Dynamic reconstruction component and method for the dynamic reconstruction of input signals in an open loop coriolis vibration gyroscope
The dynamic reconstruction component for OLCVGs addresses frequency-dependent compensation issues by using adjustable filters and correlation circuits to enhance sensor accuracy and reliability, improving performance parameters and reducing crosstalk.
Patent Information
- Application Number
- PCT/EP2026/062026
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-12
- Filing Date
- 2026-04-27
- Publication Date
- 2026-11-19
AI Technical Summary
Conventional methods for open-loop Coriolis vibrating gyroscopes (OLCVGs) face challenges in dynamically compensating for frequency-dependent amplitude and phase responses, leading to erroneous measurements due to environmental changes and interference from test signals, which are not effectively addressed in the prior art.
A dynamic reconstruction component with adjustable filters is implemented to reconstruct input signals of OLCVGs, using a reconstruction filter matrix to compensate for amplitude and phase responses, and a correlation circuit to evaluate and correct sensor parameters, thereby improving sensor transfer function, bandwidth, and reducing crosstalk.
The dynamic reconstruction component enhances the accuracy and reliability of OLCVGs by dynamically compensating for frequency-dependent responses, reducing crosstalk, and enabling continuous monitoring and improvement of performance parameters such as sensitivity and zero-rate output.
Smart Images

Figure EP2026062026_19112026_PF_FP_ABST
Abstract
Description
[0001] R.417457
[0002] - 1 -
[0003] Description
[0004] title
[0005] Dynamic reconstruction component and method for the dynamic reconstruction of input signals in an open-loop Coriolis vibrating gyroscope
[0006] The invention relates to the reconstruction of input signals of an open Coriolis vibration gyroscope (OLCVG) from demodulated signals and a method for operating an open Coriolis vibration gyroscope, wherein reconstructed signals are calculated by applying a reconstruction filter matrix which correspond to the input signals applied to the at least one detection oscillator of the open Coriolis vibration gyroscope.
[0007] State of the art
[0008] The invention relates to Coriolis vibratory gyroscopes (CVG) in which an angular velocity to be measured is determined directly from the amplitude of the resulting detection oscillation, i.e., no position control of the detection oscillation is used.
[0009] These rotation rate sensors, referred to as "open loop" or open Coriolis vibrating gyroscopes (CVGs) (hereinafter abbreviated as OLCVG), are distinguished from so-called "closed loop" CVGs with position control of the detection oscillation primarily by their lower manufacturing costs. Despite inherent disadvantages, open loop Coriolis vibrating gyroscopes achieve R.417457
[0010] - 2 -
[0011] OLCVG increasingly shows similar performance parameters compared to "closed loop" CVG.
[0012] The input signals of an OLCVG (Online Color Volume Verification Generator) consist of the angular velocity or rotation rate to be measured, the quadrature, forces due to external vibration, and any test signals that may be superimposed on the quadrature or rotation rate. Dynamic reconstruction means that the input signals are dynamically reconstructed from the signals demodulated according to rotation rate and quadrature using suitable transfer functions or filters. The use of dynamic reconstruction of the input signals of an OLCVG is not known in the prior art.
[0013] The transfer function of an open-loop Coriolis vibrating gyroscope (OLCVG), particularly the transfer function from input rotation rate to demodulated rotation rate signal, is frequency-dependent. This frequency dependence arises primarily from the transfer functions of the detection mode and the low-pass filters used during demodulation and potentially also as output filters. Low-pass filters can be used to compensate for the amplitude response of the detection mode's transfer function, but this simultaneously amplifies changes in the phase response. Individual compensation of the amplitude response is not currently possible and is not known in the prior art. This reconstruction allows for individual and extensive compensation of both amplitude and phase response.
[0014] Environmental changes (temperature, stress, e.g., after soldering or due to component aging) of an open-loop Coriolis vibrating gyroscope (OLCVG) lead to changes in the mechanical transfer function and thus to erroneous measurement signals. Conventional methods exist in which environmentally induced changes in phase or amplitude response are statically compensated. This means that correct compensation of phase and amplitude only occurs for static input rotation rates with a frequency equal to zero.
[0015] R.417457
[0016] - 3 -
[0017] In conventional methods, test signals are injected into test electrodes, thereby exciting the detection mode. The resulting measured sensor signal – usually after demodulation with the frequency of the drive movement – is correlated with the test signal. This correlation can be used for a functional test and / or to compensate for the rotation rate signal.
[0018] Feedback electrodes (no modulation of the force on the detection mode with the drive movement), quadrature electrodes (modulation of the force on the detection mode with the drive movement) and drive electrodes for exciting the drive oscillations can be used as possible test electrodes.
[0019] In conventional methods where test signals are fed onto quadrature electrodes, an intrinsic disadvantage is that the test signal – with the exception of DC signals, which do not allow continuous functional testing or compensation – is coupled into the rotation rate signal even in the ideal case, because the frequency-dependent transfer function of the detection oscillation is not taken into account.
[0020] In conventional circuits, interference with the gyroscope signal from the test signal is usually avoided by placing the test signal, or its response signal, outside the bandwidth of the output filter, as described in US 8910518. However, this method does not allow for the determination of all sensor parameters, such as phase and sensitivity. With broadband signals, unaccounted-for signal distortion can lead to an incorrect correlation between the measured and input signals. In principle, a very low-frequency test signal with frequency components, e.g., below 20 Hz, can also be used, for which crosstalk and thus interference in the measured gyroscope signal are small. However, this does not allow for rapid fault detection.
[0021] R.417457
[0022] - 4 -
[0023] Disclosure of the invention
[0024] According to a first aspect, the invention provides a dynamic reconstruction component for an open loop Coriolis vibrating gyroscope (OLCVG), wherein the dynamic reconstruction component has filters whose transfer functions are designed to reconstruct input signals of the open loop Coriolis vibrating gyroscope from demodulated signals of the open loop Coriolis vibrating gyroscope.
[0025] The reconstruction component of the open-loop Coriolis vibrating gyroscope can be implemented in hardware circuitry, firmware, or software.
[0026] In one possible embodiment, the reconstruction component can be formed by a digital reconstruction filter with adjustable filter coefficients.
[0027] The reconstruction preferably takes place in real time.
[0028] The open-loop Coriolis vibratory gyroscope can also be referred to as an open Coriolis vibratory gyroscope. The reconstruction component and method according to the invention perform a dynamic reconstruction of the input signals of an open-loop Coriolis vibratory gyroscope, which is not known in the prior art.
[0029] The dynamic reconstruction according to the invention can improve the sensor transfer function, bandwidth, sensor noise and size of the open loop Coriolis vibrating gyroscope.
[0030] The reconstruction component and method according to the invention enable a dynamic reconstruction of the input signals of an open-loop Coriolis vibrating gyroscope. This allows methods based on injected test signals for determining internal sensor parameters to be improved or even made possible in the first place.
[0031] R.417457
[0032] - 5 -
[0033] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, the demodulated signals of the open loop Coriolis vibration gyroscope include a demodulated rotation rate signal and a demodulated quadrature signal.
[0034] In one possible embodiment of the dynamic reconstruction component according to the invention for an open-loop Coriolis vibration gyroscope, the input signals of the open-loop Coriolis vibration gyroscope include a rotation rate input signal, a quadrature input signal, a vibration input signal, and / or a test signal. This allows for flexible use in different applications.
[0035] In one possible embodiment of the dynamic reconstruction component according to the invention for an open-loop Coriolis vibration gyroscope, the reconstruction component generates a reconstructed rotation rate signal and a reconstructed quadrature signal. These can be evaluated by downstream circuits, in particular by a correlation circuit.
[0036] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, a correlation circuit connected to the reconstruction component has correlation units designed to correlate the reconstructed rotation rate signal and the reconstructed quadrature signal emitted by the reconstruction component with signal components of at least one test signal generated by a test signal generator and applied to a quadrature signal input of the open loop Coriolis vibration gyroscope for the purpose of calculating correlation values.
[0037] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, R.417457
[0038] - 6 -
[0039] The correlation units of the correlation circuit are further designed to calculate autocorrelation values for the reconstructed rotation rate signal and the reconstructed quadrature signal output by the reconstruction component.
[0040] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, an evaluation unit is designed to evaluate the correlation values calculated by the correlation units of the correlation circuit and / or the autocorrelation values calculated by the correlation units of the correlation circuit for recalibration of filters of the dynamic reconstruction component and / or for sensor monitoring of the open loop Coriolis vibration gyroscope.
[0041] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, a vibration compensation circuit connected to the reconstruction component is designed to evaluate the reconstructed rotation rate signal and the reconstructed quadrature signal emitted by the reconstruction component for the purpose of compensating for vibrations.
[0042] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, the reconstruction component comprises compensation filters for compensating crosstalk from input signals to demodulated signals.
[0043] The reconstruction component preferably has a first compensation filter and a second compensation filter designed such that the sensor transfer function with respect to a rotation rate input signal to a rotation rate output signal is flat and without phase shift.
[0044] R.417457
[0045] - 7 -
[0046] The reconstruction component preferably includes a third compensation filter and a fourth compensation filter, which are designed such that a sensor transfer function with respect to a quadrature input signal to a quadrature output signal is flat and without phase shift.
[0047] In one possible embodiment of the dynamic reconstruction component according to the invention for an open loop Coriolis vibration gyroscope, a low-pass filter is connected downstream of each of the compensation filters of the dynamic reconstruction component.
[0048] In one possible embodiment of the dynamic reconstruction component according to the invention for an open-loop Coriolis vibration gyroscope, the first compensation filter and the second compensation filter, as well as their respective serially connected low-pass filters, provide filtered output signals which are fed to a first adder of the reconstruction component, which adds the two filtered output signals to calculate a reconstructed rotation rate signal, and wherein the third compensation filter and the fourth compensation filter, as well as their respective serially connected low-pass filters, provide filtered output signals which are fed to a second adder of the reconstruction component, which adds the two filtered output signals to calculate a reconstructed quadrature signal.
[0049] The reconstruction component can be implemented as a hardware or software module. Preferably, dynamic reconstruction is performed in real time by a hardware module. This module, for example, forms part of an ASIC.
[0050] According to a further aspect of the invention, an open loop Coriolis vibratory gyroscope (OLCVG) for measuring an angular velocity is provided, wherein the open loop Coriolis vibratory gyroscope has a dynamic reconstruction component according to the first aspect of the invention.
[0051] R.417457
[0052] - 8 -
[0053] In the preferred embodiment, the open loop Coriolis vibratory gyroscope is a MEMS open loop Coriolis vibratory gyroscope.
[0054] In one possible embodiment, the open loop Coriolis vibration gyroscope according to the invention is integrated into an IMU (Inertial Measurement Unit) of a device.
[0055] According to a further aspect, the invention provides a method for reconstructing input signals of an open loop Coriolis vibrating gyroscope (OLCVG), wherein demodulated signals of the open loop Coriolis vibrating gyroscope are filtered by adjustable filters of a reconstruction component for the reconstruction of the input signals.
[0056] According to a further aspect of the invention, a test component (TC) provides a test signal generator for generating a test signal which is applied to a signal input of an open loop Coriolis vibration gyroscope, which has a dynamic reconstruction component for calculating a reconstructed rotation rate signal and a reconstructed quadrature signal, wherein the reconstructed rotation rate signal and the reconstructed quadrature signal are correlated by a correlation circuit of the test component with signal components of the generated test signal for calculating correlation values, which are evaluated by an evaluation unit of the test component for tracking filter coefficients of compensation filters of the reconstruction component.
[0057] The test signal generator of the test component can, for example, be implemented on an ASIC that also includes the reconstruction component according to the invention.
[0058] According to a further aspect, the invention provides a method for operating an open Coriolis vibration gyroscope, wherein signals reconstructed from the signals demodulated according to gyroscope rate and quadrature are calculated by applying a reconstruction filter matrix, which R.417457
[0059] - 9 -
[0060] correspond to the input signals applied to at least one detection transducer of the open Coriolis vibration gyroscope.
[0061] According to a further aspect, the invention provides an open Coriolis vibration gyroscope with a drive control loop comprising a drive oscillator, at least one detection oscillator and a reconstruction component comprising an adaptive reconstruction filter matrix designed to calculate reconstructed signals from signals demodulated according to gyroscope rate and quadrature, which correspond to the input signals applied to at least one detection oscillator of the open Coriolis vibration gyroscope.
[0062] Further possible embodiments, developments, and implementations of the invention also include combinations of features of the invention described previously or subsequently with regard to the exemplary embodiments, even if not explicitly mentioned. In particular, the person skilled in the art will also be able to combine individual aspects as improvements or additions to the respective embodiments and developments described above, as appropriate, and add them to the basic form of the present invention.
[0063] The present invention will be explained in more detail below with reference to the exemplary embodiments shown in the schematic figures of the drawings.
[0064] This shows:
[0065] Fig. 1 shows a schematic block diagram of a single-axis open
[0066] Loop Coriolis vibration gyroscope OLCVG with separated resonance frequencies;
[0067] Fig. 2 shows a simplified block diagram of a single-axis open circuit.
[0068] Loop Coriolis vibration gyroscope OLCVG with separated resonance frequencies;
[0069] R.417457
[0070] - 10 -
[0071] Fig. 3 shows a circuit diagram of a conventional static
[0072] Compensation circuit;
[0073] Fig. 4A, 4B shows an amplitude and phase response of transfer functions of a sensor core;
[0074] Fig. 5 shows a simplified block diagram of the detection of a single-axis open-loop Coriolis vibrating gyroscope OLCVG with separate resonance frequencies and subsequent reconstruction without compensation of vibration forces;
[0075] Fig. 6 shows a block diagram for compensating for
[0076] Vibration input signals using a possible embodiment of a reconstruction component according to the invention;
[0077] Fig. 7A, 7B shows an amplitude and phase response of transfer functions of a sensor core and of filters in a possible embodiment of a reconstruction component according to the invention;
[0078] Fig. 8 shows a block diagram of a possible embodiment of a reconstruction component according to the invention for selective reconstruction where crosstalk is compensated;
[0079] Fig. 9A, 9B shows an amplitude and phase response of transfer functions;
[0080] Fig. 10 shows a block diagram illustrating sensor monitoring and tracking of sensor parameters under
[0081] R.417457
[0082] - 11 -
[0083] Use of a reconstruction component according to the invention;
[0084] Fig. 11 shows a block diagram illustrating sensor monitoring and tracking of sensor parameters with additional compensation of vibrations using a reconstruction component according to the invention;
[0085] Fig. 12 shows a block diagram illustrating sensor monitoring and tracking of sensor parameters with selective compensation using a reconstruction component according to the invention;
[0086] Figs. 13 to 21 are block diagrams illustrating further possible embodiments of an open-loop Coriolis vibration gyroscope according to the invention.
[0087] The accompanying drawings are intended to provide a further understanding of the embodiments of the invention. They illustrate embodiments and, in conjunction with the description, serve to explain the principles and concepts of the invention. Other embodiments and many of the advantages mentioned will become apparent with reference to the drawings. The elements of the drawings are not necessarily shown to scale.
[0088] In the figures of the drawing, identical, functionally equivalent and similarly acting elements, features and components - unless otherwise stated - are each provided with the same reference symbols.
[0089] The REK reconstruction component and the method according to the invention enable the reconstruction of the
[0090] R.417457
[0091] - 12 -
[0092] Input signals of an open loop Coriolis vibration gyroscope, which can also reduce and ideally completely compensate for interfering crosstalk of a test signal TS into the rotation rate output signal.
[0093] In drive mode, the vibration gyroscope is set into oscillation by an excitation unit. This oscillation can occur along a specific axis or in a mode that provides the desired sensitivity for measuring the rotational speed. In one possible embodiment, the excitation occurs close to the resonant frequency of the drive mode of the Open Loop Coriolis Vibration Gyroscope OLCVG to maximize the oscillation amplitude. Once the Open Loop Coriolis Vibration Gyroscope OLCVG is set into rotation, the Coriolis force acts on an oscillating mass inside the gyroscope in detection mode. This force causes a change in the oscillation amplitude that is directly proportional to the rotational speed. The change in the oscillation amplitude is measured and recorded to determine the rotation rate.
[0094] The following section primarily describes embodiments in which test signals are fed onto quadrature electrodes. Dynamic reconstruction is independent of this. The application of dynamic reconstruction described below can be transferred to methods in which other electrodes, or more generally, force transmitters, are used for feeding test signals.
[0095] In particular, the compensation of vibration patterns V is independent of the force transmitters used to input the test signals TS. It may be necessary to adapt the methods described below to increase robustness.
[0096] In one possible embodiment, the reconstruction component REK and the method according to the invention comprise filtering the measured quadrature signal before subtracting it from the R.417457.
[0097] - 13 -
[0098] Rotation rate signal, which allows the transfer function of the detection oscillation to be included by the filter.
[0099] The reconstruction thus allows test signals TS to be injected, for example, tones or narrow- or broadband noise onto the quadrature electrodes of the open-loop Coriolis vibration gyroscope OLCVG, and largely prevents their crosstalk onto the measured gyroscope rate signal. Therefore, the reconstruction according to the invention fulfills a prerequisite for the use of test signals TS whose frequency components lie within the bandwidth of the gyroscope rate sensor.
[0100] Simultaneously, the reconstruction largely prevents crosstalk between rotation rate input signals R and the reconstructed quadrature signal □'. Thus, in the above-mentioned embodiment, the reconstructed quadrature signal corresponds to the superposition of intrinsic quadrature force, which changes only very slowly (e.g., due to temperature changes), and the injected test signal Ts, and is not, or only very minimally, disturbed by rotation rate input signals. Since the transmitted test signal TS can be used to monitor sensor characteristics of the open-loop Coriolis vibrating gyroscope OLCVG or to adjust filter coefficients W to the sensor parameters, this suppression of crosstalk is an essential prerequisite for a robust method that delivers reliable results even at high input rotation rates.Such compensation is not known in the prior art and also leads to disturbances with test signals TS outside the bandwidth of the output filter, since the coupling of the rotation rate input signals to the quadrature signal occurs in the mechanics and cannot be avoided.
[0101] The inventive reconstruction component REK and the inventive method further enable the vibration robustness of an open loop Coriolis vibrating gyroscope OLCVG by compensation R.417457
[0102] - 14 -
[0103] to increase the false signals VF ZU. Such a method for compensating for false signals caused by vibrations is not known in the prior art.
[0104] The inventive reconstruction component REK and the inventive method allow monitoring methods of open loop Coriolis vibratory gyroscopes OLCVG to be improved, as well as performance parameters of open loop Coriolis vibratory gyroscopes OLCVG, in particular sensitivity, zero-rate output, vibration sensitivity and frequency-dependent transfer function of the open loop Coriolis vibratory gyroscope OLCVG.
[0105] The inventive reconstruction component REK and the inventive method further enable technical systems which have an open loop Coriolis vibrating gyroscope OLCVG to be improved in their reliability and in their performance parameters.
[0106] The inventive reconstruction component REK and the inventive method make it possible to directly improve performance parameters by dynamically reconstructing the OLCVG input signals, in particular to compensate for the frequency dependence of the sensor transfer function from input rotation rate to measured rotation rate and thus to achieve a flat transfer function of both amplitude and phase.
[0107] Furthermore, by compensating the sensor transfer function, the frequency difference between the drive and detection resonance frequencies can be reduced without limiting the possible sensor bandwidth of the OLCVG open-loop Coriolis vibrating gyroscope. A reduced frequency difference improves the signal-to-noise ratio. This can be used to improve sensor noise and / or to achieve a smaller size for the OLCVG. The size directly impacts the unit cost.
[0108] R.417457
[0109] - 15 -
[0110] In so-called dual-resonant operation (where the resonant frequencies of the drive and detection modes are the same within the manufacturing tolerances and, if necessary, after adjustment, which can also be performed continuously), the inventive reconstruction component REK and the inventive method can be used to increase the sensor bandwidth of the open-loop Coriolis vibrating gyroscope OLCVG. Among other things, the limited bandwidth of dual-resonant operation is often a significant disadvantage. The inventive reconstruction component REK and the inventive method eliminate this disadvantage, thereby making the key advantage of dual-resonant operation—namely, a significantly increased sensor signal (e.g., by a factor of 50)—utilizable.
[0111] The reconstruction component REK and the method according to the invention enable the monitoring and, if necessary, the improvement of the performance parameters of an open-loop Coriolis vibrating gyroscope OLCVG. By injecting test signals TS onto quadrature electrodes, as shown in Fig. 10, the phase and sensitivity of the open-loop Coriolis vibrating gyroscope OLCVG can be determined. This determination is significantly more accurate and robust than with conventional methods due to the reconstruction. This allows for improved monitoring in terms of robustness and accuracy. By adjusting the sensitivity and phase, the zero-rate output and the sensor's sensitivity can be improved. These procedures can be performed continuously, upon detection of a rest position, or initiated by the user.
[0112] The inventive reconstruction component REK and the inventive method also allow the vibration robustness of an open loop Coriolis vibrating gyroscope OLCVG to be improved.
[0113] Vibrational forces with frequencies close to the drive frequency can directly and unintentionally excite the DM detection mode. In the case of so-called "unclosed" modes, linear or rotational accelerations can excite the DM detection mode. (See vibration-resistant R.417457.)
[0114] - 16 -
[0115] In designs with "closed modes," excitation can occur due to asymmetries caused by manufacturing tolerances. This excitation leads to erroneous signals in the demodulated signals with the difference frequency between the drive frequency and the vibration frequency. It has been shown that the signals in the gyroscope and quadrature channels are phase-shifted by 90° and each exhibits the same amplitude. The method according to the invention can largely eliminate the erroneous signal in the gyroscope output. The method is only reliable if dynamic reconstruction, or at least parts thereof, are used. Furthermore, under certain conditions, the method according to the invention can also largely eliminate erroneous signals in the reconstructed quadrature Q'. This further increases the robustness and accuracy of methods in which the reconstructed quadrature Q' is used for monitoring or improving performance parameters.
[0116] The reconstruction component and method according to the invention perform a dynamic reconstruction of the input signals of an open-loop Coriolis vibrating gyroscope (OLCVG) not known in the prior art. Using the dynamic reconstruction according to the invention, the sensor transfer function, bandwidth, sensor noise, and size of the OLCVG can be improved.
[0117] The inventive method for the dynamic reconstruction of the input signals of an open-loop Coriolis vibration gyroscope (OLCVG) can improve or even enable methods based on injected test signals for determining internal sensor parameters. When a quadrature test signal is injected, conventional methods do not dynamically compensate for crosstalk between the test signal TS and the gyroscope's rotation rate signal. Therefore, the test signal TS must lie sufficiently far outside the sensor's bandwidth to avoid appearing as interference in the rotation rate output. Dynamic reconstruction according to the inventive method can address this issue.
[0118] R.417457
[0119] - 17 -
[0120] In contrast, the crosstalk of the test signal TS into the measured or reconstructed rotation rate signal R' is almost completely suppressed.
[0121] The reconstructed quadrature signal Q' is preferably correlated with the test signal TS, e.g., by multiplication or cross-correlation. In one possible embodiment, the result is used to monitor the sensor function and to determine or correct parameters such as phase and sensitivity. Input rotation rates inevitably cross-reference to the quadrature signal on the mechanical sensor element in a frequency-dependent manner. With conventional methods, this disrupts the correlation between the test signal and the quadrature signal. This can lead to faulty monitoring or incorrectly determined phase and sensitivity. The dynamic reconstruction according to the invention also largely suppresses cross-referencing from a rotation rate input signal R into the measured (reconstructed) quadrature signal Q'. This results in significantly more robust monitoring and / or more accurate determination of phase and sensitivity.
[0122] In one possible embodiment of the method according to the invention, the vibration robustness of the open-loop Coriolis vibrating gyroscope OLCVG is improved. Vibrational forces with frequencies close to the drive frequency can undesirably excite the detection mode DM. This excitation leads to erroneous signals in the demodulated signals with the difference frequency between the drive frequency and the vibration frequency. The signals in the rotation rate and quadrature channels are phase-shifted by 90° but have the same amplitude. According to the invention, in one possible embodiment, this fixed phase-amplitude relationship is exploited, and the quadrature signal is phase-shifted by 90°, e.g., using a Hilbert transform filter, and then subtracted from the rotation rate signal.
[0123] R.417457
[0124] - 18 -
[0125] This allows the erroneous signal in the gyroscope output to be largely eliminated. The method is only reliable if the quadrature signal used for compensation essentially contains only the signal generated by the vibration, as otherwise the compensation will generate erroneous signals in the gyroscope output. By using the dynamic reconstruction component REK according to the invention, or at least parts thereof, it can be largely ensured that the crosstalk from gyroscope input signals R to the measured or reconstructed quadrature signal Q' is compensated before the reconstructed quadrature signal Q' is used to compensate the vibration signals in the gyroscope output. Furthermore, under certain conditions, the method according to the invention also makes it possible to largely eliminate an erroneous signal in the reconstructed quadrature signal Q.This is advantageous when the reconstructed quadrature signal Q' is used to monitor or improve performance parameters.
[0126] Figure 1 shows a highly simplified block diagram of a single-axis open-loop Coriolis vibrating gyroscope (OLCVG) with separate resonant frequencies for the drive and detection modes described below. The method is also applicable to open-loop Coriolis vibrating gyroscopes (OLCVG) with so-called double-resonant operation, meaning the two resonant frequencies are designed to be as close to the same as possible. The descriptions apply analogously to multi-axis open-loop Coriolis vibrating gyroscopes (OLCVG).
[0127] The mechanical component MECH of the open-loop Coriolis vibrating gyroscope OLCVG consists of a drive mode AM and a detection mode DM. The latter is described by a transfer function Hsense, as shown schematically in Fig. 1.
[0128] A phase-locked loop (PLL) and an amplitude gain control (AGC) advantageously excite the AM drive mode with a fixed amplitude at its resonant frequency. The phase-locked loop (PLL) and the R.417457
[0129] - 19 -
[0130] The amplitude controller (AGC) of the open-loop Coriolis vibration gyroscope (OLCVG) shown in Fig. 1 can be implemented analogously or digitally. The output signals of the phase controller (PLL) and the amplitude controller (AGC) are converted into a force on the drive mode (AM) via force transducers. The displacement of the drive mode (AM) is converted into a current or voltage by a signal converter. The signal converter and force transducer are not shown in Fig. 1.
[0131] An angular velocity R* around the sensitive axis (perpendicular to the drive motion) results in a Coriolis force Fc on the detection mode DM perpendicular to the angular velocity and to the drive oscillation. In vector notation, equation (1) applies:
[0132]
[0133] 15
[0134] This is m sense The mass of the detection mode DM and Vdrive is the speed of the detection mode DM in the drive direction.
[0135] Is the deflection of the drive mode with
[0136] x drive = x0sin(ω drive t)
[0137]
[0138] As described, the magnitude of the Coriolis force Fc can be applied to the detection mode DM with
[0139] F c = k c x0cos(ω drive t)R*
[0140] (3)
[0141] be specified (kc = 2 msense codrive).
[0142] This is shown in Fig.1 by multiplying the amplitude of the drive deflection xO with kc, cos (the argument (wdrivet) is omitted in Fig.1) and R*.
[0143] R.417457
[0144] - 20 -
[0145] In addition to the Coriolis force Fc, quadrature forces act on the detection mode DM, primarily caused by manufacturing tolerances. These forces are proportional to the displacement of the drive mode AM and, for example, to asymmetries represented by Q, hereinafter referred to as quadrature. In the block diagram of Fig. 1, they are shown as the multiplication of the amplitude of the drive displacement xO (the argument (co drive t) is omitted) and Q.
[0146] Furthermore, Fig. 1 shows vibration forces V that act directly on the detection mode DM and cause its deflection. Vibration forces V with a frequency close to the drive resonance frequency Wdrive are particularly important here, since other frequencies are suppressed by the output filter (not shown).
[0147] The deflection of the detection mode DM results from the acting quantities R*, V, and Q, as well as the transfer function Hsense. For the sake of clarity, it is assumed that the same transfer function can be used for R*, V, and Q. If this does not describe the system with sufficient accuracy, the steps below must be performed for different transfer functions.
[0148] The deflection of the detection mode DM is converted into a current or voltage, for example, using a signal converter. Signal converters are not shown in Fig. 1. After possible amplification and possibly also digitization (not shown), the signal is converted with cos(ω). drive t) and sin(ω drivet) multiplied. After adjusting the sensitivity by the gains GR and GQ, the demodulated rotation rate signal RD or the demodulated quadrature signal QD is obtained. In a dual-resonant operation, this corresponds to sin(ω). drive t) multiplied signal to the rotation rate signal RD and that with cos(ω drive t) multiplied signal to the quadrature signal QD.
[0149] The following section considers the case with separated resonant frequencies.
[0150] R.417457
[0151] - 21 -
[0152] The double-resonant operation can essentially be calculated with little effort by interchanging the meaning of the derived transfer function.
[0153] Low-pass filtering (LP1 and LP2) preferably removes frequencies at twice the drive frequency that arise during multiplication, yielding signals for the rotation rate R* and the quadrature Q. These signals contain errors because the transfer function Hsense leads to a frequency-dependent amplitude and phase, a rotation rate R* partially cross-references into the demodulated signal QD, and a quadrature Q partially cross-references into the demodulated signal RD.
[0154] With the reconstruction component REK according to the invention, these errors can be largely suppressed in one possible embodiment by reconstructing the rotation rate R* and the quadrature Q from the demodulated signals RD and QD. By using the reconstruction component REK according to the invention, extensive compensation of crosstalk can also be achieved without complete reconstruction.
[0155] The derivation of the reconstruction according to the invention, based on the block diagram of the open loop Coriolis vibration gyroscope (OLCVG) according to Fig. 1, consists of the following derivation steps or points:
[0156] • Conversion into a block diagram for detection consisting of transfer functions;
[0157] • Calculation of the transfer functions of the detection;
[0158] • Calculation of transfer functions for complete reconstruction
[0159] For points 2 and 3, the transfer function for an arbitrary transfer function Hsense of the detection mode DM is calculated in the first derivation step.
[0160] R.417457
[0161] - 22 -
[0162] In a second derivation step, the transfer function of a damped harmonic oscillator is used for the detection mode DM.
[0163] In the first derivation step for the derivation of the reconstruction, the block diagram of Fig.1 is transformed into a block diagram consisting of transfer functions that explicitly include crosstalk.
[0164] In the block diagram of Fig. 2, the drive mode AM is not explicitly shown, but is implicitly included in the transfer functions. The bold dashed outline of the block diagram in Fig. 2 corresponds to the bold dashed outline of the block diagram in Fig. 1, each labeled with sensor core SEN-K ("Sense Core").
[0165] To preserve the symmetry of the equations that will be apparent later, the rotation rate R* is divided by the rotation rate R scaled with codrive.
[0166]
[0167] replaced. Here, ωdrive is the actual drive frequency, which is temperature-dependent, among other things, and is normalized with ωdrive0, e.g., the value of ωdrive determined during final adjustment.
[0168] In practice, the input variables R*, Q, and V are scaled to units of a rotation rate, e.g., degrees per second (dps). If R* were multiplied by kc (and not by kc / kcO), there would be an exact equivalence between Fig. 1 and Fig. 2. However, the scaling mentioned above would no longer be valid.
[0169] The transfer function HR2RD (2 corresponds to “to”) shown in Fig.2 describes the transfer H of the rotation rate input signal R to the demodulated rotation rate signal RD.
[0170] R.417457
[0171] - 23 -
[0172] The transfer function HR2QD shown in Fig.2 describes the transfer H of the rotation rate input signal R to the demodulated quadrature signal QD, i.e. the crosstalk of the rotation rate input signal R.
[0173] The transfer function HQ2QD shown in Fig.2 describes the transfer of a quadrature input signal Q to the demodulated quadrature signal QD.
[0174] The transfer function HQ2RD shown in Fig.2 describes the transfer of a quadrature input signal Q to the demodulated rotation rate signal RD, i.e. the crosstalk of the quadrature input signal Q.
[0175] The transfer function HV2RD shown in Fig. 2 describes the transfer of a vibration force V with a frequency close to the drive frequency codrive to the demodulated rotation rate signal RD. It consists of the transfer function Hsense and subsequent multiplication by cos(o)drivet) (dash-dotted outline).
[0176] The transfer function HV2QD shown in Fig. 2 describes the transfer of a vibration force V with a frequency close to the drive frequency codrive to the demodulated quadrature signal QD. It consists of the transfer function Hsense and subsequent multiplication by sin(o)drivet (outlined with a semicolon).
[0177] Before the calculation of the transfer functions H is presented, a conventional compensation circuit is described with reference to Fig. 3. To compensate for static crosstalk, the quadrature is multiplied by the phase error pO and subtracted from the rotation rate. The quadrature can be compensated analogously. Alternatively, the demodulation phase can also be adjusted (not shown).
[0178] Since the quadrature usually changes very slowly, this compensation is in many cases initially sufficient for an improvement of R.417457.
[0179] - 24 -
[0180] Zero-rate output. However, the conventional static compensation shown in Fig. 3 is insufficient if test signals TS are to be used for continuous monitoring of the sensor function or performance parameters, or for continuous tracking of performance parameters such as phase or sensitivity. The phase and amplitude response of the rotation rate input signal R to the demodulated rotation rate RD are not compensated in the conventional circuit shown in Fig. 3.
[0181] General transfer function of the DM detection mode:
[0182] For the rotation rate R, multiplication with cos(wdrivet) yields the modulated, Laplace-transformed signal Rm in the s-domain:
[0183]
[0184] Accordingly, for the quadrature Q, after multiplication with sin(a)drivet, one obtains the modulated signal Qm:
[0185]
[0186] First, the transfer functions H are defined as generally valid for a
[0187] Hsense calculates any transfer function before using a specific function for it.
[0188] For the resulting displacements Ur(s) and Uq(s) of the detection mode DM, one obtains
[0189]
[0190] R.417457
[0191]
[0192] It is assumed that vibration forces V with frequencies close to the drive frequency directly and undesirably excite the detection mode DM.
[0193] The resulting displacement is given by:
[0194]
[0195] Multiplying the signals Ur(s), Uq(s), and Uv(s) by cos(a)drivet yields the signals RrD, QrD, and VrD demodulated with "rate phase". Multiplying the signals by sin(a)drivet yields the signals RqD, QqD, and VqD demodulated with "quadrature phase". For RrD, we first obtain:
[0196]
[0197] The signals are additionally multiplied by "2" to account for the factor "1 / 2" by which the useful signals are scaled through multiplication. This factor can also be considered in the gains GR and GQ. Taking into account that the signals are removed by a low-pass filter LP at twice the drive frequency, the transfer function HR2RD is obtained as follows:
[0198]
[0199] R.417457
[0200] - 26 -
[0201] Accordingly, the further transfer functions H are obtained, as shown in Fig.
[0202] 2 are shown schematically:
[0203]
[0204] The following example assumes (without loss of generality) that the transfer function of the detection mode DM is the transfer function Hsense of a damped harmonic oscillator.
[0205] can be used:
[0206]
[0207] From equations (11) to (16) we obtain
[0208] R.417457
[0209] - 27 -
[0210]
[0211] V2QD — i ^V2RD (23)
[0212] Expanding and simplifying the expressions often yields them in the form shown below.
[0213] required form:
[0214] H(s) = N(s) / D(s), where N(s) and D(s) are polynomials of s.
[0215] From equations (18) and (20) we thus obtain the two
[0216] Transfer functions HR2RD and HR2QD:
[0217]
[0218] R.417457
[0219] Figures 4A and 4B show the amplitude response AMP (Figure 4A) and the phase response PH.
[0220] (Fig.4B) of the two transfer functions HR2RD (dashed) and HR2QD
[0221] (not shown with dashed lines), each multiplied by the gain G=GR=GQ
[0222] Figures 4A and 4B show that, for example, at f=100 Hz, approximately...
[0223] 3.5% of the rotation rate is coupled into the quadrature. The phase between the rotation rate
[0224] and the signal coupled into the quadrature is approximately 65°.
[0225] The gain G is selected or set so that when
[0226] At a static rotation rate, the overall gain is "1", which is related to the following:
[0227] Gain G is achieved:
[0228]
[0229] The approximation for the gain G is valid for sufficiently large differences in the resonance frequencies of the detection and drive modes. It is important to note that...
[0230] Note that the gain G is implemented in the electronics and in the
[0231] In practice, it is also necessary to consider the amplification of the electronics or, for example, the signal converters. The values wdriveT, wsenseT, and QsenseT correspond to these.
[0232] not the mechanical sensor element, but the gain G is used in the
[0233] Once the sensor is calibrated, what is indicated by the letter "T" (for
[0234] Trimming) is indicated in the variable names.
[0235] The following describes the calculation of the transfer functions H for the
[0236] A complete reconstruction is described.
[0237] For the sake of clarity, it is assumed in the following that the amplifications G = GR = GQ can be chosen, respectively.
[0238] are adjustable. If this does not describe the system with sufficient accuracy,
[0239] The following considerations require different approaches in each case.
[0240] Reinforcements are used.
[0241] R.417457
[0242] - 29 -
[0243] From Fig. 2, one can directly read the following for the demodulated rotation rate signal RD and the demodulated quadrature signal QD:
[0244]
[0245] Multiplying both equations by the values given after each equation and then calculating the difference between the two equations, after rearranging, we obtain the following for the first value for the rotation rate R:
[0246]
[0247] and for the second value for the squaring Q:
[0248]
[0249] and
[0250] R.417457
[0251] - 30 -
[0252]
[0253] and
[0254] Q" = HR D2 QR D + H QD2 QQ D + H ¥2 QV = Q' + H ¥2Q V (35)
[0255] For ideal coefficients of the transfer functions used here, the following applies:
[0256] R' ' = R and Q' ' = Q (36)
[0257] Initially, a complete reconstruction cannot yet be performed using equations (34) and (35), since the vibration force V is generally unknown. Setting V=0 yields the block diagram of the reconstruction according to the invention shown in Fig. 5.
[0258] The actual reconstruction component REK is outlined with a dashed line in Fig. 5. In addition to the four transfer functions H of the filters contained in the reconstruction component REK, four low-pass filters LP1 to LP4 are also included in Fig. 5. These are necessary because unstable transfer functions are calculated for the exemplary transfer function Hsense of a damped harmonic oscillator.
[0259] R.417457
[0260] - 31 -
[0261] For the circuit arrangement shown in Fig. 5, the attenuation and phase shift generated by the low-pass filters LP are identical in each of the two branches leading to the measured rotation rate R' and the measured quadrature Q', respectively, if the two low-pass filters LP1 and LP3, as well as the two low-pass filters LP2 and LP4, are identical. This can be advantageous for achieving the best possible compensation. However, if the demodulated rotation rate signal RD and the demodulated quadrature signal QD already have different attenuation and phase shifts, low-pass filters LP3 and LP4, or additional filters, can be used to compensate for these differences.
[0262] In the embodiment of the dynamic reconstruction component REK according to the invention shown in Fig. 5 for an open loop Coriolis vibration gyroscope, the reconstruction component REK comprises compensation filters for compensating crosstalk from input signals to demodulated signals and thus to measured signals.
[0263] The reconstruction component REK preferably has a first compensation filter KF1 and a second compensation filter KF2, which are designed such that a sensor transfer function with respect to a rotation rate input signal R to a rotation rate output signal R' is flat and without phase shift.
[0264] The reconstruction component REK preferably has a third compensation filter KF3 and a fourth compensation filter KF4, which are designed such that a sensor transfer function with respect to a quadrature input signal Q to a quadrature output signal Q' is flat and without phase shift.
[0265] In the embodiment shown in Fig. 5, the compensation filters KF are each connected in series with a low-pass filter LP.
[0266] R.417457
[0267] - 32 -
[0268] The first compensation filter KF1 with a transfer function (HQD2R) with the low-pass filter LP3 connected in series to it and the second compensation filter KF2 with the low-pass filter LP1 connected in series are designed such that a sensor transfer function with respect to a rotation rate input signal R to a measured or reconstructed rotation rate output signal R' is flat and without phase shift.
[0269] The third compensation filter KF3 with a transfer function (HQD2Q) with the low-pass filter LP2 connected in series to it and the fourth compensation filter KF4 with the low-pass filter LP4 connected in series are designed in such a way that a sensor transfer function with respect to a quadrature input signal Q to a measured or reconstructed quadrature output signal Q' is flat and without phase shift.
[0270] In one possible embodiment of the dynamic reconstruction component REK according to the invention for an open loop Coriolis vibration gyroscope OL-CVG as shown schematically in Fig. 5, the first compensation filter KF1 with the low-pass filter LP3 connected in series to it and the second compensation filter KF2 with the low-pass filter LP1 connected in series to it provide filtered output signals, which are fed to a first adder ADD1 of the reconstruction component REK, which adds the two filtered output signals to calculate a reconstructed rotation rate signal R'.
[0271] In one possible embodiment of the dynamic reconstruction component according to the invention for an open-loop Coriolis vibration gyroscope OL-CVG, the third compensation filter KF3 and the low-pass filter LP2 connected in series with it, and the fourth compensation filter KF4 and the low-pass filter LP4 connected in series with it, provide filtered output signals that are fed to a second adder ADD2 of the reconstruction component REK, which R.417457
[0272] - 33 -
[0273] Both filtered output signals are used to calculate a reconstructed
[0274] Quadrature Q' added, as shown in Fig. 5.
[0275] By multiplying the reconstructed quantity R' with ωdrive0 / ωdrive_E, the
[0276] Inverse of the normalized drive frequency, the scaling of the
[0277] The input rotation rate is taken into account with the normalized drive frequency according to equation (4). Thus, the output-side quantity R*' shown in Fig. 5 corresponds to the reconstructed input rotation rate R*. "_E" again indicates
[0278] that ωdrive_E does not represent the actual drive frequency, but corresponds to the value determined for it by the electronics.
[0279] In equations (34) and (35), the signals HV2RV and HV2QV correspond respectively.
[0280] the negative of the vibration V transmitted by the mechanics and the reconstruction filters REK to the measured rotation rate R' or the measured quadrature Q'. After filtering the signal Q' with a
[0281] High-pass filter HP with transfer function H HP (as shown in Fig. 6)
[0282] is) with a very low cutoff frequency (e.g., "< 1 Hz"), which is the sensor's intrinsic,
[0283] Suppresses quadrature that changes very slowly, this corresponds to
[0284] filtered signal approximately only the measured quadratic Q'
[0285] transmitted vibration -HV2QV. From equation (34) we obtain the quantity
[0286] R", as shown in Fig. 6.
[0287]
[0288] This allows the vibration in the rotation rate channel to be compensated. With the
[0289] Equations (32), (33), (31) and (23) give the transfer function
[0290] HVQ2VR, as shown in Fig. 6
[0291]
[0292] R.417457
[0293] - 34 -
[0294] Accordingly, starting from equation (35) one obtains the quantity Q", as shown in Fig.6.
[0295]
[0296] with the transfer function HVR2VQ also shown in Fig. 6:
[0297] 1
[0298] H VR2VQ = 1 / H VQ2VR = i
[0299] "VQ2VR
[0300] The compensation of vibration input signals V using an embodiment of the reconstruction component REK according to the invention is shown in Fig. 6.
[0301] Equations (38) and (40) correspond—with opposite signs—to the transfer function of an ideal Hilbert transform filter for positive frequencies. The Hilbert transform shifts the phase of a signal by 90 degrees without changing its amplitude. In practice, an ideal Hilbert transform filter cannot be directly implemented because it is non-causal and has an infinite impulse response. Instead, approximations such as FIR or IIR Hilbert transform filters are used.
[0302] The compensation of the vibration signal V in the measured quadrature Q' according to equation (39) is usually faulty in practice without additional measures, since the signal R' used for compensation may contain high input rotation rates in addition to a possible vibration component.
[0303] By substituting equations (18) to (21) into the defining equations (31), one obtains for a damped harmonic oscillator:
[0304] R.417457
[0305]
[0306] RD2Q ~ QD2R (43) QD2Q ~ RD2R (44)
[0307] In equations (41) and (42), the sensor quantities could again be denoted by “_E”. This is omitted here, as these are explicitly transfer functions H that are implemented in the electronics.
[0308] In equations (41) and (42), the coefficients are substituted to represent them as a function of directly measurable quantities.
[0309] If we define coefficients c0,c1:
[0310]
[0311] and
[0312]
[0313] For equations (41), (42) we obtain:
[0314] R.417457
[0315]
[0316] It is assumed that the drive frequency codrive can be accurately determined, e.g., by a temperature model TMOD or from the phase controller output of the PLL. As described below, the two coefficients cO and c1 can be determined from the correlation of the measured rotation rate R' and the measured quadrature Q', each with the test signal TS.
[0317] The sensor sizes can then be calculated as follows:
[0318]
[0319] The phase at zero frequency for a damped harmonic oscillator is given by:
[0320]
[0321] Comparing equation (46) with equation (26) reveals that c1 for ideal coefficients of the approximation for the gain G corresponds to the reciprocal of the mechanical gain in codrive.
[0322] The amplitude response AMP and the phase response PH of HRD2R, HQD2R, G*HR2RD and G*HR2QD are shown in Fig. 7A, 7B. The R.417457 is noteworthy.
[0323] - 37 -
[0324] HRD2R and HQD2R exhibit a smooth response. Unlike HR2RD and HR2QD, they do not show a resonance point at the difference between the resonance frequencies of the drive and detection modes (3000 Hz in this example).
[0325] To assess the accuracy of the compensation, the errors caused by differences (mismatch) between the parameters corresponding to the sensor element and those used for reconstruction are calculated. The latter can be stored in an electronic circuit, such as an ASIC. For example, a temperature model (TMOD) can be stored for the parameters. The parameters can also be measured, for example, using test signals (TS). This measurement can be used for monitoring or for recalculating the parameters. The appropriate method can be determined, for example, from the subsequently calculated errors or from errors determined through measurements or simulations.
[0326] To calculate the errors, equations (27) and (28) are substituted into equation (34) with V=0. This yields
[0327]
[0328] Setting Q=0 in equation (52) yields a measure of the sensitivity error of the rotation rate, i.e., a measure of the remaining crosstalk of the rotation rate input signal R on the reconstructed rotation rate signal R'.
[0329]
[0330] Setting R=0 in equation (52) yields a measure of the remaining crosstalk of the quadrature input signal Q on the reconstructed rotation rate signal R':
[0331] R.417457
[0332]
[0333] Similarly, one obtains from
[0334] Q' ~ QD2Q \ Q2QDQ + R2QD J + RD2Q \ Q2RDQ + ^R2RD^J ©5)
[0335] a measure of the sensitivity error of quadrature Q:
[0336]
[0337] and a measure of the remaining crosstalk from
[0338] Rotation rate input signal R to the reconstructed quadrature signal Q':
[0339]
[0340] The formulas for errors due to a mismatch are used to calculate the correlation of the test signal TS with the measured or reconstructed rotation rate signal R' or the measured or reconstructed quadrature signal Q'.
[0341] It should be noted that various selective reconstructions can easily be derived from the complete reconstruction.
[0342] For example, dynamic crosstalk compensation may not be absolutely necessary, which allows the omission of the two filters KF1 and KF4 shown in Fig. 5 with the transmission functions HQD2R and HRD2Q.
[0343] R.417457
[0344] This leads to modified filters with transfer functions HRD2Rs and HQD2Qs.
[0345] If, however, the improvement of the transfer function from rotation rate input to rotation rate output is not required, the two filters KF2 and KF3 shown in Fig. 5 with the transfer functions HRD2R and HQD2Q can be omitted. This then results in other filters with the transfer functions HQD2R selectively (HQD2Rs) and HRD2Q selectively (HRD2Qs), as shown in Fig. 8. For this selective reconstruction, a selective reconstruction component REK' is obtained in the block diagram according to Fig. 8. If vibration is not taken into account, the following expressions for the demodulated rotation rate signal RD and the demodulated quadrature signal QD are obtained from equations (27) and (28):
[0346]
[0347] For the measured rotation rate signal R', the following results from Fig. 8:
[0348] ß' = RD + HQ D 2RSQD = G (HR2RDR + HQ2RDQ + HQD2RSHR2QDR + HQD2RS HQ2QDQ
[0349] With
[0350]
[0351] This results in the following for the measured rotation rate signal R':
[0352] R.417457
[0353] - 40 -
[0354]
[0355] The "_E" indicates that the corresponding transfer functions are not those of the sensor element, but are modeled in the electronics. With ideal filter coefficients of HQD2Rs, the crosstalk of the quadrature input signal Q onto the measured angular rate signal R' is completely compensated. The transfer of the input angular rate signal R to the measured angular rate signal R' is distorted by the second, smaller term, which is usually insignificant in practice. It corresponds to the angular rate input signal R, which first enters the quadrature channel via Hsense and is then transferred back to the measured angular rate R' via the filter.
[0356] The derivation of the transfer function HRD2Qs for the filter KF6 shown below in Fig. 8 for compensating the crosstalk of rotation rate input signals R into the measured quadrature signal Q' is analogous to the transfer function HQD2Rs for the compensation filter KF5 shown above in Fig. 8 for compensating the crosstalk from the quadrature input signal Q to the measured rotation rate R'.
[0357] For the measured quadrature signal Q', the following is obtained according to Fig. 8:
[0358]
[0359] The measured quadrature signal Q' is thus obtained as follows:
[0360] R.417457
[0361]
[0362] With ideal filter coefficients of HRD2Qs, the crosstalk of the rotation rate input signal R onto the measured quadrature Q' is completely compensated. The second, smaller term slightly distorts the transfer of the quadrature Q to the measured quadrature signal Q', which is practically insignificant. It corresponds to the quadrature input signal Q, which is first coupled into the rotation rate channel by Hsense and then transferred back into the quadrature via the filter.
[0363]
[0364] A filter KF5 with the transfer function HQD2Rs according to equation (66) is shown in Fig. 8 in the selective reconstruction circuit REK ' before the low-pass filter LP3.
[0365] For frequencies with a sufficiently large difference from the wsense - codrive difference frequency, the following approximation can be used:
[0366]
[0367] A filter KF6 for compensating the crosstalk from the quadrature input signal Q to the measured or reconstructed rotation rate signal R' with the transfer function HQD2Rs according to equation (67)R.417457
[0368] - 42 -
[0369] is provided in the selective reconstruction circuit REK' according to Fig.8 before the low-pass filter LP4.
[0370] Similarly, the measured quadrature signal Q' is obtained according to Fig. 8.
[0371]
[0372] the approximate compensation transfer function HRD2Qs:
[0373]
[0374] The amplitude response AMP and the phase response PH of the transfer function HQD2Rs for the filter for compensating the crosstalk from the quadrature input signal Q to the measured or reconstructed rotation rate signal R', the approximation of HQD2Rs, G*HR2RD and G*HR2QD are shown in Fig. 9A, 9B.
[0375] Similar to the case of complete reconstruction, it is helpful to substitute coefficients in the transfer functions H in order to represent them as a function of directly measurable quantities.
[0376] If one defines
[0377]
[0378] (dO corresponds to the phase at zero frequency of a damped harmonic oscillator) and
[0379] R.417457
[0380]
[0381] ^sense ^drive
[0382] one receives
[0383] H_QD2Rs = d1s + d0(72)
[0384] As described below, dO and d1 can be determined from the correlation of the measured rotation rate R' and the measured quadrature Q' with the test signal TS.
[0385] To calculate the errors, one first obtains from equations (61) and (62)
[0386]
[0387] Setting Q=0 in equation (73) yields a measure of the sensitivity error of the rotation rate R:
[0388]
[0389] Setting R=0 in equation (73) yields a measure of the remaining crosstalk from quadrature input signal Q to the measured or reconstructed rotation rate signal R':
[0390]
[0391] Similarly, one obtains from
[0392] R.417457
[0393] - 44 -
[0394]
[0395] a measure of the sensitivity error of quadrature Q:
[0396]
[0397] and a measure of the remaining crosstalk from rotation rate input signal R to the reconstituted quadrature signal Q':
[0398] 0'
[0399] — = H R2 QDG + H RD2 Q S H R2RD G (78)
[0400] K
[0401] The formulas for errors due to a mismatch are used to calculate the correlation of the test signal TS with the measured and reconstructed signals R' and Q', respectively.
[0402] In one possible embodiment, the invention provides a method for sensor monitoring as well as for monitoring and tracking the sensor parameters of an open-loop Coriolis vibrating gyroscope (OLCVG). The use of the dynamic reconstruction component REK according to the invention for monitoring the sensor function, or for monitoring or tracking sensor parameters, is illustrated in the block diagram of Fig. 10.
[0403] According to a further aspect, the invention provides a test component TK with a test signal generator TSG for generating a test signal TS, as shown in Fig. 10. In the illustrated embodiment, the generated test signal TS is applied to a quadrature signal input of an open-loop Coriolis vibration gyroscope (OL-CVG). This gyroscope has a dynamic reconstruction component REK for calculating a reconstructed rotation rate signal R' and a reconstructed quadrature signal Q'.
[0404] The reconstructed rotation rate signal R' and the reconstructed quadrature signal Q' are correlated by a KSCH of the test component TK with signal components of the generated test signal TS to calculate
[0405] R.417457
[0406] - 45 -
[0407] Correlation values X and Y are correlated and evaluated by an evaluation unit of the test component TK (not explicitly shown in Fig. 10) for sensor monitoring and / or for adjusting filter coefficients of compensation filters of the reconstruction component REK. In one possible implementation, the reconstruction component REK, the correlation circuit KSCH, and the evaluation unit can be at least partially integrated into a single component, for example, in an ASCI of an inertial measurement unit (IMU).
[0408] A test signal TS is generated using, for example, a digital test signal generator (TSG). In the exemplary representation shown in Fig. 10, the test signal TS consists of two narrowband signal components at two different center frequencies. Only one frequency range or more than two can also be used. Tones (i.e., signals with fixed frequencies), broadband signals, or non-periodic patterns can also be used.
[0409] The test signal TS, if digitally generated, can be converted into a voltage via a digital-to-analog converter (DAC) and conditioned by further circuit components (not shown). In the embodiment of the test component TK shown in Fig. 10, the generated test signal TS is applied to quadrature electrodes, thus generating a test force on the detection mode DM.
[0410] The test signal TS generated by the test signal generator TSG is also fed to the correlators designated CORR of a correlation circuit KSCH of the test component TK via the following transfer functions:
[0411] • a first transfer function HDAC, which maps the possible conversion and conditioning described above;
[0412] • Low-pass filters LP5 or LP6, which advantageously correspond to filters LP1 or LP2, to produce the same attenuation and phase shift that the measured rotation rate signal R' or the measured quadrature signal Q' experiences.
[0413] R.417457
[0414] - 46 -
[0415] • Bandpass filters (BP1, BP2). The bandpass filters BP1, BP2 are advantageously dimensioned so that they have the same center frequency as the respective test signal components of the test signal TS generated by the test signal generator TSG.
[0416] The reconstructed rotation rate signal R' or the reconstructed quadrature signal Q' are preferably fed to the correlators or correlation calculation units CORR of the correlation circuit KSCH via the same bandpass filters BP1 and BP2 respectively.
[0417] If the filter coefficients of the reconstruction component REK are ideally set, the response to the test force appears only in the reconstructed quadrature signal Q' and not in the reconstructed rotation rate signal R'. The correlations Y1 and Y2 (e.g., cross-correlation or, when using tones, multiplication correlation) of the test signal TS with the reconstructed rotation rate signal R' are zero when the input rotation rates are zero in the frequency range of the bandpass filters BP1 and BP2, respectively. The correlations X1 and X2 of the test signal TS with the reconstructed quadrature signal Q' are then equal to one when the signals are appropriately scaled.
[0418] In one possible embodiment of the correlation circuit KSCH, it may be advantageous to also calculate the auto-correlation AR1 and AR2 of the measured rotation rate signals R' and the auto-correlation AQ1 and AQ2 of the measured quadrature signals Q', as shown schematically in Fig. 10.
[0419] For the filter coefficients of the filters shown in Fig. 10, a model is preferably used that takes into account their temperature dependence and stress dependence, provided that temperature and stress sensors are available or the temperature T or the stress is obtained from other sensor parameters. R.417457
[0420] - 47 -
[0421] This can be determined, for example, from the amplitude or frequency of the drive voltage.
[0422] Deviations between values stored in the ASIC and the actual sensor values can be measured as follows. Figures 4 and 7 show that the "direct" transfer functions (HR2RD and HQ2QD) as well as the transfer functions (HRD2R and HQD2Q) are used to compensate for the amplitude response up to very high frequencies (in this example, up to approximately...).
[0423] 1000 Hz) are approximately constant, and thus determined by the constant term. Using one (or more) test signals TS in this range, the sensitivity error can be obtained from equation (56) together with equations (47) and (44).
[0424]
[0425] the ratio of the mechanical gain (at the drive frequency) of the sensor element to the gain stored in the ASIC.
[0426] With the following advantageous normalization, the correlation X1 (or X2) for the specified frequency range of the test signal TS corresponds to the value Q' / Q. Setting Q = TS, which is permissible due to the bandpass filtering, yields a multiplication correlation and normalization of 1 / TS. 2 for the amplitudes, again using equations (56), (47) and (44)
[0427] Q' = c1H Q2QD (0) TS (80)
[0428]
[0429] The correlation X1 (or X2) is a measure of the deviation of the sensor element's gain from the stored value and can be used for monitoring or recalibration. For recalibration, it is calculated using R.417457.
[0430]
[0431] the new value d from d and the correlation X1 (or X2):
[0432]
[0433] Figures 4 and 7 further show that the transfer functions HR2QD and
[0434] HQ2RD, which describes the crosstalk, as well as the transfer functions HRD2Q and HQD2R for compensating the crosstalk, for small
[0435] Frequencies (in the example up to about 20 to 30 Hz) are approximately constant, i.e.
[0436] determined by the constant term.
[0437] Using one (or more) test signals TS in this range, the crosstalk error is obtained from equations (54), (48), (51), (46) and (26).
[0438]
[0439] the difference in phase caused by the sensor element and that in the
[0440] The circuit stores a value to compensate for this phase. Normalization is achieved using 1 / TS. 2 Does the correlation Y1 (or Y2) correspond to this value?
[0441]
[0442] The correlation Y1 (or Y2) is a measure of the deviation of the phase of the sensor element (HQ2RD(0)G) from the stored value cO / G for its
[0443] Compensation and can be used for monitoring or recalibration. For recalibration, one calculates with
[0444]
[0445] R.417457
[0446] - 49 -
[0447] new values from cO and the correlation Y1 (or Y2):
[0448]
[0449] The possible embodiments of methods for monitoring, measuring or tracking phase and sensitivity as well as for tracking the filter coefficients as sensor parameters described below also apply analogously to the methods with vibration compensation and methods for selective reconstruction without vibration compensation described in the following sections.
[0450] The correlation X1 (or X2) represents a measure of the sensor's sensitivity deviation, and the correlation Y1 (or Y2) represents a measure of the phase deviation. These values can be used to
[0451] - to monitor the sensor and issue a corresponding error message if limit values are exceeded
[0452] - to recalculate the reconstruction and thereby readjust phase and sensitivity by:
[0453] • Calculation of cOneu according to equation (87) and dneu according to equation (83)
[0454] • Replacing cO and c1 with cOneu and dneu respectively in the filter according to equations (47), (48), (43) and (44)
[0455] In one possible embodiment, recalibration can be performed continuously, advantageously by implemented control loops that regulate the correlation value X1 to "1" and the correlation value Y1 to "0", or at predetermined intervals, initiated by the user or at sensor start-up.
[0456] For the robustness of monitoring or recalibration, it is advantageous to provide a measure of potential interference with signals X1 (or X2) and Y1 (or Y2), for example, caused by high rotation rates in the corresponding frequency ranges. The previously described R.417457 can serve as a measure, for example.
[0457] - 50 -
[0458] Autocorrelation AR1 (or AR2) can be used, as shown in Fig. 10.
[0459] Different limits are set, which increase in the following list:
[0460] • Limit 1 applies to the use of Y1 (or Y2)
[0461] • Limit 2 applies to the use of X1 (or X2)
[0462] If the limits are exceeded, monitoring or recalibration by Y1 (or Y2) or, if applicable, by X1 (or X2) will be suspended.
[0463] Further robustness can be achieved by using n>1 test signals TS. Sensor monitoring or recalibration can then only be suspended if all n autocorrelation measures ARi exceed their respective limits. This also reduces the probability of false error messages during sensor monitoring. For example, an error message is only triggered if all test signals Xi and Yi, for which the ARi limits are not exceeded, would simultaneously lead to an error message. Accuracy can be improved during recalibration. For example, the mean value can be calculated from the cOneui and c1 neui values for which the ARi limits are not exceeded and used for recalibration.
[0464] The concept is that the output signals from Fig. 10, together with, for example, the temperature T measured by a temperature sensor and / or the PLL output PLL_out of the phase controller PLL and / or the AGC output AGC_out of the amplitude controller AGC, are fed to a processing unit that can perform the operations described below. It is assumed, without loss of generality, that signals 1 and 2 lie within the frequency range FB suitable for measuring the phase error and that signals 3 and 4 lie within the
[0465] R.417457
[0466] - 51 -
[0467] The frequency range FB is suitable for measuring the sensitivity error. Signals 1 and 2 can be the same as signals 3 and 4.
[0468] In one possible embodiment of the method according to the invention, the filter coefficients W of the compensation filters KF are tracked, and thus the phase and the sensitivity are adjusted.
[0469] In a first step S1, codrive is calculated, e.g., from a stored temperature model TMOD together with the temperature T, or from PLL_out and a stored model.
[0470] In a second step S2, the autocorrelation measures AR1...4 are compared with predefined limits. Since the measurements of the correlations X of the test signal TS with the reconstructed quadrature signal Q' are less susceptible to interference from external rotation rates, higher limits will be appropriate here compared to the limits for the measurements of the correlations Y of the test signal TS with the reconstructed rotation rate signal R'.
[0471] In a third step S3, the assigned cOneul, c0neu2 and c1 neu3, c1 neu4 are marked as valid / invalid.
[0472] In a fourth step S4, the valid values of cOneul, c0neu2 from equation (87) and c1neu3, c1neu4 from equation (83) are calculated. If cOneul and c0neu2 are valid, the mean of cOneul and c0neu2 is used for cOneul in step S5.
[0473] If, however, cOneu and cOneu2 are invalid, the last valid value of cOneu is used in step S6. For example, in continuous tracking, a PI controller used for tracking is opened at its controller input and the controller output is held at its last value; it is advisable to apply this procedure for every invalid value.
[0474] If only cOneul or only c0neu2 is valid, the valid value cOneul or c0neu2 is used for cOneu in step S7.
[0475] Similarly, in step S8, dneu is determined using the quantities c1neu3, c1neu4.
[0476] R.417457
[0477] - 52 -
[0478] Finally, in step S9, the filter coefficients of the filters from cOneu, dneu and codrive are calculated according to equations (47), (48), (43) and (44) and the filter coefficients are set accordingly.
[0479] For sensor monitoring of the Open Loop Coriolis Vibration Gyroscope OLCVG, the following procedure can be used in one possible embodiment:
[0480] In a first step S1, the autocorrelation measures AR1...4 are compared with predefined limit values. Since the measurements of correlations X are less susceptible to disturbances from external rotation rates, higher limits will be appropriate here compared to the limits for the measurements of correlations Y.
[0481] In a second step S2, the assigned correlations Y1, Y2 and correlations X3, X4 are marked as valid / invalid.
[0482] If the correlations Y1 and Y2 are invalid, the output S3 "Monitoring phase suspended" is displayed in a third step.
[0483] If only correlation Y1 or only correlation Y2 is valid, a comparison with limit values is performed in a fourth step S4. If the limit value is exceeded, the output "Phase error" is displayed.
[0484] If the correlations Y1 and Y2 are both valid, a comparison with limit values is performed in a fifth step S5. If both signals exceed the limit values, the "phase error" is calculated.
[0485] In a further step S6, an analogous evaluation of correlations X3, X4 is carried out with corresponding outputs, whereby “phase” is to be replaced by “sensitivity”.
[0486] Additionally, it can be advantageous to evaluate the values Qsense, cosense, and cpsense analogously to S7 in a further step. For this, the values cOneul, c0neu2 from equation (87), c1neu3, c1neu4 from equation (83), and wdrive are first calculated, e.g., from a stored temperature model TMOD together with the temperature T and / or from PLL_out. From these, cosense, Qsense, and cpsense are calculated according to equations (49), (50), and (51), respectively, and compared with predefined limits. If R.417457 is exceeded...
[0487] - 53 -
[0488] Depending on the limits, the output may be “cosense error”, “Qsense error” or “cpsense error”.
[0489] In the embodiment shown in Fig. 13, the calculations of further correlations Zi and L of the reconstructed rotation rate signals R' with the reconstructed quadrature signals Q' are also performed. In contrast to the correlations Yi and Y2, these are independent of the gain of the quadrature electrodes, which can be caused, for example, by a change in their electrode spacing.
[0490] Figure 13 shows an evaluation unit (AE) that receives input signals from the correlation calculation units (CORR). Based on the obtained correlation values X, Y, Z and the obtained autocorrelation values AR, AQ, the evaluation unit (AE) calculates or estimates the filter coefficients W for the reconstruction filter (REK) and, if necessary, generates flags and control signals (VM). In the embodiment shown in Figure 13, a unit for error calculation and execution of an adaptation algorithm includes a converter (DAC) with transfer function HDAC, the two low-pass filters LP5 and LP6, the band-pass filters BP, the correlation calculation units (CORR), and the evaluation unit (AE), as schematically depicted in Figure 13.
[0491] If vibration compensation is to be used in the methods described above, it must be taken into account that the test signals TS transferred to the reconstructed quadrature Q' must be removed before the reconstructed quadrature Q' can be used to compensate for vibration signals in the reconstructed rotation rate R'. This can be done by subtracting TS_LP (see Fig. 14) from the reconstructed quadrature signal Q'. Compensating the vibration signals in the reconstructed quadrature signal Q' with the signal R'_Vcomp shown in Fig. 15 yields the reconstructed signal Q". Using Q" instead of Q' when correlating with the test signal TS can improve its robustness. The vibration components in the reconstructed quadrature Q' are only correctly compensated if the reconstructed rotation rate signal R' essentially only R.417457
[0492] - 54 -
[0493] exhibits vibration components and no external rotation rates. Therefore, at high input rotation rates, which can be detected, for example, using autocorrelation values AR1 and / or AR2, or by the autocorrelation of the unfiltered signal R" (not shown), or preferably by comparing the autocorrelations of R'_Vcomp and Q'_Vcomp (not shown; only if both autocorrelations are similar can it be assumed that the signals on R' are essentially caused by vibrations V), the compensation of the vibration components in the reconstructed quadrature signal Q' with the reconstructed rotation rate signal R' is interrupted, as indicated in Fig. 14 by the inputs AR1 and AR2 to HVR2VQ. In Fig. 14, the multiplication of R" by a)driveO / a)drive_E (compare with Fig. 5) and the "Sense Core" SEN-K are not shown again.
[0494] For sensor monitoring or tracking of sensor parameters of the open-loop Coriolis vibration gyroscope OL-CVG, selective reconstruction can also be performed using a selective reconstruction circuit REK', omitting the "direct compensations" HRD2R and HQD2Q. The corresponding block diagram is shown in Fig. 12. Compared to the method with complete reconstruction (see Fig. 10), the direct transfer functions HR2RD_E and HQ2QD_E can be considered in the signal path from the test signal TS to the correlators CORR. "_E" denotes the transfer functions used in the electronics, which correspond to the mechanical transfer functions HR2RD and HQ2QD only with ideal coefficients. If the transfer functions HR2RD_E and HQ2QD_E are required, the complexity is comparable to the method using complete reconstruction, which may then be preferable.
[0495] Fig. 4 shows that the “direct” transfer functions (HR2RD and HQ2QD) are approximately constant up to very high frequencies (in the example up to about 1000 Hz), i.e., they are determined by the constant term. If one (or more) test signals in this range are used, the sensitivity error is obtained from equation (77).
[0496] R.417457
[0497]
[0498] the ratio of the mechanical gain (at the drive frequency) of the sensor element to the gain 1 / G stored in the ASIC.
[0499] In the following advantageous normalization, the correlation X1 (or X2) for the specified frequency range of the test signal corresponds to the value Q' / Q.
[0500] Setting Q = TS, which is permitted due to the bandpass filtering, yields a multiplication correlation and normalization of 1 / TS. 2 for the amplitudes, again using equations (77)
[0501] Q' = GHQ2QD (0) TS (89)
[0502]
[0503] The calculated correlation value X1 (or X2) is a measure of the deviation of the sensor element's gain from the stored value and can be used for monitoring or recalibration. For recalibration, one calculates...
[0504]
[0505] the new value Gnew from G and the correlation X1 (or X2):
[0506]
[0507] For test signals TS with large frequency components (from approx. 500 Hz), the approximation of equation (77) to (88) may be invalid. In equation (92) and in the subsequent derived equations, the value calculated according to equation (77) must then be used for "1".
[0508] R.417457
[0509] - 56 -
[0510] Gneu can be used to adjust the gain. This can be done using the GR and GQ blocks in Fig. 12, or the block not shown for scaling with (drive_E or additional gain).
[0511] The filter coefficient d1 can be calculated from equations (71) and (26):
[0512]
[0513] wdrive can be determined using a stored model, as with other
[0514] The position is explained.
[0515] The signal diagrams according to Fig. 4 and Fig. 9 show that the transfer functions HR2QD and HQ2RD, which control crosstalk
[0516] The transfer functions HRD2Q and HQD2R for crosstalk compensation are described and are approximately constant for small frequencies (in the example up to approximately 20 to 30 Hz), i.e., determined by the constant term. Using one (or more) test signals TS in this range, the crosstalk error is obtained from equations (75), (72), and (70).
[0517]
[0518] The difference between the phase caused by the sensor element and the value stored in the circuit to compensate for this phase. Normalized with 1 / TS 2 Does the correlation Y1 (or Y2) correspond to this value?
[0519] ® ÜQ 2RD (0) G + d0(95)
[0520] Y1 (or Y2) is a measure of the deviation of the sensor element's phase from the stored value for its compensation and can be used for monitoring or recalibration. For recalibration, it is calculated using R.417457.
[0521]
[0522] new values from dO and the correlation Y1 (or Y2):
[0523]
[0524] dOneu can be used to monitor the phase or together with d1 neu to recalculate the transfer functions HQD2Rs according to equation (72) and HRD2Qs according to equation (69), and thus to recalibrate the phase.
[0525] According to one aspect of the invention, an open-loop Coriolis vibration gyroscope (OLCVG) provides a drive control loop comprising a drive oscillator and at least one detection oscillator. The open-loop Coriolis vibration gyroscope OLCVG includes a reconstruction component REK, which has a reconstruction filter matrix MDE' and is designed to calculate reconstructed signals VE' ZU from signals VD demodulated according to gyroscope rate and quadrature, corresponding to the input signals VE applied to the at least one detection oscillator of the open-loop Coriolis vibration gyroscope OLCVG.
[0526] The transformation of the input signals R and Q to the demodulated signals RD and QD according to equations [Error! Reference source not found, until [Error! Reference source not found., and [Error! Reference source not found]] can be represented with a 2x2 filter matrix MED:
[0527]
[0528] This includes
[0529] R.417457
[0530]
[0531] From equation (2) we obtain:
[0532]
[0533] where detMgß is the determinant and adj M ED the adjoints of the filter matrix M ED are. Considering that the filter matrix MED cannot be determined with arbitrary precision, and denoting the determined model of the filter matrix by M, ED ,m is obtained for the reconstruction in matrix notation:
[0534]
[0535] The reconstruction filter matrix MDE is thus given by:
[0536]
[0537] Equation (4) leads to an error message! Reference source could not be found. if one sets G = GR = GQ, as in the case of an error message! Reference source could not be found.
[0538] The selective reconstruction corresponds to the following approximation in matrix representation:
[0539] R.417457
[0540]
[0541] The MED filter matrix and the associated MDE reconstruction filter matrix can be extended to any number of detection transducers.
[0542] For example, the vector VE of input quantities or input signals shown in Fig. 16 for an integer number n of detection transducers of the open Coriolis vibration gyroscope OLCVG can be defined as follows:
[0543]
[0544] where the signals Ri and Qi are the inputs to the i-th detection transducer of the open Coriolis vibration gyroscope OLCVG. It is important to note that in this representation, the inputs must be used in the direction of the respective sensitive axes, and non-orthogonality, for example, must be taken into account beforehand.
[0545] The vector VD of the demodulated signals shown in Fig. 16 is then given by:
[0546]
[0547] R.417457
[0548] - 60 -
[0549] where RDi and QDi are the signals of the i-th detection oscillator demodulated according to rotation rate and quadrature.
[0550] This yields a 2n x 2n filter matrix M ED, which describes the transfer of the input quantities of the vector VE to the demodulated signals of the vector VD:
[0551] V D = MED V E 0°)
[0552] The calculation of the vector VD from the vector VE using the filter matrix M ED is shown schematically in Fig. 16.
[0553] The reconstruction is given by:
[0554] 1
[0555] g / — M DE IV D — M ED m V D — adj M ED m V D « V E (11) aetM ED m
[0556] The filter matrix M ED m can be calculated analytically, as in the previously shown examples, or determined by simulations or measurements.
[0557] The reconstruction matrix M used in Fig. 16 to calculate the reconstructed signals DE > can also be calculated analytically, or determined by simulations or measurements. The "approximately equal to" symbol in equation (77) takes into account that the filter matrix M ED m generally not exactly the sensor properties MED reflects or in the inversion of M ED m Approximations are made, and thus the reconstructed quantities V E > only approximately with the input variables V E agree.
[0558] The reconstruction filter matrix MDE' can account for crosstalk between the detection transducers of the open Coriolis vibration gyroscope OLCVG. For example, if a quadrature motion of a first detection transducer coincides with the input gyroscope rate of a second R.417457
[0559] - 61 -
[0560] If the detection oscillator is used to generate a rotation rate signal in a third detection oscillator, this can be compensated.
[0561] The block diagram of the general reconstruction is shown in Fig. 16.
[0562] It is important to note that the filter coefficients Wo of the reconstruction filter matrix MDE' are advantageously tracked via a temperature model TMOD, and / or, as will be shown later, adaptively adjusted. In addition to an input for the vector VD of the demodulated signals, the reconstruction matrix MDE' advantageously has further inputs for the filter coefficients.
[0563] The stored temperature model TMOD for the filter coefficients Wo (T) shown schematically in Fig. 16 can use as input a measured temperature T and / or the force amplitude of the drive oscillator AGC_out and / or a measure of the frequency of the drive oscillator, e.g. the PLL output PLL_out, as shown schematically in Fig. 16.
[0564] The compensation of vibration signals V for a detection oscillator of the open Coriolis vibration gyroscope OLCVG can be generalized in matrix notation as follows:
[0565]
[0566] The matrix M v is used to compensate for vibration signals V (see equation [Error! Reference source could not be found, and error!]).
[0567] (Reference source could not be found.) Provided:
[0568]
[0569] It should be noted that the compensation of the quadrature is preferably used in conjunction with an error calculation by a
[0570] R.417457
[0571] - 62 -
[0572] The error calculation unit FBE and the provision of a monitor unit MON are used. Otherwise, the following matrix Mv is preferably used:
[0573]
[0574] In this case, the procedure corresponds to the block diagram shown in Error! Reference source could not be found.
[0575] Furthermore, it should be noted that when test signals TS are fed in for the compensation of reconstructed rotation rate signals R', the input signals Q'-£TS and not Q' are preferably used, i.e.
[0576]
[0577] That is, all associated test signals TS are subtracted from the reconstructed quadrature signal Q' before it is used for compensation.
[0578] The extensions to any number of detection oscillators of equation (72) are given by:
[0579] v E " = M V V E , = M V M DE , V E
[0580]
[0581] Equation (75) becomes:
[0582] v E " = M v (y E , - V TS )
[0583]
[0584] The vector VTS is defined as follows:
[0585] R.417457
[0586] - 63 -
[0587]
[0588] where 'Sij' represents the sum of all test signals TS fed to the i-th detection transducer.
[0589] The following describes in more detail the evaluation of the reconstructed signals VE' for sensor monitoring with the input of a test signal TS.
[0590] The generalization of the evaluation of the reconstructed signals VE- when using test signals TS is shown in the embodiment according to Fig.18.
[0591] The error calculation unit FBE shown in Fig. 18 calculates the error signals of the error signal vector VF. These error signals primarily exhibit deviations of the reconstructed quantities from the target values, particularly for the rotation rate signals, which are measures of the respective signal amplitudes. The target values for the reconstructed quadrature signals Q' are the corresponding test signals TS, with zero expected for the reconstructed rotation rate signals R'. Advantageous preconditioning can also be performed within the error calculation unit FBE, for example, bandpass filtering of the signals or additional filtering of the test signals TS of the test signal vector VTS. This preconditioning takes into account the transmission of the test signals TS up to the interface of the detection transducers, or, especially in the case of selective reconstruction, also the transfer functions of the detection transducers.
[0592] The actual error calculation by the error calculation unit FBE shown in Fig. 18 can, for example, include the following in possible embodiments:
[0593] R.417457
[0594] - 64 -
[0595] • a previously presented cross-correlation (of the components of Vp with the corresponding components of the test signal vector VTS). For quadrature components, the expected value of the cross-correlation is "1", for rotation rate components it is "0";
[0596] • the autocorrelation described above as a measure of the amplitude of a quantity;
[0597] • for monofrequency test signals, demodulation (of the components of Vp with the associated components of the test signal vector VTS);
[0598] • a differentiation.
[0599] In the monitor unit MON shown in Fig. 18, the functions described above are performed in possible embodiments:
[0600] • Comparison of the error VF with predefined limits;
[0601] • Comparison of the errors VF for different test signals TS applied to a detection transducer;
[0602] • Use of autocorrelation or measures for the quantities of the quadrature and / or rotation rate signals of Vp belonging to the respective test signals to decide whether the error signals may be disturbed by high input rotation rates or by vibrations;
[0603] • Make a decision based on the points mentioned above,
[0604] whether and for which sizes an error flag (VM) should be set;
[0605] whether, if necessary, the compensation of the quadrature components of VE' should be stopped during vibration compensation by the matrix Mv (see Fig. 19 and Error! Reference source could not be found.) (VM3).
[0606] • When adaptively tracking the filter coefficients W (see Fig. 20 and Error! Reference source not found.), VM2 makes a decision as to whether the currently determined values for the filter coefficients W are reliable and accordingly controls the adaptation algorithm ANP-ALGO executed on the adaptation calculation unit ABE.
[0607] R.417457
[0608] - 65 -
[0609] The following section describes in more detail the evaluation of the reconstructed signals VE' for sensor monitoring and tracking of the filter coefficients W with the input of a test signal TS.
[0610] In order to continuously adjust the filter coefficients W, the one shown in Fig.
[0611] The adaptation calculation unit ABE, shown in Figures 20 and 21, is provided for the execution of an adaptation algorithm ANP-ALGO. The adaptation calculation unit ABE calculates the optimal filter coefficients Wopt while minimizing the deviations between the reconstructed signals VE and the expected values.
[0612] In possible embodiments, the following adaptation algorithms ANP-ALGO can be used, although the method according to the invention is not limited:
[0613] • Least Mean Squares (LMS) algorithm
[0614] • Normalized Least Mean Squares (NLMS) algorithm
[0615] • Recursive Least Squares (RLS) algorithm
[0616] • Kalman filter
[0617] • Affine Projection Algorithm (APA)
[0618] • Frequency-Domain Adaptive Filter (FDAF)
[0619] • Controller, e.g. B. I, PI, PID controllers
[0620] The calculation of the filter coefficients W can be supported by a temperature model TMOD for the filter coefficients WO (T), which allows the optimal filter coefficients Wopt to be determined faster and more reliably when the rotation rate sensor is switched on.
[0621] As schematically shown in Fig. 20, when tracking the filter coefficients W, the reconstruction component REK, which multiplicatively applies the reconstruction matrix MDE- to the vector VD of the demodulated signals RD, QD to calculate the vector VE- of the reconstructed signals R', Q', forms the error calculation unit FBE and the adjustment R.417457
[0622] - 66 -
[0623] The processing unit ABE, which executes the adaptation algorithm ANP-ALGO, forms an adaptive filter, ADAP-FIL, through which the filter coefficients W of the reconstruction matrix MDE' of the reconstruction component REK are continuously updated. In the embodiment shown in Fig. 20, the vector VD is the vector of demodulated signals, VE is the input signal vector, VTS is the reference signal vector, and VF is the error vector. W represents the weighted filter coefficients, and VE- is the output signal vector of the adaptive filter ADAP-FIL.
[0624] In one possible embodiment, the error signals of the error vector VF are output to a monitor unit MON. Preferably, temperature models TMOD for the filter coefficients and control signals VM2, which are generated by the monitor unit MON, are additionally used as input variables for the adaptation calculation unit ABE, thereby improving the robustness of the method according to the invention.
[0625] As shown in Figures 19 and 21, in one possible embodiment, vibrations V are additionally compensated to provide vibration-compensated signals VE-. These vibration-compensated signals VE- preferably form further input variables for the error calculation unit FBE. In a preferred embodiment, the vibration-compensated signals VE- are additionally used in the error calculation by the error calculation unit FBE, as schematically illustrated in Figures 19 and 21.
[0626] According to a further aspect, the invention provides a method for operating an open Coriolis vibration gyroscope or an open-loop Coriolis vibration gyroscope OLCVG with at least one sensitive axis. In the inventive method for operating an open Coriolis vibration gyroscope OLCVG, signals R', Q' are reconstructed from the signals VD demodulated according to gyroscope rate and quadrature by applying a reconstruction filter matrix MDE', which then calculates the signals R', Q'.
[0627] - 67 -
[0628] corresponding to the input signals VE applied to at least one detection transducer of the open Coriolis vibration gyroscope OLCVG.
[0629] The open Coriolis vibration gyroscope OLCVG has a drive control loop, wherein the drive control loop has a drive oscillator. In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has at least one or more detection oscillators j. In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has devices for generating detection signals Uj, which are a measure of the displacement of the respective detection oscillator j.
[0630] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG also includes devices with which the signals VD, demodulated according to gyroscope rate and quadrature, are generated from the detection signals Uj. In one possible embodiment, the open Coriolis vibration gyroscope OLCVG according to the invention includes a reconstruction filter matrix MDE'.
[0631] The inventive method for operating an open Coriolis vibration gyroscope OLCVG is characterized in particular by the fact that, by applying a reconstruction filter matrix MDE', the reconstructed signals VE' are generated from the signals VD demodulated according to gyroscope rate and quadrature, which, with ideal filter coefficients Wopt and with negligible disturbance, e.g. by vibrations or sound waves, largely correspond to the input signals VE.
[0632] In one possible embodiment of the method for operating an open Coriolis vibration gyroscope OLCVG, the reconstruction filter matrix MDE' of the reconstruction component REK is determined by the following steps:
[0633] R.417457
[0634] - 68 -
[0635] In a first step, the transfer functions MED from the input signals VE to the demodulated signals VD are determined experimentally, by simulation or computationally.
[0636] In a second step, the filter structure and the coefficients of the individual filter matrix elements of the reconstruction filter matrix MDE' are determined experimentally, by simulation, numerically or computationally, for example by inverting the transfer function matrix MED.
[0637] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, temperature models TMOD are determined for the filter coefficients W of the reconstruction filter matrix MDE'. These models, together with the measured temperature and / or a temperature-dependent sensor parameter, in particular the output of the AGC controller or the PLL controller, are used to adapt the filter coefficients W of the reconstruction filter matrix MDE' to a changed temperature T. The temperature is the ambient temperature of the open Coriolis vibration gyroscope OLCVG.
[0638] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, at least one test signal vector VTS is fed into the at least one detection transducer via an interface. The test signals TS of the test signal vector VTS can be fed into the at least one detection transducer via an interface which has quadrature compensation electrodes.
[0639] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, errors VF of the reconstructed quantities VE- are calculated with respect to expected values.
[0640] R.417457
[0641] - 69 -
[0642] For error calculation by the error calculation unit FBE of the adaptation filter ADAP-FIL, one of the following calculation methods can be used:
[0643] • Cross-correlation of components of the vector Vp of the reconstructed signals with the corresponding components of the test signal vector VTS. For quadrature components, the expected value of the cross-correlation is "1", for rotation rate components it is "0".
[0644] • Autocorrelation of the signals reconstructed from the components of the subject as a measure of the amplitude of a quantity;
[0645] • Cross-correlation of the rotation rate-demodulated components RDJ of VE' with the quadrature-demodulated components QDJ of Vp. The expected value of the cross-correlations is "0";
[0646] • For monofrequency test signals, a demodulation of the components of the vector Vp of the reconstructed signals with the corresponding component of the test signal vector VTS. ;
[0647] • Differentiation
[0648] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, monofrequency signals, and / or narrowband noise, and / or broadband noise, and / or pseudo-random sequences are used for the at least one test signal TS of the test signal vector VTS.
[0649] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, the signal vectors VE' and VTS are filtered by bandpass filters BP before being used for error calculation. The bandpass filters BP preferably each have a center frequency which is R.417457
[0650] - 70 -
[0651] The center frequency of the respective test signal TS corresponds to this. The bandpass filters BP preferably also have a frequency bandwidth that ensures that the respective test signal TS remains virtually unchanged.
[0652] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, the error signals of the error signal vector VF are used to monitor the sensor.
[0653] Furthermore, a monitor unit MON is preferably used, with which the following, but not necessarily all, functions can be performed:
[0654] • Comparison of the error signals of the error signal vector VF with predefined limits;
[0655] • Comparison of the errors of the error signal vector VF for different test signals applied to a detection transducer;
[0656] • Use of the measures for the quantities of the quadrature and / or rotation rate signals of the vector Vp belonging to the respective test signals to decide whether the error signals may be disturbed by high input rotation rates or by vibrations;
[0657] The monitor unit MON can decide, based on the points above, whether and for which quantities an error flag should be set, with a corresponding output of an error flag or first control signal (VM1). The monitor unit MON can also decide whether and for which detection transducers the adjustment of filter coefficients W should be suspended and output a corresponding second control signal (VM2). The monitor unit MON can also decide whether the vibration compensation of the quadrature signals of the vector VE' should be suspended for which detection transducers.
[0658] - 71 -
[0659] is to be and output a corresponding third control signal (VM3).
[0660] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, the quadrature components Q' of the vector Vp of the reconstructed signals, after subtracting the respective test signals TS, are filtered with a filter that approximates an ideal Hilbert transform filter, for example, an FIR or IIR Hilbert transform filter. The filtered signal disturbances on the gyroscope components R' of the vector Vp of the reconstructed signals, caused by vibrations, are then compensated. The gyroscope components of the vector VE- of the reconstructed signals, thus compensated, can be used for error calculation by the error calculation unit FBE. The monitor unit MON can optionally disable the vibration compensation via the third control signal VM3.
[0661] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has at least one adaptive filter ADAP-FIL. The filter matrix of the adaptive filter ADAP-FIL is formed by the reconstruction filter matrix MDE-, which has inputs VD and outputs VE-.
[0662] In one possible embodiment, the error calculation unit FBE is part of the adaptive filter ADAP-FIL, which has the inputs VE, VTS and, in the case of vibration compensation, additionally the inputs VE-, and which contains as outputs the calculated errors of the error vector VF, as shown in the Fig.
[0663] Figures 19 and 21 are shown. The error vector VF calculated by the error calculation unit FBE is fed to the monitor unit MON.
[0664] The adaptive filter ADAP-FIL preferably includes a fitting calculation unit ABE, which executes a fitting algorithm ANP-ALGO that determines the optimal filter coefficients W of the reconstruction filter matrix R.417457.
[0665] - 72 -
[0666] MDE is calculated while minimizing the error signals of the error vector VF. In one possible embodiment, the adaptation calculation unit ABE of the adaptive filter ADAP-FIL optionally receives a signal that represents a measure of the ambient temperature T of the open Coriolis vibration gyroscope OLCVG.
[0667] In one possible embodiment, the ANP-ALGO adaptation algorithm optionally receives the second control signal VM2 from the MON monitor unit, whereupon the adaptation of the filter coefficients W of the reconstruction filter matrix MDE' is stopped or resumed.
[0668] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, the reconstruction filter matrix MDE' of the reconstruction component REK only compensates the crosstalk of quadrature input signals Q on the measured gyroscope oscillator R' of the same detection oscillator.
[0669] In another alternative embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, the reconstruction filter matrix MDE' of the reconstruction component REK additionally compensates the crosstalk of gyroscope input signals R on the measured quadrature signal Q of the same detection oscillator.
[0670] In another possible embodiment of the inventive method for operating an open Coriolis vibration gyroscope OLCVG, transfer functions with respect to the test signals TS of the test signal vector VTS up to the interface to the at least one detection transducer, for example due to transducers or RC elements, or up to the detection signals Uj, additionally due for example to the transfer function of the respective detection transducer, are taken into account by filtering them by means of filters before they are used for the error calculation R.417457
[0671] - 73 -
[0672] are used, or rather, filtered within the adaptive filter using a filter.
[0673] In one possible embodiment, the invention provides an open Coriolis vibration gyroscope OLCVG with a drive control loop, wherein the drive control loop includes a drive oscillator. The open Coriolis vibration gyroscope OLCVG also includes at least one detection oscillator. In one possible embodiment, the open Coriolis vibration gyroscope OLCVG includes a device for generating a detection signal U, which is a measure of the displacement of the detection oscillator.
[0674] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has devices with which the detection signal U is demodulated with a signal in phase with the speed of the drive oscillator, which in double-resonant operation yields the quadrature demodulated signal QD and in non-double-resonant operation yields the gyroscope demodulated signal RD.
[0675] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has devices with which the detection signal U is demodulated with a signal in phase with the deflection of the drive oscillator, which in double-resonant operation yields the signal RD demodulated according to gyroscope rate and in non-double-resonant operation yields the signal QD demodulated according to quadrature.
[0676] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG includes a reconstruction filter REK. The open Coriolis vibration gyroscope OLCVG has a reconstruction filter REK, or a reconstruction component REK. The reconstruction filter REK generates the reconstructed signals R' and Q' of the vector VE' from the detection signals RD and QD, respectively, which are demodulated according to gyroscope rate and quadrature. The signals generated by the
[0677] R.417457
[0678] - 74 -
[0679] Reconstruction filters REK reconstructed signals R' and Q' with ideal filter coefficients W. opt and in the case of negligible disturbance, especially by vibrations or sound waves, the input rotation rate signal R or the input quadrature signal Q.
[0680] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG has an interface which includes quadrature compensation electrodes, as well as a device for generating test signals TS or a test signal generator TSG.
[0681] In one possible embodiment, the open Coriolis vibration gyroscope OLCVG incorporates an adaptive filter ADAP-FIL. The reconstruction filter REK has a reconstruction filter matrix MoE and is preferably a component of the adaptive filter ADAP-FIL. In another possible embodiment, the adaptive filter ADAP-FIL receives the test signal TS as a further input signal. The desired signals are the test signal TSi for Q' or Q" and zero for R' or R". The adaptation algorithm ANP-ALGO, executed on the adaptation calculation unit ABE of the adaptive filter ADAP-FIL, adaptively calculates the filter coefficients W of the reconstruction filter REK, which can also be supported by a temperature model TMOD.
[0682] The method and device according to the invention can be used, for example, for open-loop gyroscopes in the automotive and consumer electronics sectors. Due to its simple design, low energy consumption, recalibration capability, monitoring capabilities, and high robustness, the open-loop Coriolis vibratory gyroscope according to the invention is used in many applications. Some important application areas of an open-loop Coriolis vibratory gyroscope (OLCVG) according to the invention are listed below.
[0683] Use of the Open Loop Coriolis Vibration Gyroscope (OLCVG) according to the invention in the automotive industry:
[0684] R.417457
[0685] - 75 -
[0686] • Electronic Stability Control (ESC): To detect rotational movements of the vehicle and stabilize it in case of skidding.
[0687] • Navigation systems (INS, GPS support): Supplementing GPS for more accurate positioning, especially in tunnels or parking garages.
[0688] • ADAS (Advanced Driver Assistance Systems): For autonomous driving functions and lane keeping assist.
[0689] Use of the open loop Coriolis vibration gyroscope according to the invention in aerospace:
[0690] • Drones (UAVs): For situation control and position determination.
[0691] • Aircraft navigation: Support for inertial navigation in aircraft.
[0692] • Satellites: For attitude control and stabilization in orbit.
[0693] Use of the Open Loop Coriolis Vibration Gyroscope (OLCVG) according to the invention in robotics & industrial automation:
[0694] • Industrial robots: For precise control and motion tracking.
[0695] • Autonomous vehicles & mobile robots: For path navigation and stabilization
[0696] • Machine alignment: In CNC machines for detecting vibrations and inclinations.
[0697] Use of the Open Loop Coriolis Vibration Gyroscope (OLCVG) according to the invention in medical technology:
[0698] • Endoscopy & surgical instruments: For precise movement control in minimally invasive surgeries.
[0699] • Portable medical devices: For recording movement data in patients 5.
[0700] Use of the Open Loop Coriolis Vibration Gyroscope (OLCVG) according to the invention in consumer electronics and sports:
[0701] • Smartphones & Tablets: For screen rotation and motion tracking in AR / VR applications.
[0702] • Gaming controllers & VR systems: For precise tracking of movements.
[0703] • Sports analysis & wearables: For movement studies and training recording.
[0704] R.417457
[0705] - 76 -
[0706] Although the present invention has been fully described above with reference to preferred embodiments, it is not limited thereto, but can be modified in many ways.
Claims
R.417457 - 77 - Claims:
1. Dynamic Reconstruction Component (REK) for an Open Loop Coriolis Vibration Gyroscope (OLCVG), wherein the dynamic reconstruction component (REK) is designed to reconstruct input signals (VE) of the Open Loop Coriolis Vibration Gyroscope (OLCVG) from demodulated signals (VD) of the Open Loop Coriolis Vibration Gyroscope (OLCVG).
2. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to claim 1, wherein the demodulated signals (VD) of the open loop Coriolis vibration gyroscope (OLCVG) comprise a demodulated rotation rate signal (RD) and a demodulated quadrature signal (QD).
3. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to claim 1 or 2, wherein the input signals (VE) of the open loop Coriolis vibration gyroscope (OLCVG) comprise a rotation rate input signal (R), a quadrature input signal (Q), and a vibration input signal (V) and / or a test signal (TS).
4. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to one of claims 1 to 3, wherein the reconstruction component (REK) is designed to provide reconstructed signals (VE), in particular a reconstructed rotation rate signal (R') and a reconstructed quadrature signal (Q').
5. Dynamic reconstruction component for an open-loop Coriolis vibration gyroscope according to one of claims 1 to 4, wherein a correlation circuit connected to the reconstruction component (REK) R.417457 - 78 - (KSCH) Correlation calculation units (CORR) designed to correlate the reconstructed rotation rate signal (R') and the reconstructed quadrature signal (Q') emitted by the reconstruction component (REK) with signal components of at least one test signal (TS) generated by a test signal generator (TSG) to calculate correlation values (X, Y).
6. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to claim 5, wherein correlation calculation units (CORR) of the correlation circuit (KSCH) are further designed to calculate autocorrelation values (AR) for the reconstructed rotation rate signal (R') output by the reconstruction component (REK) and autocorrelation values (AQ) for the reconstructed quadrature signal (Q') output by the reconstruction component (REK).
7. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to claim 5, wherein an evaluation unit is designed to evaluate the correlation values (X1, X2, Y1, Y2) calculated by the correlation calculation units (CORR) of the correlation circuit (KSCH) and / or the autocorrelation values (AR, AQ) calculated by the correlation calculation units (CORR) of the correlation circuit (KSCH) for recalibration of filters of the dynamic reconstruction component (REK) and / or for sensor monitoring of the open loop Coriolis vibration gyroscope (OLCVG).
8. Dynamic reconstruction component for an open-loop Coriolis vibration gyroscope according to any one of the preceding claims 1 to 7, wherein a vibration compensation circuit connected to the reconstruction component (REK) is designed to compensate for the reconstructed rotation rate signal (R') emitted by the reconstruction component (REK) and the vibration compensation circuit emitted by the reconstruction component (REK). R.417457 - 79 - to evaluate the reconstructed quadrature signal (Q') to compensate for vibration interference signals (V).
9. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to any one of the preceding claims 1 to 8, wherein the reconstruction component (REK) comprises compensation filters (KF1, KF2, KF3, KF4) which are provided to compensate for crosstalk from input signals (VE) to reconstructed signals (VE ).
10. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to any one of the preceding claims 1 to 9, wherein the reconstruction component (REK) comprises a first compensation filter (KF1) and a second compensation filter (KF2) designed such that a sensor transfer function with respect to a rotation rate input signal (R) to a reconstructed rotation rate signal (R') is flat and without phase shift.
11. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to any one of the preceding claims 1 to 10, wherein the reconstruction component (REK) comprises a third compensation filter (KF3) and a fourth compensation filter (KF4) designed such that a sensor transfer function with respect to a quadrature input signal (Q) to a reconstructed quadrature signal (Q') is flat and without phase shift.
12. Dynamic reconstruction component for an open loop Coriolis vibration gyroscope according to claims 9 to 11, wherein a low-pass filter (LP) is connected downstream of each of the compensation filters (KF) of the dynamic reconstruction circuit (REK) in such a way that crosstalk from input signals (R, Q) to reconstructed signals (R', Q') is compensated.
13. Dynamic reconstruction component for an open-loop Coriolis vibration gyroscope according to claim 12, wherein the first compensation filter R.417457 - 80 - (KF1) and the low-pass filter (LP3) connected in series with it, as well as the second compensation filter (KF2) and the low-pass filter (LP1) connected in series with it, provide filtered output signals that are fed to a first adder (ADD1) of the reconstruction component (REK), which adds the two filtered output signals to calculate a reconstructed rotation rate signal (R'), and wherein the third compensation filter (KF3) and the low-pass filter (LP2) connected in series with it, as well as the fourth compensation filter (KF4) and the low-pass filter (LP4) connected in series with it, provide filtered output signals which are fed to a second adder (ADD2) of the reconstruction component (REK), which adds the two filtered output signals to calculate a reconstructed quadrature signal (Q').
14. Open Loop Coriolis Vibrating Gyroscope (OLCVG) for measuring an angular velocity, wherein the Open Loop Coriolis Vibrating Gyroscope (OLCVG) comprises a dynamic reconstruction component (REK) according to any one of claims 1 to 13.
15. Open Loop Coriolis Vibrating Gyroscope (OLCVG) for measuring an angular velocity according to claim 14, wherein demodulated signals (VD) of the Open Loop Coriolis Vibrating Gyroscope are filtered by adjustable filters of the reconstruction component (REK) for the reconstruction of the input signals (VE) of the Open Loop Coriolis Vibrating Gyroscope (OLCVG).
16. Test component (TK) with a test signal generator (TSG) for generating a test signal (TS) which is applied to a signal input of an open loop Coriolis vibration gyroscope (OLCVG) which has a dynamic reconstruction component (REK) for calculating a reconstructed rotation rate signal (R') and a reconstructed quadrature signal (Q'), wherein the reconstructed rotation rate signal (R') and the reconstructed quadrature signal (Q') are correlated by a correlation circuit (KSCH) of the test component (TK) with signal components of the generated test signal (TS) to calculate correlation values which are determined by an R.417457 - 81 - The evaluation unit of the test component (TK) is used to track the filter coefficients (W) of the compensation filters (KF) of the reconstruction component (REK).
