PRIOR ART There are several known methods for determining the position of a railbound vehicle. The position can, for example, be deter- mined point-wise by having position sensors arranged on the train, reading stationary position indicators which are placed along the length of the track. The current position of the train is then calculated by integration of the velocity of the train from the read track position. Another way to determine the position of the train is to use satellite navigation, Global Positioning System (GPS). In some applications, a very high degree of accuracy in the determination of the position is needed. The drawback with the known methods is that they cannot determine the position of the vehicle with sufficient accuracy. To determine the position of the vehicle through integration of the velocity gives large errors with time, since systematic sources of error, such as miscali-

brated instruments, are also integrated. GPS gives a poor accu- racy for moving objects and neither is the information available everywhere, such as in tunnels.

Examples of applications that demand a very high accuracy for the determination of the position are such where the position of the carriage body relative the boogie is controlled during the journey of the vehicle in order to increase the comfort for pas- sengers in the vehicle. With boogies are meant the at least in the horizontal plane rotatable wheel undercarriages, in which wheels and axes are mounted. The comfort for the passengers can, for example, be increased by compensating for the irregu- larities in the track and for acceleration stress during a journey through a curve of the track. In its applications, it is the local, position, for example the vehicle's position relative to the ir- regularities in the track or relative the curves of the track, which is important to determine with high accuracy. The global position of the vehicle, for example relative some larger city is not inter- esting in these applications.

In order to prevent that irregularities in the track are propagated to the carriage body, a spring system is arranged between the boogie and the carriage body. The spring system can be passive or active. For an active spring system, which can, for example, be constituted by a hydraulic cylinder, the spring system is con- trolled by a control signal based on measurement values during the journey from a detector, the output signal of which depends on irregularities in the track. The detector is usually constituted by an accelerometer, which measures the lateral acceleration of the vehicle. One drawback with controlling the spring system in dependence on measurements measured directly during the journey of the vehicle is that when a disturbance is detected, it is too late to make a good compensation. Also, the measure- ment signal is delayed by the treatment of the signal needed in order to obtain a suitable control signal for the spring system. In order to make a really good compensation, it is necessary to

know in advance when the next disturbance is coming. In order to predict when the next disturbance comes, the position of the vehicle needs to be known with a very high accuracy.

During a journey through a curve with a vehicle which is equipped with a spring system between the carriage body and the boogie, the acceleration force tends to push out the carriage body laterally. In order to prevent that the carriage body sur- passes the allowed vehicle profile and bumps into a lateral stop that is arranged on the vehicle, a so-called HOD (hold of device) is used. The HOD has as its assignment to hold the carriage body centered over the boogie during the cornering? The HOD is controlled by measurement values from a detector established on the train, the output signal of which depends on the curvature of the track, for example, and accelerometer, and shows the same drawback as the active spring system.

In order to decrease the stress from the lateral acceleration during a cornering, the carriage bodies are inclined towards the curve during the vehicle's journey through the curve. In order to achieve an inclination of the carriage bodies in the passenger vehicles, which are part of a connected train set, the vehicle is equipped with an especially established body inclination system, which produces the inclination of the carriage body relative to the boogies of the vehicle. The control of the body inclination for each vehicle in a train set can be achieved in different ways. An often used way is that when the control signal for the inclination of the bodies is based on the measurement values from an ac- celerometer that measures the lateral acceleration in the first boogie of the train set.

In the Swedish patent application No. 96039037, a method is shown for controlling the inclination of the carriage body in de- pendence on the geometry of a track curve, which is based on registration and storage of the ideal curve geometry of the track in advance. The position of the train during a journey is deter-

mined point-wise with some known method and the previous curve geometry data are used for controlling the inclination of the carriage body during the passage of the curves. The ideal curve geometry is obtained by registering measurement values from a previous journey filtered from noise. The ideal curve ge- ometry of the track is registered and stored in a database for each journey of the vehicle. In such a way, the geometrical data of the track is continuously updated and changes in the geome- try of the track can be used already during the next pass of the train of the section of the track. In order to provide good comfort for passengers in the vehicle, it is important that the inclination of the carriage body is correct in relation to the curve geometry of the track. If the carriage body begins to incline too late or too early, it results in less travelling comfort. Therefore, it is impor- tant that the position is determined with high accuracy. The drawback with the known methods for determining the position of the train is that they cannot determine the position with suffi- cient accuracy.

SUMMARY OF THE INVENTION The purpose of the invention is to provide a method and a de- vice which makes it possible to determine the position of a rail- bound vehicle with high accuracy during a journey and make it possible to create a control signal to compensate for irregulari- ties in the curve geometry of the track.

What characterizes a device and a method according to the in- vention is clear from the appended claims.

According to the invention, a parameter is determined during a journey over a section of the track starting from comparisons between during a journey registered measurement values and previously stored measurement values dependent on the move- ments of the vehicle caused by the curve geometry of the track and irregularities in the track. By continuously comparing, the

latest measurement values during the journey with measurement values from a previous journey over the same section of the track, a parameter can be determined that depends on the posi- tion of the vehicle within the section of the track. The deter- mined parameter contains information about the local position of the vehicle, i. e. the position relative to irregularities in the track or relative to the curve geometry of the track, and can, for ex- ample, be used to determine the global position of the vehicle or to generate a signal to control the position of the carriage body relative to the boogie in order to compensate for the curve ge- ometry of the track and irregularities in the track.

The registered measurement values depend on the movements of the vehicle caused partly by the curve geometry of the track and partly by the irregularities of the track. In one embodiment of the invention, a series of measurement values are separated from the registered measurement values, the first series of which depends on the movements of the vehicle caused by the irregularities, wherein said comparison is carried out between the measurement values from the first series. What is separated from the measurement signal is actually the noise, and the pa- rameter is determined by comparing measured and stored noise.

In the noise, there is a characteristic pattern that looks different dependent on where in the section of the track the train is, but that looks the same each time the train passes the same spot on the track. By continuously comparing the pattern in the new noise with the pattern in the stored noise, a parameter that de- pends on the local position of the vehicle can be determined with high accuracy.

In one embodiment of the invention, the parameter constitutes a control parameter for controlling of the position of the carriage body relative to the boogie. In one embodiment example, the control parameter is used to correct for irregularities in the track. The control parameter is then generated in dependence on the stored measurement values that depend on the irregu-

larities in the track and the comparison. A control parameter can, for example, be used to generate a control signal to control an active spring system between the carriage body and the boogie. By controlling the active spring system between the car- riage body and the boogie with a control signal that is based on stored and already treated measurement values instead of di- rectly measured values, the compensation for irregularities in the track is carried out in the right time and thus becomes bet- ter. It is the accurate local determination of the position that makes it possible that the control signal can be based on stored measurement values instead of on directly measured measure- ment values.

In another embodiment of the invention, the control parameter is used to correct for the curve geometry of the track. From the registered measurement values, a second series of measure- ment values is separated, which depends on the movements of the vehicle caused by the curve geometry of the track. The con- trol parameter is generated in dependence on these measure- ment values of the curve geometry of the track stored during a previous journey and said comparison. The control parameter can be used to generate a control signal to control the inclina- tion of the carriage body in dependence of the curve geometry of the track. The control parameter can also be used to generate a control signal to center the carriage body over the boogie thus avoiding that the carriage body surpasses the profile of the ve- hicle. In that the local position of the vehicle can be determined with high accuracy, the position of the carriage body can be controlled in time, relative to its entrance into the curves, so that maximum comfort is achieved.

BRIEF DESCRIPTION OF THE DRAWINGS One embodiment example of the device according to the inven- tion shall hereafter be described with support from the appended

drawing. The figure shows a device to determine a parameter for a railbound vehicle.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS OF THE INVENTION The ideal geometry of a track consists of straight sections, cir- cular curves with a constant radius, and transition curves that connects the straight sections with the circular curves. In the curves, the track is cumbered, which means that the inner rail is lower than the outer rail. A device for determining a parameter according to the invention is shown in the figure. The device comprises a detector member 1, the output signal S of which depends of the movements of the vehicle when it passes over the track. The movements of the vehicle depends in turn on the radius of curvature of the curves or the rail elevation angle, i. e. the difference in height between the inner and the outer rail. The detector member can, for example, be a accelerometer, which measures the lateral acceleration of the vehicle, or a gyro that measures angular velocity either along a length axis (the roll angular velocity) or around a vertical axis (the turn angular ve- locity).

The signal (S) from the detector shown in the figure depends on the inverse of the radius of the curve and has dynamical distur- bances caused by irregularities in the track super-positioned.

The dynamical disturbances in the signal is a form of noise and will hereafter be named as noise. According to the invention, the noise is used to determine a parameter that depends on the po- sition of the vehicle and therefore the noise is separated from the measurement signal. The noise is separated by passing the measurement signal through a filter member 2 where the noise is filtered off from the measurement signal. Since the curves of the track ideally have a constant radius, the signal that repre- sents the ideal track geometry will consist of a number of straight lines that are tied together in certain points. In one

preferred embodiment, the filtering takes place in a computer by adapting straight lines to the curve in an optimal way. This adaption can, for example, be carried out by using the least square method. When the signal representing the ideal curve geometry, hereafter called the ideal signal A, has been pro- duced, the noise signal B can be separated from the original measurement signal S by subtraction of the ideal signal A from the measurement signal in a member 3.

When the signals have been separated, the noise signal B and the ideal signal A are stored in a database in a memory 4. In the database there are already separated noise signals BL and ideal signals AL stored from a previous journey over the same section of the track. The stored signals are updated each time the vehi- cle passes over the section of the track so that updated signals are always available. The new signals A, B and the signals from the previous journey AL, BL are stored separately. A section of the new noise signal, for example, the latest ten seconds, are compared with the stored noise signal BL from the previous journey in a comparison member 5. The comparison can, for ex- ample, be carried out in such a way that the difference is calcu- lated between the latest section of the new noise signal and a number of sections of the noise signal from the previous journey that are somewhat displaced in time relative to each other. The calculated differences are compared in order to find the section of the stored signal that gives the smallest difference between the signals. When the section that gives the smallest difference has been found, the position in the stored noise signal is known.

In such a way, it is possible to follow the stored noise signal during the journey and thus obtain a position determination relative to the noise. The noise in turn depends on irregularities in the track and is therefore obtained from an indirect position determination relative to the track. In the comparison member 5 is concluded which section that has the smallest difference and

a parameter p is generated that depends on the position of the vehicle relative to the noise.

If the difference never falls below a given minimum value, it can depend on the fact that a change of the track has taken place, since the last journey. Now, it is important not to loose the posi- tion and to find the correct location in the noise signal after the vehicle has passed the change. It can therefore be advanta- geous if there is a second coarser position determination that nevertheless gives an indication of where the vehicle is posi- tioned on the track. For this purpose, the stored ideal signal can be used. By localizing the break points in the ideal curve, a rough position determination is obtained which can be used to find a suitable starting-point to find the correct location in the noise signal. The break points can be positioned by comparing the stored ideal signal from the previous journey with the new ideal signal.

In this embodiment example, the parameter p is used which is a measure on the position of the vehicle relative to the noise, in order to control the inclination of a railway carriage body of the vehicle relative to the boogie during the passage of the vehicle through a curve of the track. In that the position relative to the noise is known, the position relative to the stored ideal signal can be determined. The stored ideal signals AL in the database and the parameter p from the comparison member 5 constitute input data to a computing member 6 that computes a control pa- rameter q which is a measure of the position of the vehicle rela- tive to the ideal curve of the track. From the control parameter q, the inclination of the carriage body is then controlled in a known way by a body inclination system.

In another embodiment of the invention, the control parameter q can be used to center the carriage body over the boogie during a cornering of a curve with the vehicle. The control parameter q then constitutes input data to a system that has at its assign-

ment to keep the carriage body in place during the passage of the curve (HOD). The system thus obtains information that makes it possible to control the carriage body relative to the boogie in dependence on the position of the vehicle relative to the ideal curve geometry.

The position determination relative to the noise can also be used to control an active spring system between the carriage body and the boogies, see the figure. The control of the spring system is carried out in dependence on the stored noise BL from a previous journey and the parameter p. Thanks to the fact that the position relative to the noise is known through the parameter p, it is possible to follow in the stored noise signal and to control the spring system in time so as to compensate for irregularities in the track. In a computing member 7, a control parameter r is produced from the parameter p and the stored noise BL from the database. The control parameter r constitutes input data to a system that controls the active spring system.

In another application of the invention, the parameter p can be used to determine the global position of the vehicle, that is the position relative to a fixed point in the surroundings. In order to be able to determine the global position of the vehicle there must be position data stored in the database that can be cou- pled to the parameter p. The global position of the vehicle is computed in a computing member and the parameter p and po- sition data from the data base constitutes the input to the com- puting member.