotives;

[0118] Step S42: The state transfer equation and the measurement equation are respectively expressed as:

[0119]

[0120] In the formula, I and 0 are the identity matrix and the zero matrix respectively; is the time interval; a is the acceleration measurement vector of all locomotives; F , B They are the state transfer matrix and the control matrix respectively;u is the control vector; H is the measurement matrix. Only when the measurement quantity describes the same locomotive status information as the state quantity, the corresponding element in the matrix is ​​1, and the rest are 0. It is a sparse matrix. w and v They represent process noise and measurement noise respectively, which satisfy the zero mean and covariance. Q and R Gaussian white noise; subscript t Indicates the moment;

[0121] Step S43: The position estimation model based on Kalman filtering includes two steps: prediction and update, which can be expressed as:

[0122]

[0123] In the formula, Represent the predicted value vector, the prediction error covariance matrix and the state transition covariance matrix respectively; and P Represent the estimated value vector and the estimated error covariance matrix respectively; R represents the measurement error covariance matrix; K represents the Kalman gain matrix.

[0124] Step S5 specifically includes the following steps:

[0125] Step S51: The residual vector measured at any time is V , the weight matrix of the quantity measurement is B , the equivalent weight matrix is , each measurement is independent of each other, and there are:

[0126]

[0127] In the formula, n Indicates the number of measured values; ω i Represents the equivalent weight factor, which can be determined by the equivalent weight function; the IGGⅢ equivalent weight function is used:

[0128]

[0129] In the formula, Respectively represent i The measured residuals and their standardized residuals, omitting the time subscripts t ; c 0 and c 1 is the harmonic coefficient;

[0130] Step S52: At any time, the predicted value is used as the initial value of the robust estimation iteration, and the initial residual value is obtained as follows:

[0131] Formula XVII;

[0132] Step S53: The measurement values ​​of different locomotive positions are processed separately. For all measurement residuals of the same locomotive position, the absolute values ​​are taken and normalized at the measurement residuals corresponding to the predicted values ​​to further unify the error measurement standard:

[0133] Formula XVIII;

[0134] In the formula, the superscript p Indicates the locomotive number, omitting the iteration number superscript ( k ) and time subscript t ; In the first iteration, ;

[0135] Step S54: Measure the minimum residual value of the normalized sensor As harmonic coefficient c 0 and c 1 Selection basis:

[0136] Formula 19;

[0137] Step S55: Obtaining the equivalent weight factor of each measurement value ω i And remove ω i <0.1 measurement value;

[0138] Step S56: The robust estimation result is solved by iterative method:

[0139] Formula 20;

[0140] In the formula, the superscript k Represents the number of iterations, and the following is similar; when Stop the iteration when , and get the robust solution at the current moment , is a minimum value;

[0141] Step S57: Maintain the initial weight, i.e. , the operation is iterated until the convergence condition is met, and the retained measurement value has a new equivalent weight factor ;

[0142] Step S58: Calculate the latest equivalent weight of each measurement using the formula As the measurement value weight of the subsequent filtering process, the calculation formula obtains the robust solution at the current moment .

[0143] Step S6 specifically includes the following steps:

[0144] Step S61: The innovation vector is the actual measurement value The error vector between the measured predicted value is:

[0145] Formula 21;

[0146] The theoretical covariance matrix of the innovation and the actual covariance matrix They are:

[0147]

[0148] In the formula, N is the sliding window length;

[0149] Step S62: Constructing adaptive factors α 1t for:

[0150]

[0151] In the formula, represents the trace of a matrix;

[0152] Step S63: Modify the formula and obtain:

[0153]

[0154] In the formula, is the multiplication factor, Q t The value is determined;

[0155] Step S64: Status mismatch statistics by Reflection as a benchmark The degree of deviation:

[0156] Formula XXVII;

[0157] Step S65: Constructing adaptive factors α 2t :

[0158]

[0159] In the formula, m and n is a constant, Approaching 0; , m Take the middle and late Maximum experience value;

[0160] Step S66: further modify the formula and obtain:

[0161]

[0162] In the formula, Depend on The value is determined and satisfies ;

[0163] Step S67: Estimating the locomotive position and saving the estimation result.

[0164] The locomotive is an urban rail locomotive driven by electricity and using a direct current traction power supply system.

[0165] In step S1 , sensor measurement data is selected to determine the segment positioning result.

[0166] Based on the above model and process design, this example implements the proposed urban rail locomotive position estimation method taking into account electrical measurement information in the MATLAB environment. The modeling and solution process is shown in Figure 1 shown.

[0167] According to one embodiment of the present invention, in order to further verify the effectiveness and accuracy of the proposed urban rail locomotive position estimation method, a simulation example was designed based on the Beijing Metro Line 13A data to perform simulation verification of the position estimation algorithm. The line has a total of 23 traction depots and 46 locomotives in operation.

[0168] The sensor measurement is selected as the segment positioning result, and the random error in the range of [-0.2, 0.2] km is used for approximation. The simulation position estimation result is as follows: Figure 3 And as shown in Table 1.

[0169] Table 1 Comparison table of average root mean square error at different times (unit: meter)

[0170]

[0171] From Table 1 and Figure 3 It can be seen that in the proposed urban rail locomotive position estimation method, the method of calculating the locomotive position by using electrical measurement values ​​combined with circuit characteristic constraints can effectively solve the problem of insufficient real-time measurement of the locomotive position, and the reliability of the robust estimation result is high. At the same time, in the case of inaccurate measurement, this method can effectively improve the accuracy of locomotive positioning, reducing the error from 100 meters to 10 meters.

[0172] The main processes implemented by the method shown in this example include the correlation analysis and calculation formula establishment method between the electrical measurement information in the traction power supply system and the locomotive position, as well as the establishment and solution method of the urban rail locomotive position estimation model based on the robust adaptive Kalman filter.

[0173] This example shows that relevant research on urban rail locomotive position estimation was carried out based on circuit constraints and locomotive operation process, and an urban rail locomotive position estimation method considering electrical measurement information was proposed, which is conducive to improving the accuracy of urban rail locomotive positioning under limited measurement conditions, thereby ensuring the safe and efficient operation of urban rail transportation.

[0174] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of urban rail locomotive position estimation methods that take into account electrical measurement information under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A method for estimating the position of a rail vehicle taking into account electrical measurement information, characterized in that: When establishing a locomotive position estimation model taking into account electrical measurement information, the method first determines the number of locomotives in operation between adjacent traction depots according to the locomotive section positioning result, and constructs a circuit equivalent model of the traction power supply system under different numbers of locomotives; then, based on the circuit theorem, the multi-equation relationship between the locomotive position and each electrical measurement value in the traction power supply system is analyzed to obtain the locomotive position calculation value as a supplement to the position measurement; then, considering the influence of measurement error and modeling error, combined with the locomotive operation process, the locomotive position is optimally estimated by using the locomotive position estimation method based on robust adaptive Kalman filtering; The method comprises the following steps: Step S1, based on the locomotive section positioning result of the trackside equipment based on the axle counter and the transponder, obtaining the number of locomotives between adjacent traction depots; Step S2, dividing the section by adjacent traction depots, and constructing a circuit equivalent model of the traction power supply system with different numbers of locomotives in the section; Step S3, analyzing the correlation between the electrical measurement values ​​in the traction power supply system and the locomotive position in the section based on circuit theory, and obtaining a locomotive position calculation formula; Step S4, constructing a rail locomotive position estimation model based on Kalman filtering, and adding the locomotive position calculation value obtained in step S3 to the measured value; Step S5: Considering the influence of measurement errors, based on the robust estimation concept, identify and remove bad data in the measurement values, while reducing the influence of low-quality data; Step S6: Considering the influence of modeling error, construct an adaptive adjustment factor to act on the estimation model to achieve optimal estimation of the locomotive position and output the estimation result.

2. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 1, characterized in that: Step S1 includes the following steps: Step S11, obtaining the section location information of the running locomotive in the traction power supply system in real time from the locomotive automatic monitoring system; Step S12: Acquire the physical location information of each section and traction depot, and obtain the number of locomotives in operation between adjacent traction depots according to the section location of the locomotive.

3. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 1, characterized in that: Step S2 specifically includes the following steps: Step S21: The traction station and the locomotive are regarded as power supply nodes, and the overhead contact network and rails between the nodes are regarded as branches. The resistance of each branch is: In the formula, R c and R r are the branch resistances of the overhead contact network and rails respectively; r c and r r are the resistance per unit length of the overhead line and the rail respectively; l is the branch length, i.e., the distance between adjacent nodes; the subscript i indicates the branch number; Step S22: Ignore the longitudinal resistance and stray current between the rail and the ground, and merge the R c With R r , and further construct the chain circuit model of the traction power supply system under different numbers of locomotives.

4. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 2, characterized in that: Step S3 specifically includes the following steps: Step S31: Using the traction voltage U tss , locomotive voltage U train The electrical quantities of the feeder current I and the multi-coupling relationship between the locomotive position in the section and the electrical measurement values ​​are established according to the circuit theorem. The formula is as follows: m=f(U tss , U train ,I,r c , r r , L) Formula 2; In the formula, m is the distance from the locomotive to the traction station at the head of the section; L is the distance between adjacent traction stations; Step S32: When there is a locomotive on both the up and down routes in the section, the position of the up locomotive relative to the front traction station is: In the formula, r = r c +r r , U1, U2, U3, U4 are the voltages of the first and last traction stations and the voltages of the up and down locomotives respectively, I1, I2, I3, I4 are the four feeder currents drawn from the first and last traction stations; Step S33: The situation where the interval contains the remaining number of locomotives is the same as the above steps, and the locomotive voltage value needs to be additionally added or deleted, and the feeder current remains unchanged.

5. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 4, characterized in that: Step S4 The specific steps include: Step S41: Take the state variable x and the quantity measurement z as: x=[m,v] T Formula 4; z=[z m，mea ,z m,cal ,z v，mea ] T Formula 5; Where m is the position vector of all locomotives; v is the velocity vector of all locomotives; z m,mea is the position sensor measurement value vector of all locomotives; v,mea is the speed measurement vector of all locomotives; Step S42: The state transfer equation and the measurement equation are respectively expressed as: z t =H t x t +v t公式七 ; Where I and 0 are the unit matrix and the zero matrix respectively; Δt is the time interval; a is the acceleration measurement vector of all locomotives; F and B are the state transfer matrix and the control matrix respectively; u is the control vector; H is the measurement matrix. Only when the measurement quantity and the state quantity describe the same locomotive state information, the corresponding element in the matrix is ​​1, and the rest are 0. It is a sparse matrix; w and v represent the process noise and measurement noise respectively, which are Gaussian white noises with zero mean and covariances of Q and R respectively; the subscript t represents the time; Step S43: The position estimation model based on Kalman filtering includes two steps: prediction and update, which can be expressed as: P t - =FP t-1 F T +Q t Formula 9; P t =(IK t H t ) t - Formula 12; In the formula, P - and Q represent the predicted value vector, the prediction error covariance matrix, and the state transition covariance matrix, respectively; and P represent the estimated value vector and the estimated error covariance matrix respectively; R represents the measurement error covariance matrix; K represents the Kalman gain matrix.

6. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 5, characterized in that: Step S5 specifically includes the following steps: Step S51: The residual vector measured at any time is V, the weight matrix of the measurement is B, and the equivalent weight matrix is Each measurement is independent of each other and includes: Where n represents the number of measured values; ω i Represents the equivalent weight factor, which is determined by the equivalent weight function; the IGGⅢ equivalent weight function is used: Where V i and They represent the i-th measurement residual value and its standardized residual, respectively, and the time subscript t is omitted; c0 and c1 are the harmonic coefficients; Step S52: At any time, the predicted value is used as the initial value of the robust estimation iteration, and the initial residual value is obtained as follows: Step S53: The measurement values ​​of different locomotive positions are processed separately. For all measurement residuals of the same locomotive position, the absolute values ​​are taken and normalized at the measurement residuals corresponding to the predicted values ​​to further unify the error measurement standard: In the formula, the superscript p represents the locomotive number, the iteration number superscript (k) and the time subscript t are omitted; in the first iteration, Step S54: Measure the minimum residual value of the normalized sensor The basis for selecting the harmonic coefficients c0 and c1 is: c ＝ k*c0 Formula 19; Step S55: obtaining the equivalent weight factor ωi of each measurement value and eliminating the measurement value with ωi<0.1; Step S56: The robust estimation result is solved by iterative method: In the formula, the superscript k represents the number of iterations; when Stop the iteration when , and get the robust solution at the current moment ε is a minimum value; Step S57: Maintain the initial weight, i.e. The operation is iterated until the convergence condition is met and the retained measurement value has a new equivalent weight factor Step S58: Calculate the latest equivalent weight of each measurement using Formula 15 and Formula 16 As the measurement value weight of the subsequent filtering process, the calculation formula obtains the robust solution at the current moment 7. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 5, characterized in that: Step S6 specifically includes the following steps: Step S61: The innovation vector is the actual measurement value z t The error vector between the measured predicted value is: The theoretical covariance matrix C of the innovation t and the actual covariance matrix They are: Where N is the sliding window length; Step S62: construct an adaptive factor α 1t for: Where tr(·) represents the trace of the matrix; Step S63: Correct Formula 9 and Formula 10 to obtain: P t - =FP t-1 F T +α 1t Q t Formula 25; In the formula, κ1 is the multiplication factor, which is determined by the value of Qt; Step S64: Status mismatch statistics by Reflect as a benchmark The degree of deviation: Step S65: Construct the adaptive factor α2t: In the formula, m and n are constants, and ε approaches 0; let m takes the middle and late stages Maximum experience value; Step S66: Further modifying Formula 25 and Formula 26 to obtain: P t - =(FP t-1 F T +α 1t Q t ) / α 2t Formula XXIX; In the formula, κ2 is composed of α 2t The value is determined and satisfies P1 - / α 2t >>κ2R1; Step S67: Estimating the locomotive position and saving the estimation result.

8. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 1, characterized in that: The locomotive is an urban rail locomotive driven by electricity and using a direct current traction power supply system.

9. The method for estimating the position of a rail vehicle taking into account electrical measurement information according to claim 2, characterized in that: In step S1 , sensor measurement data is selected to determine the segment positioning result.

Images