EP4194313A1

EP4194313A1 - On-board detection device for a railway vehicle, railway vehicle comprising such device and associated railway system

Info

Publication number: EP4194313A1
Application number: EP22212010.7A
Authority: EP
Inventors: Muniandi GANESAN; Arun RANJAN; Arvind Muthukrishnan; Manoj NAGARAJAN; Arunava CHAKRABORTY
Original assignee: Alstom Holdings SA
Current assignee: Alstom Holdings SA
Priority date: 2021-12-09
Filing date: 2022-12-07
Publication date: 2023-06-14

Abstract

The present disclosure relates to an on-board detection device (20) for a railway vehicle (4), the detection device (20) being configured for detecting a start and/or end of a neutral section (N) of a railroad infrastructure (2) on which the railway vehicle (4) is intended to run, a neutral section (N) being a non-powered section of a power line (8) of the railroad infrastructure (2), the detection device (20) comprising a wheel wear database (36) comprising a wear coefficient relative to the wear of wheels (15) of the railway vehicle (4), the detection device (20) furthermore comprising a receiving module (22) configured for receiving a plurality of speed signals of the speed of the railway vehicle (4) from on-board speed sensors (16, 18) of the railway vehicle (4), and at least one reference database (32) comprising start and end positions of the neutral sections (N) of the railroad infrastructure (2).

Description

The present invention relates to an on-board detection device for a railway vehicle.
The present invention also relates to a railway vehicle comprising such a device, and to a railway system comprising a railroad infrastructure and an associated railway vehicle.
Railroad infrastructures generally have power lines for electrical supply of railway vehicles, such as overhead lines. In large infrastructures, the power lines are often divided in a plurality of sections for electrical supply, which are separated by non-powered sections, which are also called neutral sections. This allows preventing for example multiple failures of power supply or avoiding that electrical phase synchronization cannot be guaranteed in the whole power line.
When a railway vehicle approaches a neutral section, circuit breakers of the vehicle must be opened and the vehicle must be coasting in an idle mode, so as to prevent the formation of electrical arcs. In an analog manner, in the end of the neutral section, the circuit breakers are closed again so that traction of the vehicle continues.
In order to detect neutral sections, generally Automatic Train Control (ATC) uses track-based beacons and speed sensors. It sends the information about the neutral section start and end to the railway vehicle for automatic removal and reapplication of power.
If the ATC fails or is bypassed, other systems using trackside mounted permanent magnets, transponders such as RFID or automated equipment identifier tags are implemented for detection of the neutral sections.
However, such systems have some drawbacks, because providing track-based transponders is costly. In addition, such systems are not always reliable, as track-based material can fail or may be damaged.
It is therefore an aim of the present invention to obtain a backup device or system for detection of neutral sections, which is easy and cost-efficient, while guaranteeing a high position accuracy for detecting a neutral section.
To this end, the present disclosure concerns an on-board detection device for a railway vehicle, the detection device being configured for detecting a start and/or end of a neutral section of a railroad infrastructure on which the railway vehicle is intended to run, a neutral section being a non-powered section of a power line of the railroad infrastructure, the detection device comprising a wheel wear database comprising a wear coefficient relative to the wear of wheels of the railway vehicle, the detection device furthermore comprising:

a receiving module configured for receiving a plurality of speed signals of the speed of the railway vehicle from on-board speed sensors of the railway vehicle;
at least one reference database comprising start and end positions of the neutral sections of the railroad infrastructure;
at least one memory module comprising at least an initial position of the railway vehicle;
an estimation module comprising a Kalman filter forming a kinematic model of the railway vehicle, the Kalman filter being configured for providing an estimated distance of the railway vehicle from the initial position based on said speed signals; and
a determination module configured for correcting the estimated distance in function of the wear coefficient so as to obtain a corrected estimated distance, the determination module furthermore configured for determining an estimated position of the railway vehicle based on the corrected estimated distance and the initial position, and for comparing said estimated position with the start and end positions of the reference database to detect a start and/or end of a neutral section.

Further embodiments may relate to one or more of the following features, which may be combined in any technical feasible combination:

the detection device furthermore comprises a gating module configured for receiving, at a current time step, a speed value of the railway vehicle estimated by the Kalman filter for a time step prior to said current time step, and for validating each speed signal received at the current time step having a difference to said speed value lower than an error value, so as to obtain validated speed signals, the Kalman filter being configured for providing the estimated distance as a function of the validated speed signals;
the gating module is configured to determine the error value as equal to N times of a standard deviation of the speed signals at the current time step, N being an integer greater than 1;
the detection device furthermore comprises a weightage module configured for receiving, at a current time step, a speed value of the railway vehicle propagated by the Kalman filter for a time step prior to said current time step, and for multiplying each speed signal received at the current time step with a respective weightage coefficient, which is a function of the difference between the speed signal at the current time step and said speed value of the time step prior to said current time step, so as to obtain weighted speed signals, the Kalman filter being configured for providing the estimated distance in function of the weighted speed signals;
the wear coefficient is susceptible to be updated as a function of a difference between the estimated distance provided by the Kalman filter and a predetermined reference distance;
the Kalman filter is configured to provide the estimated distance in the absence of a correction of said estimated distance, during operation of the railway vehicle, by any measurement received from a track-based beacon or a track-based transponder;
each speed sensor is chosen from a first speed sensor type operative to obtain a first speed signal from measuring a rotation of one of the wheels and an axle connected to the wheel, and a second speed sensor type operative to obtain a second speed signal from a measurement of a traction current and a voltage of a traction motor of the railway vehicle, the traction motor being connected to the wheel, the speed signals comprising the first speed signal and the second speed signal;
the estimation module is configured for determining an intermediate estimated distance from the initial position in function of said speed signals, and for determining an estimation difference between the intermediate estimated distance at a current time step and an estimated distance at a time step previous to the current time step;

if the estimation difference is lower than or equal to a predetermined threshold, the estimation module is configured for providing the estimated distance, which is the intermediate estimated distance in this case,
if the estimation difference is higher than the predetermined threshold, the estimation module is configured for providing the estimated distance, which is the estimated distance at the time step previous to the current time step in this case.

According to another aspect, the present disclosure concerns a railway vehicle comprising a plurality of on-board speed sensors and a detection device as described above, the detection device being configured for receiving speed signals from the on-board speed sensors.
According to another aspect, the present disclosure concerns a railway system comprising a railroad infrastructure and a railway vehicle as described above, the railroad infrastructure comprising a trackside equipment configured for determining a wheel wear condition of the railway vehicle intended to run on the railroad infrastructure, the trackside equipment comprising:

a trackside receiving module configured for receiving an estimated distance from the railway vehicle, and for receiving a plurality of speed signals of the speed of the railway vehicle from on-board speed sensors of said railway vehicle,
a trackside estimation module comprising a trackside Kalman filter forming a kinematic model of the railway vehicle, the trackside Kalman filter being configured for providing a wheel wear value for each axle of the railway vehicle having a corresponding speed sensor, in function of the estimated distance and said speed signals, and
a trackside determination module configured for determining a wheel wear condition for said each axle in function of the corresponding wheel wear value, the wheel wear condition being chosen from among a plurality of predetermined wheel wear conditions.

These features and advantages of the invention will be further explained in the following description, given only as a non-limiting example, and with reference to the attached drawing, on which the single figure 1 is a schematic view of a railway system comprising a railway vehicle having an on-board detection device according to the invention.
With reference to figure 1, a railway system 1 comprises a railroad infrastructure 2 and a railway vehicle 4 which is intended to run on the railroad infrastructure 2.
The railroad infrastructure 2 comprises rails 6 and a power line 8, such as an overhead line, for providing electrical power to the railway vehicle 4. The power line 8 has neutral sections N. According to an example, the railroad infrastructure 2 furthermore comprises trackside equipment 10.
The railway vehicle 4 comprises for example at least one traction motor 12 configured to be connected via a pantograph 14 to the power line 8, a plurality of wheels 15 connected to corresponding axles, at least one of the plurality of wheels being driven by the traction motor 12, a first and a second speed sensors 16, 18 configured to respectively provide one speed signal S1 and S2 representative of the speed of the railway vehicle 4, and an on-board detection device 20 configured to receive the first and second speed signals S1, S2.
In the example of the figure, the railway vehicle 4 comprises two different types of speed sensors 16, 18. First speed sensor 16 is of a first type which is functionally mounted so as to obtain the first speed signal S1 from measuring a rotation of the axle of the rail vehicle 4. Second speed sensor 18 is of a second type configured to obtain a second speed signal S2 from a measurement of a traction current and a voltage of the traction motor 12 of the railway vehicle 4.
The on-board detection device 20 comprises a receiving module 22, a gating module 24, a weightage module 26, an estimation module 28 comprising a Kalman filter 29 and a determination module 30. The on-board detection device 20 furthermore comprises at least one reference database 32 comprising start and end positions of the neutral sections N of the railroad infrastructure 2, at least one memory module 34 comprising at least an initial position of the railway vehicle 4, and a wheel wear database 36 comprising a wear coefficient relative to the wear of the wheels 15 of the railway vehicle 4.
The detection device 20 is for example implemented by a computer having at least one processor and corresponding memory. In this case, the receiving module 22, the gating module 24, the weightage module 26, the estimation module 28 and the determination module 30 are each, for example, executable by the processor and stored in the memory of the computer.
The receiving module 22 is configured for receiving a plurality of speed signals S of the speed of the railway vehicle 4 from the on- board speed sensors 16, 18 of the railway vehicle 4. The receiving module 22 is configured for transmitting the speed signals S to the gating module 24, to the weightage module 26, or directly to the estimation module 28 in the case of the absence of the gating module 24 and the weightage module 26. The transmission between the modules is in particular performed via corresponding connections 38.
In particular, the speed signals S comprise the first speed signal S1 and the second speed signal S2.
Each speed signal S comprises, for a current time step, a measured speed of the vehicle 4.
The value of the first and second speed signals S1, S2 may be different one from another for example due to slipping or sliding of the wheel 15 to which the corresponding first speed sensor 16 is operatively connected.
By "slipping" or "slip condition", it is understood an excess rotation of a wheel 15, also called overrun. This means for example that a wheel 15 rotates so that a relative movement occurs between the rail 6 and a point of a wheel surface in contact with the rail 6 at a given time step.
This condition may for example occur when setting in movement the vehicle 4. Slip leads to a higher measured speed value by the first speed sensor 16, functionally connected to the corresponding wheel 15, than the actual speed of the vehicle 4.
According to an example, the slip condition leads furthermore to a higher speed value measured by the second speed sensor 18, electrically connected to the motor 12, than the actual speed of the vehicle 4, as a traction of the motor 12 turns the wheels 15, but the vehicle 4 does not or only slightly move.
By "sliding" or "slide condition", it is understood that a wheel 15 does not rotate or at least rotate so that a relative movement occurs between the rail 6 and a point of a wheel surface in contact with the rail 6 at a given time step, in particular that it rotates slower than the actual movement of the vehicle 4. For example, in this condition, the wheel 15 is dragged along the rails 6. In this case, the first speed sensor 16 of the corresponding wheel 15 measures a speed value which is lower than the actual movement of the vehicle 4, for example close to zero.
In the present description, the expressions "propagated by the Kalman filter 29" and "estimated by the Kalman filter 29" are used as equivalents. In particular, by these expressions is understood that a corresponding state of the Kalman filter is propagated over a time step, which leads to a propagated or estimated state provided by the Kalman filter 29.
The gating module 24 is configured for receiving, at a current time step, a speed value of the railway vehicle 4 propagated by the Kalman filter 29 for a time step prior to said current time step, and for receiving the speed signals S from the receiving module 22.
The gating module 24 is configured for validating each speed signal S received at the current time step having a difference to said speed value lower than an error value, so as to obtain validated speed signals VS. The error value is a margin which is for example predefined in function of characteristics of the first speed sensors 16, the wheels 15, the time duration between two subsequent time steps, etc.
The gating module 24 is for example configured to determine the error value as equal to N times of a standard deviation of the speed signals S at the current time step. N is an integer of at least 1. For example, the error value is equal to three times the standard deviation of the received speed signals S.
The gating module 24 allows eliminating erroneous values, also called outliners, of the speed signals S, which are caused by slipping or sliding for example. In particular, only speed values S close to the estimated speed value of the previous time step are used for further treatment as validated speed signals VS.
According to an example, the gating module 24 is configured for implementing a nearest neighborhood algorithm for eliminating erroneous values of the speed signals S. The nearest neighborhood algorithm is known as such. In the present case, the gating module 24 is configured for only letting pass, as the validated speed signals VS, the or each speed signal S having a difference from each other lower than a threshold according to the nearest neighborhood algorithm.
The weightage module 26 is configured for receiving, at a current time step, a speed value of the railway vehicle 4 propagated by the Kalman filter 29 for a time step prior to said current time step, and for receiving either the speed signals S from the receiving module 22, or the validated speed signals VS from the gating module 24.
Hereafter, the operation of the weightage module 26 for treatment of received speed signals S is described. The operation of the module 26 for the validated speed signals VS instead of the speed signals S is however identical.
The weightage module 26 is configured for multiplying each speed signal S received from the receiving module 22 at the current time step with a respective weightage coefficient. The weightage coefficient is a function of the difference between the corresponding speed signal S at the current time step and the speed value propagated by the Kalman filter 29 for a time step prior to said current time step. For example, if the difference between a speed signal S and the propagated speed value is very small, the weightage coefficient may be large, because the probability that this speed value S is correct is high.
The weightage coefficient is for example a linear or quadratic function of the difference. However, other types of function may be used.
The weightage module 26 obtains, by applying the weightage coefficient to the speed signals S (or to the validated speed signals VS), weighted speed signals WS.
The weightage module 26 allows thus to stronger consider, for further treatment by the estimation module 28, the speed signals S that are closer to the propagated speed value than the speed signals S that are very different from the propagated speed value, and thus may be erroneous.
The estimation module 28 comprises the Kalman filter 29 which forms a kinematic model of the railway vehicle 4.
The Kalman filter 29 is configured for providing an estimated distance of the railway vehicle 4 from the initial position based on the speed signals.
The Kalman filter 29 is in particular a linear Kalman filter.
Kalman filters are known as such for estimation of states in function of a measurement and/or a model by using the Kalman gain, known as such. In the present case, the Kalman filter is preferably configured for estimating a distance from a starting point of the railway vehicle 4, a speed of the vehicle 4, and a drift of the speed over time steps of the Kalman filter.
According to an example, the Kalman filter 29 is configured for providing the estimated distance in the absence of a correction of said estimated distance, during operation of the railway vehicle 4, by any measurement received from a track-based beacon or a track-based transponder.
Kalman filters estimates drift generally, and in order to keep a high estimation precision, a reset of the drift is often used. According to the present description, such reset is preferably not implemented, and a high estimation precision of speed and/or distance estimation of the railway vehicle 4 is obtained solely based on the speed signals S, eventually processed prior to the Kalman filter 29 by the gating module 24 and/or the weightage module 26.
In the case of a reception of validated speed signals VS by the estimation module 28 from the gating module 24, the Kalman filter 29 is configured for providing the estimated distance in function of these validated speed signals VS instead of the speed signals received from the reception module 22.
In the case of a reception of weighted speed signals WS, the Kalman filter 29 is configured for providing the estimated distance in function of these weighted speed signals WS instead of the speed signals received from the reception module 22.
According to an example, the estimation module 28 is configured for checking the plausibility of the estimated distance, and optionally also of the estimated speed and drift of the speed of the vehicle 4. In this case, the estimation module 28 is configured for determining an intermediate estimated distance of the vehicle 4 from the initial position in function of the speed signals S, by using the Kalman filter 29. Also, the estimation module 28 is configured for determining an estimation difference between the intermediate estimated distance at a current time step and an estimated distance at a time step previous to the current time step, i.e. in the time step just before the current time step.
The "estimation difference" is the distance in meters on the rails 6 between a position calculated from the intermediate estimated distance and a position calculated from the estimated distance at the previous time step.
If the estimation difference is lower than or equal to a predetermined threshold, the estimation module 28 provides at its output the estimated distance, which is the intermediate estimated distance in this case.
If the estimation difference is higher than the predetermined threshold, the estimation module 28 provides at its output the estimated distance, which is the propagated distance at the time step previous to the current time step in this case.
According to another example, if the difference is higher, the estimation module 28 is configured for not providing any estimated distance and for sending a signal to the gating module 24 in order to reduce the error value of the module 24, so that more speed signals S are filtered. This allows then to estimate the distance-based validated speed signals VS which are only some of the speed signals S, and filtered a second time by the module 24.
The above described plausibility check allows providing only a current estimated distance at an output of the estimation module 28 if its value is plausible, i.e. not too far from the previous estimated distance. If this is not the case, as a backup, the estimation module 28 may provide the previously estimated distance.
The determination module 30 is configured for correcting the estimated distance in function of the wear coefficient so as to obtain a corrected estimated distance. For example, the corrected estimated distance is determined as equal to (1 + wear coefficient) multiplied by the estimated distance.
This allows taking into account wear of the wheels 15, which results in faster or slower rotation of the wheel 15 depending on its wear. As a consequence, the corrected estimated distance is more accurate than the estimated distance.
The determination module 30 is further configured for determining an estimated position of the railway vehicle 4 based on the corrected estimated distance and the initial position, received from the memory module 34.
The determination module 30 is further configured for comparing the estimated position with the start and end positions of the reference database 32 to detect a start and/or end of a neutral section N.
For example, the determination module 30 is configured for transmitting a signal to a driver of the vehicle 4 or to a control system of the vehicle 4 if it detects the start and end of a neutral section N. The wear coefficient of the wheel wear database 36 is reset for example at every stop of the vehicle 4 at a station and/or every passage at a start or end of a neutral section N and/or at the a final station of a mission of the vehicle 4.
The wear coefficient is susceptible to be updated as a function of a difference between the estimated distance provided by the Kalman filter 29 and a predetermined reference distance.
An example of determination of the wear coefficient is described in the following.
The determination module 30 is configured for determining a true distance error as the difference between the Kalman filter 29 estimated distance and the reference distance. The reference distance is pre-determined and stored in the wheel wear database 36, for example for a station and every passage at a start or end of a neutral section N. The determination module 30 is for example configured for determining the wear coefficient of the wheel wear database 36 on real-time in function of the true distance error.
The determination module 30 is, for example, configured for the calculation of the mean true distance error from the mean value of all the stored true distance error values. The variance of true distance is calculated by N times the standard deviation of all the true distance error values. According to an example, the wear coefficient of the wheel wear database 36 is calibrated as the multiplication of signature of mean true distance error with (variance of true distance error divided by the reference distance). If the tolerance of variance of true distance error is lower than or equal to a predetermined threshold, the wear coefficient of the wheel wear database 36 is set as zero. The gating module 24 and the weightage module 26 are optional modules.
As described above, according to an example, the railway system 1 comprises also the trackside equipment 10 which is described hereafter.
The trackside equipment 10 is configured for determining a wheel wear condition of the railway vehicle 4. The trackside equipment 10 comprises the same modules, databases, and internal connections as the detection device 20. In figure 1, the same elements have the same reference signs, increased by 100. The operation of the trackside equipment 10 is the same as the operation of the detection device 20, with the exceptions described hereafter.
The trackside equipment 10 is configured for providing a wheel wear condition for at least one wheel 15 of the vehicle 4, preferably for all wheels 15 having a first speed sensor 16.
The trackside equipment 10 comprises a trackside receiving module 122 configured for receiving the estimated distance from the detection device 20. The trackside receiving module 122 is further configured for receiving the plurality of speed signals S from on- board speed sensors 16, 18.
The trackside equipment 10 comprises a trackside estimation module 128 comprising a trackside Kalman filter 129 forming the kinematic model of the railway vehicle 4. The trackside Kalman filter 129 is configured for providing a wheel wear value W for each axle of the railway vehicle 4 having a corresponding first speed sensor 16, in function of the estimated distance and the speed signals S.
Finally, the trackside equipment 10 comprises a trackside determination module 130 configured for determining the wheel wear condition for each axle in function of the corresponding wheel wear value W. According to an example, the trackside determination module 130 is configured for emitting an alarm to the driver, to a control system of the vehicle 4 and/or to a maintenance center in function of the determined wheel wear condition.
The wheel wear condition is chosen from among a no wheel wear condition, i.e. a wheel 15 having no abrasion or nominal abrasion only, the slip condition, the slide condition and a flat wheel condition.
The slip condition and the slide condition are described above.
The flat wheel condition is a condition in which the radius of the wheel is not constant, caused for example by slipping.
The invention provides many advantages.
Thanks to the invention, the accuracy of the position detection of a neutral section N is for example lower than 80 m, preferably around 30 m.
A particular high detection accuracy is obtained by correcting the estimated distance of the Kalman filter 29 by taking into account the wear coefficient received from the wheel wear database 36. Optionally, the accuracy is further improved by filtering the signals S via the modules 24 and 26 before propagation in the Kalman filter 29.
The device 20 allows preferably using the speed signals S as only measurement input, so that track-based transponders can be omitted.
The skilled person understands that further embodiments different from those disclosed above are conceivable.

Claims

An on-board detection device (20) for a railway vehicle (4), the detection device (20) being configured for detecting a start and/or end of a neutral section (N) of a railroad infrastructure (2) on which the railway vehicle (4) is intended to run, a neutral section (N) being a non-powered section of a power line (8) of the railroad infrastructure (2), the detection device (20) comprising a wheel wear database (36) comprising a wear coefficient relative to the wear of wheels (15) of the railway vehicle (4), the detection device (20) furthermore comprising:
- a receiving module (22) configured for receiving a plurality of speed signals of the speed of the railway vehicle (4) from on-board speed sensors (16, 18) of the railway vehicle (4);

- at least one reference database (32) comprising start and end positions of the neutral sections (N) of the railroad infrastructure (2);

- at least one memory module (34) comprising at least an initial position of the railway vehicle (4);

- an estimation module (28) comprising a Kalman filter (29) forming a kinematic model of the railway vehicle (4), the Kalman filter (29) being configured for providing an estimated distance of the railway vehicle (4) from the initial position based on said speed signals; and

- a determination module (30) configured for correcting the estimated distance in function of the wear coefficient so as to obtain a corrected estimated distance, the determination module (30) furthermore configured for determining an estimated position of the railway vehicle (4) based on the corrected estimated distance and the initial position, and for comparing said estimated position with the start and end positions of the reference database (32) to detect a start and/or end of a neutral section (N).
The detection device according to claim 1, furthermore comprising a gating module (24) configured for receiving, at a current time step, a speed value of the railway vehicle (4) estimated by the Kalman filter (29) for a time step prior to said current time step, and for validating each speed signal received at the current time step having a difference to said speed value lower than an error value, so as to obtain validated speed signals;
the Kalman filter (29) being configured for providing the estimated distance as a function of the validated speed signals.
The detection device according to claim 2, wherein the gating module (24) is configured to determine the error value as equal to N times of a standard deviation of the speed signals at the current time step, N being an integer greater than 1.
The detection device according to any one of claims 1 to 3, furthermore comprising a weightage module (26) configured for receiving, at a current time step, a speed value of the railway vehicle (4) propagated by the Kalman filter (29) for a time step prior to said current time step, and for multiplying each speed signal received at the current time step with a respective weightage coefficient, which is a function of the difference between the speed signal at the current time step and said speed value of the time step prior to said current time step, so as to obtain weighted speed signals;
the Kalman filter (29) being configured for providing the estimated distance in function of the weighted speed signals.
The detection device according to any one of claims 1 to 4, wherein the wear coefficient is susceptible to be updated as a function of a difference between the estimated distance provided by the Kalman filter (29) and a predetermined reference distance.
The detection device according to any one of claims 1 to 5, wherein the Kalman filter (29) is configured to provide the estimated distance in the absence of a correction of said estimated distance, during operation of the railway vehicle (4), by any measurement received from a track-based beacon or a track-based transponder.
The detection device according to any one of claims 1 to 6, wherein each speed sensor (16, 18) is chosen from a first speed sensor type (16) operative to obtain a first speed signal from measuring a rotation of one of the wheels (15) and an axle connected to the wheel (15), and a second speed sensor type (18) operative to obtain a second speed signal from a measurement of a traction current and a voltage of a traction motor (12) of the railway vehicle (4), the traction motor (12) being connected to the wheel (15), the speed signals comprising the first speed signal and the second speed signal.
The detection device according to any one of claims 1 to 7, wherein the estimation module (28) is configured for determining an intermediate estimated distance from the initial position in function of said speed signals, and for determining an estimation difference between the intermediate estimated distance at a current time step and an estimated distance at a time step previous to the current time step;
if the estimation difference is lower than or equal to a predetermined threshold, the estimation module (28) is configured for providing the estimated distance, which is the intermediate estimated distance in this case,

if the estimation difference is higher than the predetermined threshold, the estimation module (28) is configured for providing the estimated distance, which is the estimated distance at the time step previous to the current time step in this case.
Railway vehicle (4) comprising a plurality of on-board speed sensors (16, 18) and a detection device (20) according to any one of the claims 1 to 8, the detection device (20) being configured for receiving speed signals from the on-board speed sensors (16, 18).
Railway system (1) comprising a railroad infrastructure (2) and a railway vehicle (4) according to claim 9, the railroad infrastructure (2) comprising a trackside equipment (10) configured for determining a wheel wear condition of the railway vehicle (4) intended to run on the railroad infrastructure (2), the trackside equipment (10) comprising:
- a trackside receiving module (122) configured for receiving an estimated distance from the railway vehicle (4), and for receiving a plurality of speed signals of the speed of the railway vehicle (4) from on-board speed sensors (16,18) of said railway vehicle (4),

- a trackside estimation module (128) comprising a trackside Kalman filter (129) forming a kinematic model of the railway vehicle (4), the trackside Kalman filter (129) being configured for providing a wheel wear value for each axle of the railway vehicle (4) having a corresponding speed sensor (16), in function of the estimated distance and said speed signals, and

- a trackside determination module (130) configured for determining a wheel wear condition for said each axle in function of the corresponding wheel wear value, the wheel wear condition being chosen from among a plurality of predetermined wheel wear conditions.