JP2006234710A - Device for detecting gyroscope signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To eliminate signals generated in a detecting element by electric charges (unnecessary electric charge components) regardless of Coriolis force by leakage signals that is generated the leakage of the vibration energy of a tuning-fork vibrator, without synchronously detecting for generating unnecessary noise. <P>SOLUTION: A device for detecting a gyroscope signal includes vibration gyros 11a, 11b that a pair of activating arms respectively having a pair of vibrators on an actuation side are oppositely arranged sandwiching a detecting arm with a pair of vibrators on a detection side that are arranged sandwiching an objective axis, a differential amplifier 21 for differentially amplifying analog signals outputted from a pair of the vibrators on the detection side, an A/D converter 32 for converting the differential amplification signals outputted from the differential amplifier 31, and a noise eliminating section 33 for processing digital detection signals outputted from the A/D converter with a Kalman filter to eliminate noise. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a gyro signal detection device suitable for a vibration gyro using a tuning fork type crystal resonator made of crystal as an angular velocity sensor.

As this type of gyro signal detection device, for example, in a vibration gyro provided with a synchronous detection unit that synchronously detects a sense signal output from a sensing element using a reference signal synchronized with a monitor signal, the detection signal after synchronous detection is used. In order to remove high frequency noise components, instead of using a CR filter having a large time constant, an analog moving average filter that performs a moving average process for each detection signal in a period synchronized with the reference signal is used. A synchronous detection method and apparatus and a sensor signal detection apparatus are known in which unnecessary noise components having the same frequency as the harmonics are efficiently attenuated in an infinite attenuation region of an analog moving average filter (for example, Patent Document 1). reference).
JP 2003-65768 A (first page, FIG. 1)

  However, in the conventional example described in Patent Document 1, in order to remove a high frequency noise component from a detection signal after synchronous detection of a sense signal output from a gyro with a reference signal synchronized with a monitor signal. Since the moving average processing is performed by the analog moving average filter, it is possible to efficiently attenuate the unnecessary noise component of the same frequency as the reference signal and its harmonics generated by the synchronous detection. In order to attenuate the signal, it is necessary to further low-pass filter the signal after filtering with an analog moving average filter, and there is an unsolved problem that the low-pass filter cannot reduce the size and cost.

  Therefore, the present invention has been made paying attention to the above-mentioned unsolved problems of the conventional example, and a so-called leakage signal is generated by leakage of vibration energy of the tuning fork vibrator without performing synchronous detection that generates unnecessary noise. An object of the present invention is to provide a gyro signal detection device capable of removing a signal due to an electric charge (unnecessary charge component) generated in a detection element regardless of Coriolis force.

  In order to achieve the above object, a gyro signal detection device according to a first technical means includes a vibration gyro provided with a moving-side vibrator and a detection-side vibrator, and an analog detection signal output from the detection-side vibrator. An A / D conversion unit that converts the digital detection signal into a digital detection signal and a noise removal unit that removes noise by subjecting the digital detection signal output from the A / D conversion unit to Kalman filtering are provided.

In the first technical means, the analog detection signal output from the vibration-detecting vibrator of the vibration gyro is converted into a digital detection signal by the A / D converter, and the digital signal detection signal is subjected to Kalman filter processing to thereby convert the gyro. The complex amplitude of each frequency component can be estimated from the signal, and a gyro signal that is not affected by noise can be detected.
In the gyro signal detection device according to the second technical means, a pair of drive arms each having a pair of drive-side vibrators are arranged with a target axis interposed therebetween, and the pair of detection-side vibrators are placed on the target axis. A vibrating gyroscope having a detection arm having a differential amplification section that differentially amplifies analog signals output from the pair of detection-side vibrators, and a differential amplification signal output from the differential amplification section. An A / D converter for converting into a digital signal and a noise removing unit for removing noise by subjecting a digital detection signal output from the A / D converter to Kalman filtering are provided.

  In the second technical means, the vibration gyro is arranged with the drive arm sandwiching the detection arm, so that the leakage vibration transmitted from the drive arm at the target axis position can be canceled and each detection can be performed. The detection signal output from the side vibrator is in reverse phase, and this is amplified by the operational amplifier to cancel the influence of the drive vibration and obtain a genuine detection signal. This detection signal is converted into a digital detection signal. Then, the Kalman filter process is performed, so that the complex amplitude of each frequency component can be estimated from the gyro signal, and a gyro signal that is not affected by noise can be detected.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a perspective view showing an embodiment of the present invention, in which CT is a mobile phone, and a rectangular plate-like base member 1 and a lid member 2 covering the upper surface of the base member 1 are hinged. 3 connected to each other.
As shown in FIG. 1A, the base member 1 is provided with an operation portion 4 having operation buttons on its upper surface, and a microphone 5 is disposed on the side opposite to the hinge 3. As shown in b), a radio communication antenna 6 that protrudes to the outside and communicates with the radio communication base station is attached to the end face on the hinge 3 side.

As shown in FIG. 1A, the lid member 2 is provided with an image display unit 8 on the surface facing the base member 1, and a speaker 9 is provided on the opposite side of the image display unit 8 from the hinge 3. It has been. Here, as the image display unit 8, for example, a liquid crystal panel, an organic EL panel, a plasma display panel, or the like can be applied.
Further, as shown in FIG. 1B, the imaging device 10 is attached to the back surface of the lid member 2, and the angular velocity in the horizontal direction and the vertical direction is set inside the lid member 2 in the vicinity of the imaging device 10. Gyro sensors 11a and 11b for detection are provided.

As shown in FIG. 2, each of the gyro sensors 11a and 11b has a sensor main body 111 accommodated in a package 112 using a piezoelectric material or the like, and the package 112 has a configuration capable of accommodating the sensor main body 111. It is formed in a box shape.
The sensor main body 111 includes a square fixed substrate 113 formed into a thin plate by etching a crystal, and a pair of excitations arranged in parallel on one opposing side surface of the fixed substrate 113 via support portions 114 and 115. The vibration arms 116 and 117 for detection and the vibration arms 118 and 119 for detection connected to the other opposite side surface of the fixed substrate 113 are configured. Each of the vibrating arms for excitation 116 and 117 is formed with long grooves 120 and 121 arranged in the longitudinal direction, and excitation electrodes 122 and 123 having different polarities are arranged in the grooves 120 and 121, respectively. ing.

  In the gyro sensor 11, by applying a driving voltage to the excitation vibrating arms 116 and 117 from an excitation circuit 20 to be described later, as shown by an arrow E in FIG. Vibrate like moving away or away. At this time, as shown in FIG. 2, when the rotational angular velocity ω acts around the center O of the fixed substrate 113 in the plane of the paper, the Coriolis force Fc acts in the direction F of FIG. This vibration is transmitted to the detection vibrating arms 118 and 119 via the support portions 114 and 115 and the fixed substrate 113. In other words, the excitation vibrating arms 116 and 117 receive the Coriolis force Fc acting in the vector product direction of the vibration direction in the X-axis direction and the rotational angular velocity ω, and according to the following formula, It vibrates alternately in the -Y direction (walk vibration). The vibration is transmitted to the detection vibrating arms 118 and 119 via the support portions 114 and 115 and the fixed substrate 113, and the detection vibrating arms 118 and 119 vibrate as indicated by an arrow H in FIG.

Fc = 2mV · ω
Here, m is the mass of the vibrating portion of the excitation vibrating arms 116 and 117, and V is the velocity of the excitation vibrating arms 116 and 117.
For this reason, as shown in FIG. 3 showing a cross section of the detection vibrating arms 118 and 119, one detection electrode 124 of the detection vibrating arms 118 and 119 and the other detection are caused by the vibration in the H direction of FIG. An electric field as indicated by an arrow is generated between the working electrode 125. The angular velocity ω can be detected by taking out an electric field based on the vibration of the vibrating arms for detection 118 and 119 as a signal. By using the gyro sensor 11 having such a configuration, a small package of about 1 mm square can be obtained, and the size can be sufficiently mounted on the mobile phone CT.

The excitation electrodes 122 and 123 of the excitation vibrating arms 116 and 117 are connected to the amplifier 20A in the excitation circuit 23 in which the AGC circuit 22 controlled by the starting circuit 21 is interposed between the amplifiers 20A and 20B as shown in FIG. It is connected and driven between the input side and the output side of the amplifier 20B.
On the other hand, one detection electrode 124 of the detection vibrating arm 118 is connected to one input side of the differential amplifier 31 as a differential amplifier through the amplifier 30a, and the other detection electrode 125 is grounded to detect One detection electrode 124 of the vibrating arm 119 is connected to the other input side of the differential amplifier 31, and the other detection electrode 125 is grounded.

  The differential amplification output from the differential amplifier 31 is supplied to an A / D converter 32 as an A / D converter, and converted into a digital detection signal by the A / D converter 32. The detection signal is input to a microprocessor unit (MPU) 33 as a noise removal unit. Connected to the microprocessor unit 33 are a ROM 34 for storing a program executed by the microprocessor unit 33 and a RAM 35 for storing data and calculation results in the calculation process of the microprocessor unit 33.

In the microprocessor 33, the angular velocity detection process shown in FIG. 5 is executed.
This angular velocity detection process is executed as a timer interrupt process for each predetermined sampling period ΔT. First, the detection signal output from the differential amplifier 31 in step S1 and converted into a digital signal by the A / D converter 32 is used. reading, then the process proceeds to step S2, and sets the detection signal I read as an observed value y _k, observed value y _k is, as represented by the following formula (1), specifies the number n arbitrarily varying Considering the case where it is represented by the sum of complex amplitudes a _i (k), which are time-varying components (n is an arbitrary integer), and observation noise v _k , the complex amplitude of each frequency component ω _i from observation value y _k A Kalman filter process for estimating a _i (k) is executed, and then the process proceeds to step S3, where the gyro detection is performed by performing the calculation of the above equation (1) based on the complex amplitude a _i (k) of each frequency component ω _i. The signal Sj is calculated and this is corrected From the output of the equal terminates the timer interrupt processing returns to the predetermined main program.

In this Kalman filter processing in step S2, the complex amplitude a _i (k) of each frequency component ω _i is estimated from the observed value y _k, thereby removing external vibration (linear acceleration) applied to the gyro sensor 11 and noise due to disturbance. .

However, y _k , v _k and a _i (k) are observation signals, observation noises and time-varying components at time t _k = kΔT (ΔT is a sampling interval), and ω _i is a time-varying component a _i as a complex number, respectively. The angular frequency of (k), j (j ² = −1) is an imaginary unit, and v _k is a stationary complex Gaussian white noise with an average value of zero and a variance σ _v ² .

Here, the Kalman filter processing is intended for a linear stochastic system, and receives an observation signal y _i (i = 0,..., K) as an input, and sequentially outputs an optimum estimated value of the system state u _k (n × 1 vector). Filter (algorithm).
The basic model of the Kalman filter is expressed by the following equations (2) and (3) and is called a state space model.

u _{k + 1} = F _k u _k + w _k (2)
y _k = H _k u _k + v _k (3)

Here, the above equation (2) is referred to as a state equation, and the above equation (3) is referred to as an observation equation. U _k is referred to as a system state or a state vector at time t _k .
Table In this case, the state u _{k + 1} of the system at time t _{k + 1} is multiplied by the state transition matrix F _k to the state u _k of the system at time t _k, by adding white noise w _k called system noise Is done. This is generally based on the fact that a signal with arbitrary statistical properties can be expressed as the output of a linear dynamic system with white noise added.

On the other hand, under normal circumstances, is often the system state u _k of can not be observed directly, it is common to be observed as a linear function. Therefore, the output of the system, that is, the observed value y _k, is expressed as the state u _k multiplied by the observation matrix H _k plus white noise v _k called observation noise.
Here, considering the irregularity of noise, {u _k }, {y _k }, {w _k }, {v _k } are treated as stochastic processes, and generally take vector values (u _k , w _{k). ^{∈R N, y k, v k}} ∈R M). Further, the system parameters F _k and H _k may be derived from physical laws governing the phenomenon, or may be derived from correlations measured in advance in experiments, and are generally determined matrices.

In this system model, continuous k + 1 pieces of observations y _0, ..., minimum variance estimate of u _k when y _k is given _{U k | k = E {u} k | y 0, ..., y k} Alternatively, the problem of obtaining U _{k | k−1} = E {u _k | y ₀ ,..., Y _k−1 } is called a Kalman filtering problem, and an algorithm that gives the solution is called a Kalman filter.
In order to solve the Kalman filtering problem, it is assumed that the system noise {w _k }, the observation noise {v _k }, and the initial state u ₀ are as follows.

E {v _i · v _j ^T } = δ _ij R _i ,
E {w _i · w _j ^T } = δ _ij Q _i ,
E {v _i · w _j ^T } = 0,
E {u ₀ · w _k ^T } = 0
E {u ₀ · v _k ^T } = 0
E {u ₀ } = u _a0 ,
E {[u ₀ −u _a0 ] [u ₀ −u _a0 ] ^T } = p ₀ (4)

Here, δ _ij is 1 when i = j, and is 0 otherwise, una ₀ is a known mean vector, R _k , Q _k and p ₀ are known covariance matrices, u ₀ , {W _k } and {v _k } are respectively Gaussian.

The result of solving the Kalman filtering problem based on the assumption of the above equation (4) is as follows.
U _{k | k} = U _{k | k−1} + K _k (y _k −H _k U _{k | k−1} ) (5)
U _{k + 1 | k} = F _k U _{k | k} (6)
_{K k = P k | k-} 1 H k T (H k P k | k-1 H k T + R k) -1 ......... (7)
_{P k | k = P k |} k-1 -K k H k P k | k-1 ......... (8)
_{P k + 1 | k = F} k P k | k F k T + Q k ......... (9)
U _{0 | -1} = u _a0 , P _{0 | -1} = p ₀ (10)
Here, K _k is a filter gain.

Using the above (5) to (10), by observing the signal y _k while updating the time k, it is possible to estimate the state u _k of the system.
The calculation procedure of the Kalman filter is as shown in FIG. 6. First, K ₀ is obtained from p ₀ and R ₀ , P _{0 | 0} is obtained from p ₀ and K _0, and U ₀ is obtained from u _a0 , K ₀ and y _{0. 0 | 0} can be obtained. Then, P 0 _{| 0} and Q ₀ Tokyo P _{1 |} seeking _0, further P _{1 | |} U _{1 0 | 0,} U 0 seek K ₁ from ₀ and R ₁ Tokyo, P 1 _{| 0} and K U _{1 | 1} can be obtained from ₁ and P _{1 | 1} and U _{1 | 0} , K ₁ and y ₁ .

Minimum words, the average vector u _a0 covariance matrix _{R 0, ..., R k,} Q 0, ..., Q k, and p _0, observed value y _0, ..., given the and y _k, the mean square error an optimum estimated value U ₀ to _{| 0,} ..., U k _| covariance matrix of _k and the estimated error _{P 0 | 0, ..., P} k | can be sequentially determines _k. In other words, the Kalman filter would be that sequentially obtains a conditional mean vector and covariance matrix of u _k, particularly in the case of Gaussian, which is seeking the conditional probability density function of u _k that changes every moment Means that.

Next, a method for estimating the complex amplitude a _i (k) of each frequency component ω _i using this Kalman filter processing will be described.
When the gyro detection signal is expressed by (1) as described above, the detection frequency of the frequency ω _i h @ ground color detection system is expressed. In estimating the time-varying spectrum by the Kalman filter processing, the time variation of the time-varying component a _i (k) is linearly approximated as follows.

That is, the slope of the change of the time-varying component a _i (k) at time t _k−1 is almost equal to the slope at time t _k , and ω _i (k) ′ is an approximation error at that time. . At this time, {a _i (k)} is expressed by the following second-order AR model.
a _i (k + 1) = 2 a _i (k) −a _i (k−1) + w _i (k) w _i (k) = ΔTw _i (k) ′ (12)
Using this, the above equation (1) is

It can be expressed as.
Here, it is assumed that {ω _i (k)} is a stationary complex Gaussian white noise having an average of 0, a variance average of 0, and a variance σ _w ² and independent of {v _k }.
Next, {ω _i (k)} is treated as system noise, and the observation signal {y _k } can be expressed by a basic system consisting of the following state equation and observation equation by modifying equation (13). .

x _{k + 1} = Fx _k + Aw _k (14)
y _{k + 1} = H _k x _k + v _k k = 0, 1, 2,... (15)
However,

Here, I _n × _n and 0 _n × _n represent an n × n unit matrix and a zero matrix, and A ^* and ^AT represent a complex conjugate and transpose of A, respectively. Further, diag [*] represents a diagonal matrix having * as a diagonal component, δ _ij represents a Kronecker delta, and E [*] represents an expected value of *.
At this time, assuming that x ₀ is independent of {w _k } and {v _k }, the following discrete time-varying spectrum estimation method is finally obtained by the complex Kalman filter processing.

ただし、

However,

Thereby, the complex amplitude a _i (k) of each frequency component ω _i can be estimated from the detection signal (observed value) y _k of the gyro sensor 11.
Then, by substituting the estimated complex amplitude a _i (k) into the equation (1), the gyro detection signal Sj can be calculated, and this gyro detection signal Sj is applied to a processing circuit such as a camera shake correction circuit (not shown). Is output.

  As described above, according to the above embodiment, the gyro sensor 11 is estimated by reading the detection signal of the gyro sensor 11 and performing the above-described Kalman filter processing on the detection signal, so that the gyro sensor 11 is not subjected to synchronous detection. Noise can be extracted from the detected signal and only the desired signal component can be extracted. As in the case of providing a synchronous detection circuit, processing such as switch switching and signal path selection is performed in synchronization with the reference signal. Therefore, it is possible to reliably prevent high frequency noise from being generated. Therefore, a low-pass filter that removes high-frequency noise can be omitted, and downsizing and cost reduction can be achieved.

In the above embodiment, the case where the Kalman filter process is executed by applying the microprocessor has been described. However, the present invention is not limited to this, and the Kalman filter may be configured by hardware and applied. Good.
In the above-described embodiment, the case where the present invention is applied to a horizontal vibration gyroscope as the gyro sensor 11 is described. However, the present invention is not limited to this, and a pair of gyro sensors 11 as illustrated in FIG. The drive piece 201 and the pair of detection pieces 202 are arranged with the rotation axis 203 symmetrical, and the pair of drive pieces 202 and the pair of detection pieces 202 are joined by a connecting portion 204. The present invention can also be applied to. Here, the coupling part 204 has a space inside, and has a pair of driving pieces 201, a pair of detection pieces 202, and a support part 205 that supports the joint part 204 at the center part. Is bonded and fixed to the housing 206.

  A drive electrode 207 is formed on the pair of drive pieces 201, and a detection electrode 208 is formed on the pair of detection pieces 202 by vapor deposition with a metal such as Au. Then, an electrode pattern is wired from each electrode 207, 208 to the support portion 205, and from a pad provided on the support portion 205, a wire is wired to a lead (not shown) of the housing 206 by an Au wire bonding wire, and a signal is transmitted. Input / output is configured.

  Furthermore, although the case where the present invention is applied to a mobile phone has been described in the above embodiment, the present invention is not limited to this, and the present invention is applied to a digital camera, a car navigation device, and other devices that require an angular velocity sensor. Can be applied.

It is a perspective view which shows one Embodiment at the time of applying this invention to a mobile telephone. It is a block diagram which shows a gyro sensor. It is the DD sectional schematic sectional drawing of FIG. It is a block diagram which shows a gyro signal detection apparatus. It is a flowchart which shows an example of the angular velocity detection processing procedure performed with a microprocessor. It is explanatory drawing which shows the calculation procedure of a Kalman filter. It is a perspective view which shows the other example of the gyro sensor which can apply this invention.

Explanation of symbols

  CT: mobile phone, 11a, 11b: gyro sensor, 116, 117 ... vibration arm for excitation, 118, 119 ... vibration arm for detection, 20 ... excitation circuit, 31 ... differential amplifier, 32 ... A / D converter, 33 ... Microprocessor

Claims (2)

  A vibration gyro provided with a drive-side vibrator and a detection-side vibrator, an A / D converter that converts an analog detection signal output from the detection-side vibrator into a digital detection signal, and an output from the A / D converter A gyro signal detection apparatus comprising: a noise removal unit that removes noise by subjecting the digital detection signal to Kalman filtering.   A pair of drive arms each having a pair of drive-side vibrators are arranged so as to face each other with a pair of drive-side vibrators and a detection arm having a pair of detection-side vibrators. A differential amplifier that differentially amplifies the analog signal output from the vibrator, an A / D converter that converts the differential amplified signal output from the differential amplifier into a digital signal, and the A / D converter A gyro signal detection apparatus comprising: a noise removal unit that removes noise by subjecting a digital detection signal output from the unit to Kalman filtering.

Images