JP6352117B2 - Filter device and filtering method - Google Patents

Description

  The present invention relates to a sensor signal filtering device and a filtering method used between a network and various sensors related to cloud computer networks and M2M communication.

  Signals from various sensors are usually converted into electrical signals such as voltages and currents that are easy to process using a physical phenomenon that reacts to a measurement target. This electrical signal becomes a means for signal processing when quantitatively measuring a predetermined physical quantity. Moreover, the signal converted into the electric signal has a time frequency characteristic according to a measuring object.

  By the way, in the process of sensing the physical quantity to be measured and converting it into an electrical signal, it is output as an electrical signal in which unnecessary noise is superimposed on a signal other than the physical phenomenon. For example, physical quantities such as thermometers and scales are converted to a low frequency range near DC, but ambient noise such as ambient temperature and wind, and supply voltage and current noise for conversion to electrical signals are more than that. It is normal to have a relatively high frequency variation. When extracting such a low-frequency signal, the technique of extracting a low-frequency fluctuation component having a frequency higher than that of direct current has a technical problem that requires a long processing time due to the uncertainty of time frequency. is there.

  Therefore, in order to solve the above technical problem, as a research example in which processing time is reduced using learning information of noise generation, as disclosed in Patent Document 1 below, A / D conversion is performed on the sensor signal, There are reports of applying low-pass filters using digital signal processing to noise reduction. The low-pass filter can remove even low-frequency noise components by setting the cutoff frequency width low.

  For example, in the conventional weighing device of Patent Document 1 shown in FIG. 5, an A / D conversion unit 51 that converts an analog weighing signal (input signal X) output from a weighing unit (not shown) into a digital signal, The vibration cycle calculation unit 52 that calculates the period of steady low-cycle vibration possessed by the means and generates a trigger corresponding to the period, and the unloaded weighing output by the weighing means when there is no measurement object in the transport means A waveform condition storage unit 53 for calculating and storing a waveform generation condition including a phase and an amplitude in order to generate a vibration correction waveform that approximates the waveform of the signal from a basic waveform function having a period calculated by the vibration period calculation unit 52; The correction waveform generator generates a correction signal for correcting the measurement signal from the condition for generating the vibration waveform stored in the waveform condition storage unit 53 with reference to the trigger generated from the vibration cycle calculation unit 52. Unit 54, a correction unit 55 that calculates the measurement value of the measurement object based on the difference between the measurement signal (input signal X) and the correction signal output from the correction waveform generation unit 54, and the measurement object calculated by the correction unit 55 A determination unit 56 for determining non-defective / defective products based on the measured values of the control unit, and an operation display unit 57 for displaying the determination result of the determination unit 56 and performing various operations. In the measuring device of Patent Document 1, a low-pass filter (not shown) is used to remove a high-frequency component from the measuring signal, and the peak value of the AC component of the measuring signal filtered by the low-pass filter is detected to determine the vibration period. Calculated.

  As described above, research has been conventionally conducted to remove a noise component from a sensor signal by a low-pass filter and extract the signal component with high speed and high accuracy. However, the calculation method is based on a linear operation.

  In signal processing using such linear operations, the analysis and evaluation procedures are uniquely determined, but the filter response time based on processing delay time and time frequency uncertainty is limited in principle. Accompany. Specifically, when the cut-off frequency of the low-pass filter is set low, there is a problem that it is necessary to lengthen the sensing period, and the response speed is slowed accordingly.

  In addition, when removing a noise component from a sensor signal, not only a low-pass filter but also a band-pass filter, a Hilbert transform method, or the like may be used. However, with a bandpass filter, for example, when a low-pass bandpass filter having a passband of 5 Hz to 40 Hz is designed, it is extremely difficult to attenuate a DC signal. On the other hand, in the Hilbert transform method, the direct current can be attenuated, but the low-frequency signal is attenuated. In addition, when the DC component is suppressed, the number of taps of the digital filter is remarkably increased and the processing delay is remarkably increased.

  These filtering methods extract a low-frequency component, which is a target signal component, using locality and bias of noise of amplitude frequency characteristics and phase frequency characteristics of an electrical signal as a sensor signal. If these processes are regarded as a system, it is formulated by an arithmetic model that processes the frequency phase component of the signal by convolving the frequency response component of the sensor signal with the impulse response of the processing system in the time domain. it can. This convolution operation has the property that the lower the frequency of the low frequency component to be extracted, the more the desired low frequency can not be extracted unless the number of response points of the impulse response is increased. That is, the processing time is long, and the processing time is represented by (number of response points of impulse response) * (sampling time T) / 2. Even if an analog method is used for this processing delay, τ is proportional to exp (−t / τ) in terms of the time constant τ of the processing system that is the main mode of analog processing, and τ time is the delay processing time. Therefore, in the low-frequency processing, τ increases, and the delay time until a normal signal is transmitted is increased.

  If the processing delay of such a sensor signal is allowed, the nature of the signal that changes from moment to moment is overlooked, and there is a possibility that the acquisition opportunity of the signal that fluctuates with time may be lost. Processing is an extremely important technical issue.

  Therefore, in order to solve the important technical problem described above, the following Patent Document 2 is disclosed as a filtering device and a filtering method that realize signal processing that reduces processing delay.

  In the filtering device and the filtering method disclosed in Patent Document 2, a desired low-frequency signal is not performed on the spatio-temporal characteristics of the sensor signal without performing the low-frequency signal extraction processing by the above-described convolution integration processing that takes a long processing time. The idea is to extract the high-frequency signal on the opposite side and extract the low-frequency signal by computing the high-frequency processed signal from the entire signal. Moreover, in Patent Document 2, a signal processing form is not a form in which a result is obtained by sequentially processing from the initial stage of a signal, but a real-time measurement result can be extracted for each section and section no matter where the signal section starts. The method is adopted.

JP 2008-268068 A JP2013-192020A

  However, although the filtering device and the filtering method disclosed in Patent Document 2 described above are methods for extracting a DC component of the current signal, calculation by correlation processing is performed to obtain DC, and an optimum value is obtained by a search method. Since the means for obtaining the value is used, it cannot be said that the amount of calculation is small, and there is a problem that processing takes time. For this reason, it has been desired to provide a filtering device and a filtering method capable of greatly reducing the amount of calculation and obtaining an accurate DC component while taking advantage of the filtering device and the filtering method disclosed in Patent Document 2. .

  Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to provide a filtering device and a filtering method that can significantly reduce the amount of calculation and obtain an accurate DC component.

In order to achieve the above object, a filtering device according to claim 1 of the present invention is for extracting a signal Y of a desired DC component from an analog input signal X including a noise signal component and a desired DC component. In the filtering device 1,
An A / D converter 2 that samples the input signal at a predetermined sampling period and converts the input signal into a digital full-band signal A including substantially all signal bands of the input signal;
Calculate the difference between the full-band signal from the A / D converter and the one-sample delayed signal of the full-band signal, or the difference between the one-sample delay of the differential signal and output the differential signal B A differentiator 3 to
A reference point detection unit 4 for detecting a preset reference point P from the difference signal;
An integrator 7 for integrating the difference signal when the reference point detection unit detects the reference point from the reference point and outputting an integrated signal C;
And a first subtracter 9 for subtracting the integral signal from the full-band signal by matching input timings of the full-band signal and the integral signal.

The filtering device according to claim 2 is the filtering device according to claim 1,
A reference point detection signal generator 10 for generating a reference point detection signal E for detecting the reference point P from the full-band signal A;
When the reference point detection unit 4 detects the reference point, routes R1 and R3 for inputting the full-band signal to the subtractor 3 and inputting the full-band signal to the first subtractor 9 are provided. And when the reference point detection unit does not detect the reference point, the entire band signal is input to the reference point detection signal generation unit and the reference point detection signal generation unit from the reference point detection signal generation unit It is characterized by comprising switchers 11 and 12 for switching paths so as to select paths R2 and R4 for inputting signals to the first subtractor and the differencer.

The filtering method according to claim 3 is a filtering method for extracting the signal Y of the desired DC component from the analog input signal X including a noise signal component and a desired DC component.
Sampling the input signal at a predetermined sampling period and converting it to a digital full-band signal A including substantially all signal bands of the input signal;
Calculating a difference between the full-band signal and a signal of one sample delay of the full-band signal, or a difference of one sample delay of the difference signal due to the difference;
Detecting a preset reference point P from the difference signal B based on the difference;
Integrating the difference signal when the reference point is detected from the reference point;
And subtracting the integration signal from the full-band signal by matching the input timing of the integration signal C by the integration with the full-band signal.

The filtering method according to claim 4 is the filtering method according to claim 3,
Generating a reference point detection signal E for detecting the reference point P from the full-band signal A;
When the reference point is detected, a difference based on the entire band signal is calculated, and the integration signal C is subtracted from the entire band signal by matching the input timing of the integration signal C by the integration with the entire band signal. Based on a subtract signal obtained by selecting the paths R1 and R3 and generating the reference point detection signal from the entire band signal and subtracting the reference point detection signal from the entire band signal when the reference point is not detected. A step of calculating a difference and switching the path so as to select paths R2 and R4 for subtracting the integral signal from the difference signal by matching an input timing of the difference signal based on the difference and an integral signal obtained by integrating the difference signal It is characterized by including.

  According to the filtering device and the filtering method of the present invention, an AC signal specified by a preset reference point is generated, and the generated AC signal is removed from the input signal (current signal: full-band signal). Thereby, compared with the conventional filtering apparatus and the filtering method, a calculation amount can be reduced significantly and an exact DC component can be extracted and output.

  Further, according to the filtering device described in claim 2 and the filtering method described in claim 4, even if the reference point cannot be detected because the desired DC component is buried in low-frequency noise having a long period, A signal for detecting the reference point is generated using the input signal, and an AC signal specifying the reference point can be removed from the current signal to obtain an accurate DC signal.

It is a schematic block diagram which shows the apparatus structure of 1st Embodiment of the filtering apparatus which concerns on this invention. It is a flowchart figure of the filtering method using the filtering apparatus of FIG. It is a schematic block diagram which shows the apparatus structure of 2nd Embodiment of the filtering apparatus which concerns on this invention. It is a flowchart figure of the filtering method using the filtering apparatus of FIG. It is a schematic block diagram which shows the apparatus structure of the conventional weighing | measuring apparatus.

[Technical Field of the Invention]
DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be specifically described with reference to the drawings.
The filtering device and filtering method according to the present invention relates to information processing technology for sensor signals used in various sensor networks related to cloud computer networks and M2M (Machine to Machine) communication, and is a signal processing technology for enhancing sensor signals. is there.

[Overview of the present invention]
The filtering device and the filtering method according to the present invention extract in a short time in a measurement section in which low-frequency signal components included in sensor signals and the like used between a cloud computer network and various sensors related to M2M communication and the network are cut out. A technique for processing is provided, and a desired signal is filtered without attenuating a low-frequency signal component.

  Furthermore, the filtering device and filtering method according to the present invention is not a low-frequency signal extraction process by convolution integration processing that takes a long processing time for the spatio-temporal characteristics of the sensor signal, but is opposite to a desired low-frequency signal. The high frequency signal on the side is extracted, the high frequency processing signal is calculated from the whole signal, and the concept of extracting the low frequency signal is adopted. In addition to adopting a real-time technique that can be extracted, a decisive technique for determining a reference point (reference value) is employed to solve the problem of Patent Document 2, thereby realizing a significant reduction in calculation amount.

[Whole configuration of the filtering device according to the first embodiment]
As shown in FIG. 1, the filtering device 1 (1 </ b> A) according to the first embodiment includes an A / D converter 2, a subtractor 3, a reference point detector 4, a timing generator 5, a first delay device 6, An integrator 7, a second delay unit 8, and a first subtracter 9 are provided for schematic configuration.

  The A / D conversion unit 2 samples the analog input signal X at a predetermined sampling period and converts it into a digital signal, and generates a full-band signal A including almost all signal bands of the input signal X. The full-band signal A is input to the subtractor 3 and is also input to the first subtracter 9 via the second delay unit 8.

  The differentiator 3 calculates the difference between the full-band signal A input from the A / D converter 2 and a signal obtained by delaying the full-band signal A by one sample. The difference signal B obtained by this calculation is an AC signal without DC, and is input to the reference point detection unit 4 and input to the integrator 7 via the first delay unit 6.

  The reference point detection unit 4 detects whether or not there is a preset reference point P for the difference signal B input from the differentiator 3. The reference point P is the initial initial value of the output of the subtractor 3, and is set to a value near 0 (voltage value or current value) including 0 in consideration of reducing the amount of calculation and obtaining an appropriate alternating current. It is preferable to do this. In this example, the reference point P = 0.

  The timing generator 5 is a timing signal T1 that gives a timing for delaying the output of the differentiator 3 when the reference point detector 4 detects the reference point P, and a timing at which the integrator 7 starts to integrate the difference signal B. The timing signal T2 is generated.

  The first delay unit 6 uses the timing signal T1 generated by the timing generation unit 5 as a trigger, and outputs the difference signal B input from the difference unit 3 for a predetermined time (corresponding to the time required for processing of the reference point detection unit 4). The delay time of the differential signal B input to the integrator 7 is adjusted by delaying.

  When the difference signal B is input from the difference unit 3 via the first delay unit 6, the integrator 7 starts from the time when the reference point P is detected using the timing signal T <b> 2 generated by the timing generation unit 5 as a trigger. The difference signal B is integrated. Since the integrator 7 reproduces the profile of the alternating current component of the current waveform of the input signal X, the integrator 7 has a frequency characteristic opposite to that of the subtractor 3. The integration value starting from the reference point P by the integrator 7 is input to the first subtractor 9 as an integration signal C by an AC signal having a clear reference point P.

  The second delay unit 8 uses the time required for the difference process of the difference unit 3 for the entire band signal A input from the A / D conversion unit 2, the time required for the detection process of the reference point P of the reference point detection unit 4, The first delay device 6 is delayed while maintaining its phase characteristic by a time equivalent to the sum of the delay time of the first delay device 6 and the time required for the integration processing of the integrator 7. In other words, the second delay unit 8 is arranged so that the input timings of the full-band signal A input to the first subtracter 9 and the integration signal C input from the integrator 7 to the first subtracter 9 match. The delay time of the entire band signal A is adjusted.

  The first subtracter 9 subtracts the integration signal C from the integrator 7 from the full-band signal A input from the A / D conversion unit 2 via the second delay unit 8 to obtain the input signal X which is the current signal. The desired DC signal Y based on the accurate DC component included in the signal is output.

[Example of processing operation by the filtering device of the first embodiment]
As a filtering method using the filtering device 1A of the first embodiment, a processing operation until an accurate DC component is extracted from the input signal X and a desired DC signal Y is output will be described with reference to the flowchart of FIG. To do.

  When the input signal (sensor electrical signal) X is input, the A / D converter 2 samples the input signal X at a predetermined sampling period, and the full-band signal A including almost all the signal band of the input signal X. (ST1: Sampling process). The full-band signal A generated by the A / D conversion unit 2 is input to the differentiator 3 and the second delay unit 8, respectively.

  When the full-band signal A is input from the A / D converter 2, the subtractor 3 calculates a difference between the full-band signal A and the full-band signal that is one sample delayed from the full-band signal A (ST2: Differential processing). The difference signal B calculated by the difference unit 3 is input to the reference point detection unit 4 and the first delay unit 6, respectively.

  The reference point detection unit 4 receives the difference signal B from the difference unit 3 and outputs a signal indicating that the reference point P has been detected to the timing generation unit 5 when a preset reference point P is detected from the difference signal B. (ST3: reference point detection process).

  When the signal indicating that the reference point P is detected is input from the reference point detection unit 4, the timing generation unit 5 starts the integration of the timing signal T <b> 1 that gives the timing for delaying the output of the differencer 3 and the integrator 7. And a timing signal T2 that gives the timing to be generated.

  Then, the difference signal B of the difference unit 3 is adjusted by the first delay unit 6 using the timing signal T 1 generated by the timing generation unit 5 as a trigger, and is input to the integrator 7.

  When the difference signal B is input from the difference unit 3 through the first delay unit 6, the integrator 7 uses the timing signal T <b> 2 of the timing generation unit 5 as a trigger to detect the difference signal B from the time when the reference point P is detected. Is integrated (ST4: integration processing).

  The second delay unit 8 includes the time required for the difference process of the difference unit 3 for the entire band signal A branched from the A / D conversion unit 2, the time required for the detection process of the reference point P of the reference point detection unit 4, The first delay device 6 outputs the signal while delaying it while maintaining its phase characteristic for a time equivalent to the sum of the delay time of the first delay device 6 and the time required for the integration processing of the integrator 7.

  Then, the first subtracter 9 subtracts the integration signal C from the integrator 7 from the full-band signal A input from the A / D converter 2 via the second delay unit 8, and inputs the current signal input signal. An accurate DC component included in X is extracted to output a desired DC signal Y (ST5: DC component extraction process).

[Overall Configuration of Filtering Device According to Second Embodiment]
As shown in FIG. 3, the filtering device 1 (1B) according to the second embodiment includes an A / D conversion unit 2, a subtractor 3, a reference point detection unit 4, and the filtering device 1A according to the first embodiment described above. In addition to the configuration of the timing generator 5, the first delay unit 6, the integrator 7, the second delay unit 8, and the first subtracter 9, the reference point detection signal generator 10, the first switch 11, A second switch 12 is further provided.

  In the filtering device 1B according to the second embodiment, the filtering device 1A according to the first embodiment described above is configured by the additional configuration of the reference point detection signal generation unit 10, the first switch 11, and the second switch 12. The problem when the reference point detection unit 4 cannot detect the reference point P is solved. In the filtering device 1B according to the second embodiment, the same components as those of the filtering device 1A according to the first embodiment are denoted by the same reference numerals.

  The A / D conversion unit 2 samples the analog input signal X at a predetermined sampling period and converts it into a digital signal, and generates a full-band signal A including almost all signal bands of the input signal X. This full-band signal A is input to the reference point detection signal generator 10 and is selected by the differencer 3 and the second delayer 8 by switching the contacts of the first switch 11 and the second switch 12. Input.

  The difference unit 3 calculates the difference between the full-band signal A input from the A / D converter 2 via the second switch 12 and a signal obtained by delaying the full-band signal A by one sample. . The difference signal B obtained by this calculation is an AC signal without DC, and is input to the reference point detection unit 4 and integrated via the first delay device 6 by the timing signal generated by the timing generation unit 5. Is input to the device 7.

  The reference point detection unit 4 detects whether or not there is a preset reference point P for the difference signal B input from the differentiator 3. The reference point P is the initial initial value of the output of the subtractor 3, and is set to a value near 0 (voltage value or current value) including 0 in consideration of reducing the amount of calculation and obtaining an appropriate alternating current. It is preferable to do this. In this example, the reference point P = 0.

  Further, the reference point detection unit 4 uses the switching signal S1 for switching the contact of the first switch 11 and the switching signal S2 for switching the contact of the second switch 12 as to whether the reference point P is detected. Outputs accordingly.

  The timing generator 5 is a timing signal T1 that gives a timing for delaying the output of the differentiator 3 when the reference point detector 4 detects the reference point P, and a timing at which the integrator 7 starts to integrate the difference signal B. The timing signal T2 is generated.

  The first delay unit 6 uses the timing signal T1 generated by the timing generation unit 5 as a trigger, and outputs the difference signal B input from the difference unit 3 for a predetermined time (corresponding to the time required for processing of the reference point detection unit 4). The delay time of the differential signal B input to the integrator 7 is adjusted by delaying.

  When the difference signal B is input from the difference unit 3 via the first delay unit 6, the integrator 7 starts from the time when the reference point P is detected using the timing signal T <b> 2 generated by the timing generation unit 5 as a trigger. The difference signal B is integrated. Since the integrator 7 reproduces the profile of the alternating current component of the current waveform of the input signal X, the integrator 7 has a frequency characteristic opposite to that of the subtractor 3. The integration value starting from the reference point P by the integrator 7 is input to the first subtractor 9 as an integration signal C by an AC signal having a clear reference point P.

  The second delay unit 8 is connected between the A / D conversion unit 2 and the second delay unit 8 via the first switch 11, and between the A / D conversion unit 2 and the difference unit 3. Are connected to each other via the second switch 12, equivalent to the sum of the time required for the differential processing of the differentiator 3, the delay time of the first delay device 6, and the time required for the integration processing of the integrator 7. Is delayed for a period of time while maintaining the phase characteristics of the entire band signal A input from the A / D converter 2 via the second switch 12.

  The second delay unit 8 is connected between the reference point detection signal generation unit 10 and the second delay unit 8 via the first switch 11, and the difference unit 3 and the reference point detection signal. When the generator 10 is connected via the second switch 12, the time required for the difference process of the difference unit 3, the delay time of the first delay unit 6, and the integration process of the integrator 7 The reference point detection signal E input from the reference point detection signal generation unit 10 via the second switch 12 is delayed while maintaining the phase characteristic by a time equivalent to the sum of the times.

  The first subtracter 9 is delayed by the second delay unit 8 when the A / D converter 2 and the second delay unit 8 are connected via the first switch 11. Then, the integrated signal C by the integrator 7 is subtracted from the entire band signal A, and a desired DC signal Y based on an accurate DC component included in the input signal X which is the current signal is output.

  In addition, the first subtracter 9 is configured so that the second delay device is used when the reference point detection signal generator 10 and the second delay device 8 are connected via the first switch 11. The integration signal C by the integrator 7 is subtracted from the reference point detection signal E delayed by 8, and a DC signal Y based on the DC component included in the reference point detection signal E is output.

  The reference point detection signal generation unit 10 generates a reference point detection signal E for detecting the reference point P. As shown in FIG. 3, the signal generation unit 10a, the third delay unit 10b, 2 subtractors 10c.

  In the signal generation unit 10a, the second subtractor 10c and the second delay unit 8 are connected via the first switch 11, and the second subtractor 10c and the difference unit 3 are connected to each other. 2, the I signal and the Q signal of the entire band signal A input from the A / D conversion unit 2 are extracted, and based on the extracted I signal and the Q signal, An inverted signal D having a phase inverted by 180 degrees with respect to the entire band signal A is generated. The inverted signal D is a signal that is input to the second subtractor 10c and serves as a reference point P when subtracted from the full-band signal A by the second subtractor 10c.

  The third delay unit 10 b is connected between the second subtractor 10 c and the second delay unit 8 via the first switch 11, and between the second subtractor 10 c and the difference unit 3. Are connected via the second switch 12, the entire band signal A input from the A / D conversion unit 2 is delayed by a time equivalent to the time required for the signal generation processing of the signal generation unit 10a. Output.

  The second subtractor 10b is connected between the second subtractor 10c and the second delay unit 8 via the first switch 11, and between the second subtractor 10c and the difference unit 3. Are connected via the second switch 12, the inverted signal D generated by the signal generator 10a is subtracted from the full-band signal A delayed by the third delay 10b, The obtained signal (AC average 0 signal) is input to the second delayer 8 via the first switch 11 as the reference point detection signal E and the difference via the second switch 12. Input to the device 3.

  As shown in FIG. 3, the first switch 11 has two fixed contacts 11a and 11b and one movable contact 11c. In the first switch 11, the fixed contact 11a is connected to the A / D converter 2, the signal generator 10a, and the third delay device 10b, respectively, and the fixed contact 11b is connected to the second subtractor 10c and the second switch. The movable contact 11 c is connected to the second delay device 8.

  In the first switch 11, the movable contact 11 c is selectively switched to the fixed contact 11 a or the fixed contact 11 b by a switching signal S <b> 1 from the reference point detection unit 4 according to whether or not the reference point P is detected. More specifically, the first switch 11 is configured such that the movable contact 11c is fixed to the fixed contact 11a by the initial state when the filtering device 1B is activated and the switching signal S1 when the reference point detection unit 4 detects the reference point P. Switched to the side. As a result, a path R1 connecting the A / D converter 2 and the second delay unit 8 is selected.

  Further, in the first switch 11, the movable contact 11 c is switched to the fixed contact 11 b side by the switching signal S <b> 1 when the reference point detection unit 4 does not detect the reference point P, as shown in FIG. 3. As a result, the path R2 that connects the second subtractor 10c and the second delay unit 8 is selected.

  As shown in FIG. 3, the second switch 12 includes two fixed contacts 12a and 12b and one movable contact 12c. In the second switch 12, the fixed contact 12 a is connected to the A / D conversion unit 2, the second delay unit 8, the signal generation unit 10 a, and the third delay unit 10 b, and the fixed contact 12 b is the second subtraction. The movable contact 12 c is connected to the subtractor 3.

  In the second switch 12, the movable contact 12 c is selectively switched to the fixed contact 12 a or the fixed contact 12 b by a switching signal S <b> 2 from the reference point detection unit 4 according to whether or not the reference point P is detected. More specifically, the second switch 12 is configured such that the movable contact 12c is fixed to the fixed contact 12a by the initial state when the filtering device 1B is activated or the switching signal S2 when the reference point detection unit 4 detects the reference point P. Switched to the side. Thereby, the path R3 connecting the A / D converter 2 and the differencer 3 is selected.

  Further, in the second switch 12, the movable contact 12c is switched to the fixed contact 12b side as shown in FIG. 3 by the switching signal S2 when the reference point detection unit 4 does not detect the reference point P. As a result, the path R4 connecting the second subtractor 10c and the differentiator 3 is selected.

[Example of processing operation of the filtering device according to the second embodiment]
As a filtering method using the filtering device 1B of the second embodiment, a processing operation until an accurate DC component is extracted from the input signal X and a desired DC signal Y is output will be described with reference to the flowchart of FIG. To do.

  In the filtering device 1B of the second embodiment shown in FIG. 3, the reference point P is detected, that is, the movable contact 11c of the first switch 11 is switched to the fixed contact 11a side, and the second In the state where the movable contact 12c of the switch 12 is switched to the fixed contact 12a side, the same processing operation as the filtering device 1A of the first embodiment described above is performed.

  In the filtering device 1B according to the second embodiment, when the reference point detector 4 does not detect the reference point P, the movable contact 11c of the first switch 11 is switched to the fixed contact 11b side by the switching signal S1, and the switching is performed. The movable contact 12c of the second switch 12 is switched to the fixed contact 12b side by the signal S2 (ST11: contact switching process). In this state, the connection between the A / D conversion unit 2 and the difference unit 3 and the connection between the A / D conversion unit 2 and the second delay unit 8 are disconnected, and the A / D conversion unit 2 The full-band signal A is input to the signal generator 10a and the third delay device 10b, respectively.

  When the full-band signal A is input from the A / D conversion unit 2, the signal generation unit 10a generates an inverted signal D having a phase inverted by 180 degrees with respect to the full-band signal A (ST12: Inverted signal generation processing) ). The inverted signal D generated by the signal generator 10a is input to the second subtracter 10c.

  The second subtractor 10c subtracts the inverted signal D input from the signal generation unit 10a from the entire band signal A delayed by the third delay unit 10b, and detects a reference point P for detecting the reference point P. Signal E is generated (ST13: reference point detection signal generation process). The reference point detection signal E obtained by this subtraction is an AC average 0 signal including the reference point P, and is input to the second delay device 8 via the first switch 11 and 2 is input to the differentiator 3 through the second switch 12.

  When the reference point detection signal E (an average 0 signal in terms of alternating current) is input from the second subtractor 10c, the differentiator 3 receives the reference point detection signal E (one of the full-band signal and the full-band signal). A difference (secondary difference) between the reference point detection signal E and the reference point detection signal of one sample delay of the reference point detection signal E is calculated (ST14: difference processing). The difference signal B calculated by the difference unit 3 is input to the reference point detection unit 4 and the first delay unit 6, respectively.

  When the difference signal B is input from the differentiator 3, the reference point detection unit 4 detects the reference point P from the difference signal B (ST15: reference point detection process), and outputs a signal indicating that the reference point P has been detected. Output to the timing generator 5.

  When the signal indicating that the reference point P is detected is input from the reference point detection unit 4, the timing generation unit 5 starts the integration of the timing signal T <b> 1 that gives the timing for delaying the output of the differencer 3 and the integrator 7. And a timing signal T2 that gives the timing to be generated.

  Then, the difference signal B of the difference unit 3 is adjusted by the first delay unit 6 using the timing signal T 1 generated by the timing generation unit 5 as a trigger, and is input to the integrator 7.

  When the difference signal B is input from the difference unit 3 through the first delay unit 6, the integrator 7 uses the timing signal T <b> 2 of the timing generation unit 5 as a trigger to detect the difference signal B from the time when the reference point P is detected. Are integrated (ST16: integration processing).

  When the reference point detection signal E is input from the second subtractor 10c via the first switch 11, the second delay unit 8 detects the reference point input from the second subtractor 10c. The signal E is delayed while maintaining its phase characteristic for a time equivalent to the sum of the time required for the difference processing of the difference unit 3, the delay time of the first delay unit 6, and the time required for integration processing of the integrator 7. Then, the reference point detection signal E is input to the first subtracter 9.

  Then, the first subtracter 9 subtracts the integration signal C from the integrator 7 from the reference point detection signal E input via the second delay unit 8 (ST17: subtraction processing), and detects the reference point. A DC signal Y included in the signal E is output.

  With the above operation, when the reference point detection unit 4 detects the reference point P from the difference signal B by the reference point detection signal E generated from the full-band signal A, one measurement session is ended, and the next measurement session is started. In order to shift, the movable contact 11c of the first switch 11 is switched to the fixed contact 11a side by the switching signal S1, and the movable contact 12c of the second switch 12 is switched to the fixed contact 12a side by the switching signal S2 ( (ST18: contact switching process), the above-described processes of ST1 to 5 in FIG. 2 are executed. When the reference point detection unit 4 does not detect the reference point P again, the above-described processing of ST11 to ST18 is repeated, and the reference point detection unit 4 detects the reference point P from the difference signal B by the reference point detection signal E. Then, the processes of ST1 to ST5 in FIG. 2 described above are executed.

  Note that the processing operation by the filtering devices 1A and 1B described above may be repeatedly performed in the measurement interval, and the DC signal Y obtained by this processing operation may be averaged. As a result, even when there is a fine viewpoint variation in the stable section in the measurement section, the influence of this fine viewpoint variation can be removed, and an accurate direct current component can be extracted to obtain a desired direct current signal Y, so that more accurate measurement can be performed. A value (DC signal Y) can be provided.

  Although not particularly illustrated, in the above-described filtering devices 1A and 1B, a low-pass filter may be provided immediately after the A / D converter 2 in order to remove noise superimposed on the DC component of the input signal X. .

  As described above, according to the filtering device and the filtering method according to the present invention, it is necessary to have an exploratory method for finding the estimated value that is the maximum value of the calculation or the correlation processing as disclosed in Patent Document 2. Instead, a decisive method based on a preset reference point P, that is, an AC signal specifying the reference point P is removed from the current signal (full-band signal A). Thereby, compared with patent document 2, a calculation amount can be reduced significantly and an exact direct current | flow component can be extracted, and the bandpass filter which does not impair the frequency component near direct current | flow can be comprised.

  In addition, according to the filtering device 1B and the filtering method according to the second embodiment, even when the desired DC component is buried in low-frequency noise having a long period and the reference point P cannot be detected, the current signal is used. A reference point detection signal E for detecting the reference point P is generated, and the reference point detection unit 4 detects the reference point P from the difference signal B based on the generated reference point detection signal E. As a result, the AC signal specifying the reference point P can be removed from the current signal, the calculation amount can be significantly reduced, and an accurate DC component can be extracted to obtain a desired DC signal.

  Although the best mode of the filtering device and the filtering method according to the present invention has been described above, the present invention is not limited by the description and drawings according to this mode. That is, it is a matter of course that all other forms, examples, operation techniques, and the like made by those skilled in the art based on this form are included in the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 (1A, 1B) Filter apparatus 2 A / D conversion part 3 Difference machine 4 Reference point detection part 5 Timing generation part 6 1st delay device 7 Integrator 8 2nd delay device 9 1st subtractor 10 Reference inspection Outgoing signal generator 10a Signal generator 10b Third delay unit 10c Second subtractor 11 First switcher 12 Second switcher X input signal Y DC signal A full-band signal B differential signal C integrated signal D Inverted signal E Reference point detection signal

Claims (4)

In the filtering device (1) for extracting the signal (Y) of the desired DC component from the analog input signal (X) including the noise signal component and the desired DC component,
An A / D converter (2) that samples the input signal at a predetermined sampling period and converts it into a digital full-band signal (A) including substantially all signal bands of the input signal;
The difference signal (B) is calculated by calculating the difference between the full-band signal from the A / D converter and the one-sample delay signal of the full-band signal, or the difference between the one-sample delay of the difference signal based on the difference. A differentiator (3) that outputs
A reference point detector (4) for detecting a preset reference point (P) from the difference signal;
An integrator (7) for integrating the difference signal when the reference point detecting unit detects the reference point from the reference point and outputting an integrated signal (C);
A filtering device, comprising: a first subtracter (9) for subtracting the integral signal from the full-band signal by matching input timings of the full-band signal and the integral signal. A reference point detection signal generator (10) for generating a reference point detection signal (E) for detecting the reference point (P) from the full-band signal (A);
When the reference point detection unit (4) detects the reference point, the full-band signal is input to the differentiator (3) and the full-band signal is input to the first subtracter (9). When the path (R1, R3) to be selected is selected and the reference point detection unit does not detect the reference point, the entire band signal is input to the reference point detection signal generation unit and the reference point detection signal And a switch (11, 12) for switching the path so as to select a path (R2, R4) for inputting the reference point detection signal from the generation unit to the first subtractor and the differencer. The filtering device according to claim 1, wherein: In the filtering method for extracting the signal (Y) of the desired DC component from the analog input signal (X) including the noise signal component and the desired DC component,
Sampling the input signal at a predetermined sampling period and converting it to a digital full-band signal (A) including substantially all signal bands of the input signal;
Calculating a difference between the full-band signal and a signal of one sample delay of the full-band signal, or a difference of one sample delay of the difference signal due to the difference;
Detecting a preset reference point (P) from the difference signal (B) by the difference;
Integrating the difference signal when the reference point is detected from the reference point;
And a step of subtracting the integral signal from the full-band signal by matching the input timing of the integral signal (C) by the integration with the full-band signal. Generating a reference point detection signal (E) for detecting the reference point (P) from the full-band signal (A);
When the reference point is detected, a difference based on the entire band signal is calculated, and the input signal of the integration signal (C) by the integration is matched with the input signal of the integration signal from the entire band signal. When a path (R1, R3) to be subtracted is selected and the reference point is not detected, the reference point detection signal is generated from the entire band signal, and the reference point detection signal is subtracted from the entire band signal The difference based on the subtraction signal is calculated, and the path (R2, R4) for subtracting the integration signal from the difference signal by matching the input timing of the difference signal based on the difference and the integration signal obtained by integrating the difference signal is selected. 4. The filtering method according to claim 3, further comprising the step of switching the path as described above.

Images