Method for measuring the position of a magnet using a sensor arrangement with increased magnetic field robustness, taking into account the signal-to-noise ratio.

Abstract

A method for signal processing from a magnet-sensor arrangement (1), wherein the magnet-sensor arrangement (1) comprises at least one sensor pair (3) and a magnet (2) movably mounted thereon, comprising the following steps: measuring the magnetic flux density per spatial coordinate per sensor; calculating a non-differential change in the position of the magnet (2) for each magnetic flux density measured in step a); calculating the differences of the measured magnetic flux densities per spatial coordinate; calculating a differentially calculated change in the position of the magnet (2) from the magnetic flux densities calculated in step c); normalizing the non-differentially and differentially calculated changes in the position of the magnet (2) based on a zero position of the magnet (2); calculating the differences between the differentially calculated change in the position of the magnet (2) and the non-differentially calculated changes in the position of the magnet (2);Calculate a weighting factor based on the differences calculated in step f); calculate a mean value, weighted by the weighting factor from step g), of the non-differential and differentially calculated position changes of the magnet (2) normalized in step e) in order to obtain a corrected position change.

Description

[0001] The present invention relates to a method for measuring a position of a magnet by a sensor arrangement with increased external magnetic field robustness taking into account the signal-to-noise ratio.

[0002] In many different applications, it is necessary to determine the position of a control element or sensor in the room and thus determine its operating status. This allows numerous distinguishable operating positions to be set up without the need for any mechanical locking positions or similar devices. This also has the advantage that different feedback is provided to the user depending on the current position of the control element, allowing them to react accordingly.

[0003] Typically, this position determination is carried out using an arrangement comprising a magnet, in particular a permanent magnet or electromagnet, which is usually arranged on the control element, and at least one sensor that can determine the magnetic field broken down into the relevant spatial dimensions. The measured magnetic field is usually compared with the magnetic field of the magnet located on the control element, at least in a zero position of the control element, or with a measurement under laboratory conditions, in order to determine the change in position of the magnet compared to a zero position. This determination of the position of the magnet is prone to errors in two key respects: Firstly, the measurement is distorted by signal noise, which is inherently introduced into the system by the components of the arrangement.On the other hand, the magnetic field of the magnet on the control element can be superimposed by an external field, which also distorts the measurement results of the magnetic flux density and may differ significantly from a measurement under laboratory or ideal conditions, which leads to an incorrect position determination.

[0004] For example, it is known in the prior art to use more than one sensor so that the measurement of one of the sensors can be verified by another at a different position. It is also known to use the differences in the magnetic flux densities measured by the plurality of sensors to determine the position of the magnet. This way, an external field, which ideally acts equally on both sensors, is eliminated by calculating the difference, and the calculated signal accordingly exhibits external field immunity.

[0005] Accordingly, prior art document EP 1 464 918 B1 discloses a method for measuring the position of a magnet relative to a measurement location. In this method, the magnetic field is measured by two magnetic field sensors, and the position of the magnet in space is determined by calculating the difference quotients from these measured magnetic fields. While this method has the advantage that the determined position of the magnet is corrected for external fields, the signals determined from the calculation of the difference quotients are significantly attenuated, and noise is significantly amplified relative to the signal.

[0006] In contrast, an absolutely measured signal, which is not subtracted from signals measured by other sensors, is a particularly strong signal, where noise is less significant due to component-specific tolerances. However, in this absolutely measured and unsubtracted signal, the external field remains unchanged and at its full magnitude, so that this signal may be severely distorted despite a good signal-to-noise ratio.

[0007] Consequently, the invention is based on the object of proposing a method for determining the position of a magnet, in which the determined result has both external field robustness and an advantageous signal-to-noise ratio.

[0008] This object is achieved by a method for signal processing from a magnetic sensor arrangement, wherein the magnetic sensor arrangement comprises at least one sensor pair and a magnet movably mounted thereto, comprising the following steps: a) Measurement of the magnetic flux density per spatial coordinate per sensor b) Non-differential calculation of a change in position of the magnet for each magnetic flux density measured in step a) c) Calculate the differences of the measured magnetic flux densities per spatial coordinate d) Differential calculation of a change in position of the magnet from the magnetic flux densities calculated in step c) e) Normalization of the non-differentially and differentially calculated position changes of the magnet based on a zero position of the magnet f) Calculating the differences between the differentially calculated position change of the magnet and the non-differentially calculated position changes of the magnet g) Calculating a weighting factor based on the differences calculated in step f) h) Calculating a mean value of the non-differentially and differentially calculated position changes of the magnet normalized in step e), weighted by the weighting factor from step g), in order to obtain an adjusted position change.

[0009] The magnet-sensor arrangement underlying the measurements according to step a) of the method consists of at least one sensor pair and a magnet mounted so as to be movable relative to it. For the purposes of the invention, a magnet is any device emitting a magnetic field, in particular permanent magnets or electromagnets. The mobility of the magnet is primarily relative to the at least one sensor pair, so that a change in the position of the magnet relative to a position of the sensor pair can be perceived or measured by the sensor pair. One sensor pair is advantageous for each direction along a spatial coordinate in which the magnet is movable relative to the sensor pair and the position of the magnet is to be determined. Thus, for a linear movement in one direction or for a pivoting movement around a pivot axis, only one sensor pair is required.However, if a movement in a plane or around two pivot axes is to be determined, a sensor pair is required for each movement axis, in this case two sensor pairs. Accordingly, each sensor pair is responsible for determining the position change in a corresponding direction of movement. Particularly for pivoting movements of the magnet, it is advantageous if the sensors are Hall sensors, particularly preferably 3D Hall sensors, or other 3D sensors in order to map the pivoting movement of the magnet and the corresponding magnetic flux density resulting from it in a Cartesian coordinate system. Alternatively, the use of sensors in the form of XMR sensors, for example as TMR, AMR, or GMR sensors, is also in accordance with the invention.

[0010] In the first step a) of the method, the magnetic flux density is measured in the relevant spatial directions of a Cartesian coordinate system for each sensor, so that in the case of 3D sensors, the magnetic flux densities are measured in all three spatial directions of a Cartesian coordinate system for each sensor. Thus, in the example of 3D sensors, the magnetic flux densities are obtained according to principle B. xi , B yi and B zi, where x, y and z represent the Cartesian coordinates and i is the index of the respective 3D sensor. These measured magnetic flux densities are raw data which unadulteratedly contain any external field influences and also the position of the sensors relative to a zero point of the magnet. In step b) this raw data is used to calculate a position or a change in position relative to a zero position of the magnet for each sensor. This results in a position or change in position which, due to the external influences mentioned, generally does not correspond to the actual size and therefore does not have any external field immunity. The invention defines position or change in position, or also distance, as the position in which the magnet is located in space, with the zero position of the magnet being defined as the starting point or zero point of a coordinate system in which the position is determined.Accordingly, the terms position and position change are used interchangeably in the context of this invention.

[0011] To calculate a position with external field immunity, the differences in the magnetic flux densities are calculated, whereby the difference between the magnetic flux densities measured by the two sensors of a sensor pair is calculated for each relevant spatial direction. Based on the calculation of the difference between two magnetic flux densities influenced by the same external field, this external field is removed from the result, provided the sensors are sufficiently close to each other so that the external field appears almost homogeneous in the area of ​​the sensors, or the external field is almost homogeneous, so that the result of the calculation is a magnetic flux density corrected for external fields. From these external field-corrected magnetic flux densities, a position of the magnet is determined, as in step b).Although these differences in magnetic flux densities and the resulting calculated position of the magnet are immune to external fields, the differences in the measured magnetic flux densities are significantly more susceptible to signal noise due to technical tolerances of the sensors, so that the precision of the measurements and the resulting calculated positions is reduced compared to the values ​​from process steps a) and b). The differences calculated in steps c) and d) are the differences in the measured values ​​of a sensor pair. If the process is performed for multiple sensor pairs, for example, because the position of the magnet is to be determined in multiple spatial directions, the differences in steps c) and d) are calculated for each spatial direction and for each sensor pair assigned to that spatial direction.

[0012] Although the data is largely cleared of external fields in steps c) and d), these data are still corrupted due to the position of the sensors being offset from a zero point of the magnet and do not reflect the actual position of the magnet in space. Therefore, the positions calculated in steps b) and d) are normalized based on the position of the sensors relative to the zero position of the magnet. This is particularly advantageous when done relative to an equidistance point between the two sensors of the respective sensor pair, which point is located directly below a zero position of the magnet. This enables comparability of the magnetic flux densities measured by the two sensors and a correspondingly simpler calculation of the position of the magnet.

[0013] In the next step f), the differences between the positions obtained from the different calculation methods in steps b) and d) are calculated and then normalized in step e). These results also allow conclusions to be drawn about the external field, which is present in the position from step b) but has already been eliminated in the position from step d). The difference between the positions is calculated for each sensor in step f), with the positions calculated in step b) being subtracted from the position differentially calculated in step d).

[0014] These differences in positions resulting from the different calculation methods must be taken into account when calculating the weighting factor in step g). Due to the magnitude of the difference, a correspondingly large impact of the external field on the sensor measurement results can be determined. It is therefore advantageous to incorporate as few external field influences as possible into the calculation for position determination, while simultaneously ensuring the best possible signal-to-noise ratio of the calculation factors. Accordingly, the weighting factor plays a greater role the stronger the external field.

[0015] This weighting factor is used to calculate a weighted mean of the normalized non-differentially calculated position changes from step b) and differentially calculated position changes from step d) in order to obtain a position of the magnet that is influenced as little as possible by both an external field and a deteriorated signal-to-noise ratio.

[0016] In a further development of the invention, it is proposed that the movable magnet be pivotally mounted about at least one pivot axis, and the position change is measured in the form of a deflection angle. Thus, the method is applied in a system with a pivotable lever on which a magnet is arranged. This is regularly necessary for control levers or similar devices where certain functions must be performed or specific feedback must be provided to the user based on the position of the lever.

[0017] In an embodiment of the invention, it is proposed that the deflection angle be calculated from the measured magnetic flux densities using a trigonometric function. Due to the pivoting movement of a lever on which the magnet is arranged, the magnet is movable on a circular path. This circular path of the magnet enables the deflection angle to be calculated using trigonometric functions, since the measured magnetic flux density of the magnet during a pivoting movement essentially shows the course of a sine or cosine curve as a function of the deflection angle. Since this curve is, as expected, shifted to a deflection angle of 0° due to the position of the sensors relative to a zero position of the magnet, the deflection angles calculated from the trigonometric function must be normalized accordingly in a method step e).

[0018] In a further development of the invention, it is proposed that the differentially calculated deflection angle be calculated using a trigonometric function from the differences in the measured magnetic flux densities. As with the non-differential calculation of the deflection angle, the calculation here is made possible using a trigonometric function based on the pivoting movement of the magnet.

[0019] The use of an arcsine or arccosine function is conceivable for both differential and non-differential calculations of the deflection angle. However, it is important to ensure that the result is a unique angle. This is particularly uncertain when the calculated angles approach the ±90° limit, which, due to the displacement of the sensors to a zero point, is possible even at significantly smaller, actual deflection angles of the magnet. Therefore, it is advantageous to use an arctangent function, especially an arctangent2 function, which provides a unique result.It is advisable, if possible, to calculate either the differentially or the non-differentially calculated deflection angle using a sine or cosine function, and the other deflection angle using the arctangent2 function, since the course of the differentially and non-differentially calculated magnetic flux densities as a function of the deflection angle essentially, but not exactly, corresponds to a sine or cosine curve. This allows for a diverse range of calculation methods, and any deviations from an ideal sine or cosine curve can be compensated for during processing of these calculated deflection angles through multi-point adjustment or linearization.

[0020] In an embodiment of the invention, it is proposed that the weighted mean value is formed from one of the non-differentially calculated position changes and the differentially calculated position change, with the total calculated position change being calculated using the mean value of the weighted mean values. In particular, the sums of the weighted, standardized, differentially calculated position and the weighted, standardized, non-differentially calculated positions are formed and offset against each other to form an overall result. The weighting has the effect that, in the case of a strong external field and a corresponding weighting factor, the differentially calculated position from step d) that is immune to external fields carries more weight than the non-differentially calculated position that is susceptible to external fields, so that the external field is largely eliminated and the poorer signal-to-noise ratio is accepted.However, if the external field is weak or non-existent, the external field-immune differentially calculated position is correspondingly less weighted than the non-differentially calculated position, so that the signal-to-noise ratio is improved.

[0021] In a further development of the invention, it is proposed that the standardization after step e) be carried out via a multi-point adjustment, in particular a 9-point adjustment or 15-point adjustment. By multi-point adjustment, the invention means fixing the position at several points, so that the position calculated in step b) or d) is equated with the expected position at several points, for example at nine points in a 9-point adjustment. The remaining values ​​that are not equated with the points are standardized based on a deviation of the ideal value from the calculated value, as in step e) according to the invention. As a result, the standardization is overall much more finely detailed than simple standardization and enables a much more precise determination of the position of the magnet.

[0022] In one embodiment of the invention, it is proposed that an external magnetic field influencing the measurements be determined from the difference between the two differently calculated position changes, with these position changes being corrected by the multipoint adjustment. As already explained above, the external field is determined according to the invention from the deviations in the results from the different calculation methods. The supplementation of this determination of the external field by the multipoint adjustment is understood by the invention as an application of the multipoint adjustment to the results from the different calculation methods.The data obtained using multi-point calibration almost corresponds to the actual position of the magnet, so that a deviation of the position calculated from the measured values ​​from these data allows conclusions to be drawn about distortions of the expected position, which are caused at least predominantly by an external field. In particular, multi-point calibration minimizes the influences of potential nonlinearities, gradient, or offset errors in the various methods used to calculate the position change. Accordingly, these position deviations, which must be converted back into a magnetic flux density, correspond to the external field.

[0023] In a further development of the invention, it is proposed that an external magnetic field influencing the measurements be determined from the difference between the two differently calculated position changes, whereby these position changes are corrected by comparison with a measurement under laboratory conditions. The same applies as for correction by multi-point calibration, whereby the calculated results are also corrected for a mathematical error. Such an error has already been described previously in the case of a pivoting movement of the magnet, in which the measured magnetic flux density, depending on the deflection angle, essentially, but not exactly, corresponds to a sine or cosine curve. Thus, the use of an arcsine or arccosine function to calculate the deflection angle has a mathematical error, albeit a small one.

[0024] In an embodiment of the invention, it is proposed that when calculating the weighting factor, a predefinable constant A maxis taken into account, which describes the degree of deviation between the differentially and non-differentially calculated position change, from which point onwards the signal from the differentially calculated position change according to step d) is used exclusively to calculate the position change. This constant can be predefined so that it can be changed for different applications, especially depending on the strength of an external field, but does not change dynamically due to any measurement results. By choosing the constant appropriately, the boundaries between a noisy signal and a signal that is more independent of the external field can be tuned. For example, it is advantageous if the constant is chosen in such a way that the weighting factor is correspondingly in favor of the external-field-immune, differentially calculated position and the non-differentially calculated position is correspondingly lightly weighted.

[0025] In a further development of the invention, it is proposed that when calculating the weighting factor, a predefinable constant O diff is taken into account, which can be used to specify what percentage of the differentially calculated signal is to be used to calculate the position change. Therefore, if, for example, weaker external magnetic fields are present that lead to small changes in the measured values ​​that are within an acceptable tolerance, this constant can be used to suppress excessive weighting of the differentially calculated position, so that the non-differentially calculated position with a good signal-to-noise ratio can still be used. This is particularly advantageous if the noise in the differentially calculated signal would be greater than the external field influence on the low-noise non-differentially calculated signal. This constant is the same as the constant A described above. maxpredefined and adaptable for different applications if required.

[0026] The invention is explained in more detail below with reference to the figures of a drawing. Identical parts are designated by the same reference numerals. Fig. 1: Schematic process of the procedure Fig. 2: Magnetic sensor arrangement in a preferred embodiment

[0027] Fig. 1 shows the schematic flow of the method according to the invention. The illustrated schematic flow relates to a method according to the invention in which the signals of a magnet 2 are measured and processed by two sensors 3a, 3b, or a sensor pair 3. The first two method steps a) and b) are carried out for each sensor 3a, 3b. Thus, in the first method step a), the magnetic flux density emanating from the magnet 2 is recorded for each spatial coordinate. However, the number of spatial coordinates depends on the mobility of the magnet 2 relative to the sensor 3a, 3b, so that the magnetic flux density does not necessarily have to be determined in all three spatial coordinates. Furthermore, in method step b), a non-differentially calculated change in position of the magnet 2 is calculated for each magnetic flux density measured in step a).The non-differential refers to the calculation of the position change from the absolute values ​​of the measured magnetic flux density. According to the invention, the distance of the position change is a distance or position, whereby these terms will be used interchangeably below, relative to a zero position of the magnet 2. The following steps c) and d) essentially correspond to method steps a) and b), although they are carried out only once and in particular not for each sensor. Thus, in step c), the differences between the magnetic flux densities measured in step a) are calculated for each spatial coordinate. The result is therefore, in contrast to the absolute values ​​from step a), a differentially calculated magnetic flux density, which is used in step d) analogously to step b) for the differential calculation of a position change of the magnet 2.In the subsequent method step e), both the non-differentially calculated and the differentially calculated positions of magnet 2 from method steps b) and d) are normalized based on a zero position of magnet 2. This normalization is necessary because the magnetic flux densities measured in step a) were determined by sensors 3a, 3b located at positions spaced from a zero position of magnet 2, but the position of magnet 2 is to be determined relative to this zero position. Accordingly, this error flows into the magnetic flux densities calculated from the measured magnetic flux densities in step c) and into the positions of magnet 2 calculated from them in steps b) and d), so that these values ​​must also be normalized. Subsequently, in method step f), the differences between the differentially calculated position and the non-differentially calculated positions of magnet 2 are calculated.Accordingly, each non-differentially calculated position results in a difference from the differentially calculated position of magnet 2. These differences allow conclusions to be drawn about the presence of an external field, since the non-differentially calculated position includes the external field influence, while the differentially calculated position is immune to external fields. Furthermore, these differences must be taken into account in the subsequent process step g), in which a weighting factor is calculated from these differences. Thus, based on the magnitude of the difference, a correspondingly large effect of the external field on the measurement result of the sensors can be determined, and the weighting factor is accordingly adjusted such that the differentially calculated, external-field-immune position is given greater weight.In the final method step h), a mean value of the non-differentially and differentially calculated positions of magnet 2 normalized in step e) is calculated, weighted by the weighting factor from step f), to obtain a corrected position change of magnet 2. The weighting is selected based on the calculations of method step g) such that a compromise is found between the positions in which the influences of an adverse signal-to-noise ratio and a strong external field are suppressed as advantageously as possible.

[0028] Fig.Figure 2 shows a preferred embodiment of a magnetic sensor arrangement 1 to which the method according to the invention is applied. The magnetic sensor arrangement 1 is formed from a magnet 2 and a sensor pair 3, consisting of two spaced-apart sensors 3a and 3b. The magnet is movably mounted relative to the sensors 3a, 3b. In this case, the magnet 2 is pivotally mounted about a pivot axis, with the two sensors 3a, 3b being positioned relative to one another such that their equidistance point is located below a zero position of the magnet 2. The sensor pair 3 detects the magnetic flux density emanating from the magnet 2 and, for further use of the measured data in the method, is arranged in the direction of movement of the magnet 2, so that in this case the magnet 2 can be positioned in a fully functional movement over both sensors. LIST OF REFERENCE SYMBOLS 1 magnetic sensor arrangement 2 Magnet 3 pairs of sensors 3a First sensor 3b Second sensor QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] EP 1 464 918 B1

[0005]

Claims

[1] Method for signal processing from a magnetic sensor arrangement (1), wherein the magnetic sensor arrangement (1) comprises at least one sensor pair (3) and a magnet (2) mounted movably thereto, comprising the following steps: a) Measurement of the magnetic flux density per spatial coordinate per sensor b) Non-differential calculation of a change in position of the magnet (2) for each magnetic flux density measured in step a) c) Calculate the differences of the measured magnetic flux densities per spatial coordinate d) Differentially calculating a change in position of the magnet (2) from the magnetic flux densities calculated in step c) e) Normalizing the non-differentially and differentially calculated position change of the magnet (2) based on a zero position of the magnet (2) f) Calculating the differences between the differentially calculated position change of the magnet (2) and the non-differentially calculated position changes of the magnet (2) g) Calculating a weighting factor based on the differences calculated in step f) h) Calculating a mean value, weighted by the weighting factor from step g), of the non-differentially and differentially calculated position changes of the magnet (2) normalized in step e) in order to obtain an adjusted position change. [2] Method according to claim 1, characterized by that the movable magnet (2) is pivotally mounted about at least one pivot axis and the change in position is measured in the form of a deflection angle. [3] Method according to claim 2, characterized by that the deflection angle is calculated from the measured magnetic flux densities using a trigonometric function. [4] Method according to claim 2 or 3, characterized bythat the differentially calculated deflection angle is calculated using a trigonometric function from the differences of the measured magnetic flux densities. [5] Method according to one of the preceding claims, characterized by that the weighted mean is formed from one of the non-differentially calculated position changes and the differentially calculated position change, whereby the total calculated position change is calculated using the mean of the weighted mean values. [6] Method according to one of the preceding claims, characterized by that the standardisation according to step e) is carried out via a multi-point adjustment, in particular a 9-point adjustment or 15-point adjustment. [7] Method according to claim 6, characterized bythat an external magnetic field influencing the measurements is determined from the difference between the two differently calculated position changes, whereby these position changes are corrected by the multi-point adjustment. [8] Method according to one of the preceding claims, characterized by that an external magnetic field influencing the measurements is determined from the difference between the two differently calculated position changes, whereby these position changes are corrected by a comparison with a measurement under laboratory conditions. [9] Method according to one of the preceding claims, characterized by that when calculating the weighting factor a predefinable constant A maxwhich describes the degree of deviation between the differential and non-differentially calculated position change, from which point on the signal from the differentially calculated position change according to step d) is used exclusively to calculate the position change. [10] Method according to one of the preceding claims, characterized by that when calculating the weighting factor a predefinable constant O diff is taken into account, which can be used to specify what percentage of the differential signal is to be used to calculate the position change.

Images