JP6126472B2 - Microwave Doppler detector - Google Patents

Description

  The present invention relates to a microwave Doppler detection device, and more particularly to a configuration of a detection device that is installed outdoors and is directly incorporated into a device / apparatus that may be rained, and detects the approach of a moving body such as a person using a Doppler signal.

  Conventionally, a device that uses a microwave and detects a moving body by the Doppler effect has been used. This type of detection device reduces false detection due to rainfall, reduces the influence of fluctuations in environmental noise level, and the like. .

  For example, in the security device of Patent Document 1 (Patent No. 4799173), the Doppler sensor is installed with an angle of 45 degrees with respect to the plastic window so that the rain always moves away from the Doppler sensor. Also, by using the quadrature mixer output, the approach and separation are determined, the direction in which rain flows is ignored as the separation direction from the sensor, and only the approaching person is detected.

  On the other hand, the obstacle detection sensor of Patent Document 2 (Japanese Patent Laid-Open No. 2006-275895) transmits a Doppler signal with a different amplification factor for each output frequency, and adjusts the sensitivity appropriately according to the situation. The presence or absence of is detected.

Japanese Patent No. 4799173 JP 2006-275895 A

However, in the conventional Doppler detection device, when water droplets gather and flow on the window (protective cover) due to rain or the like, there is a problem that good moving body detection cannot be performed due to the raindrop (raindrop, water film, etc.).
For example, when a person (purchaser) or the like is detected by a microwave Doppler sensor using a device such as a vending machine, the detection range is narrow and the radio wave reflected by the human body is relatively large. If the device is not directly exposed to rain, the effects of rain will not be a problem.
However, if rain is applied directly to the device and dripping onto a plastic window (protective cover) or housing for radiating microwaves, the reflected radio wave becomes very large and the mobile It will cause false detection.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide a microwave Doppler detection device that reduces the influence of dripping on a detection window and the like and reduces false detection of a moving object. is there.

To achieve the above object, the invention according to claim 1 is a microwave Doppler comprising: a Doppler sensor; and a quadrature detection circuit that generates two I and Q signals having a phase difference of 90 degrees from the output of the Doppler sensor. In the detection apparatus, the I and Q signals output from the quadrature detection circuit are sampled at a predetermined interval, and either the I signal or the Q signal is shifted by approximately ¼ period at the center frequency of the Doppler signal. A product / moving average processing circuit for obtaining a product with one signal and calculating a moving average of the product is provided. The moving average calculation of the product / moving average processing circuit is performed in a short time, and the short-time moving average is calculated. A signal when the number of times the value continuously exceeds the threshold range equal to or less than a predetermined positive threshold and equal to or greater than the negative threshold is determined as a spike, and the spike signal is determined as the threshold. And it replaces the value of the 囲内 signal.
The invention of claim 2 is characterized in that the predetermined number of times for determining the spike is set between 100 and 500 milliseconds (msec).
In the invention according to claim 3, in the short-time moving average calculation, a signal exceeding the positive threshold and a signal falling below the negative threshold or vice versa are continuously output, and the positive threshold is continuously exceeded. When each of the number of times and the number of times exceeding the negative threshold is equal to or less than the predetermined number, the positive / negative continuous signal is determined as a spike, and the spike signal is replaced with a signal having a value within the threshold range.

In the invention of claim 4, the signal obtained by the short-time moving average calculation continuously returns to the positive threshold value or less after the predetermined number of times or less and then returns to the positive threshold value or less, and the number of times equal to or less than the predetermined number of times. If it becomes equal to or greater than the positive threshold again after continuously falling below the positive threshold, the data below the positive threshold is interpolated with data that is equal to or greater than the positive threshold before and after that, and on the contrary After continuously falling below the negative threshold for the number of times less than or equal to the number of times, it returned to the negative threshold or more. Is characterized in that the data when the negative threshold value is exceeded is interpolated with data before and after the positive threshold value.
The invention of claim 5 is characterized in that the moving average calculation is executed again for the detection signal after the spike signal is replaced with a signal having a value within the threshold range.

The invention of claim 6 is a microwave Doppler detection device comprising a Doppler sensor and a quadrature detection circuit that generates two I and Q signals having a phase difference of 90 degrees from the output of the Doppler sensor. The output I and Q signals are sampled at a predetermined interval, and the product of the signal obtained by shifting either the I signal or the Q signal by approximately ¼ period at the center frequency of the Doppler signal and the other signal is obtained. And a product / moving average processing circuit for calculating a moving average of the product, performing the moving average calculation of the product / moving average processing circuit in a short time, and a plurality of times before the present time obtained by the short-time moving average calculation. The median of the data is extracted, and the median is replaced with the current data value.
The invention of claim 7 is characterized in that the moving average calculation is executed again for the detection signal (data value) after the median replacement.

  According to the above configuration, basically, I and Q signals such as rainfall are detected as random noise having no correlation between the signals, and the Doppler signal (I and Q signals) due to the movement of one person has a phase. It is observed as a signal having a correlation of 90 degrees. That is, the product of the signal obtained by shifting either the I or Q signal by approximately ¼ period at the center frequency of the Doppler signal and the other signal has a correlation between the I and Q signals, so In this case, there is a phase difference of about 90 degrees, and only a positive or negative value is obtained depending on the direction in which the phase is shifted. Therefore, a positive or negative value can be obtained by obtaining a moving average. On the other hand, when there is no correlation due to random noise, even if one is shifted and the product is taken, the result includes randomly plus and minus, so the moving average becomes a value close to zero. By such processing, noise due to rain or the like is reduced.

However, even with the above processing, if a raindrop (water film) directly hits the window immediately before the sensor to cause dripping (water film) and the reflected radio wave becomes large, erroneous detection occurs.
Therefore, the present invention performs the moving average calculation within a short set time (for example, several milliseconds to several tens of milliseconds), and sets the threshold range in which the value of the short-time moving average is equal to or less than a predetermined positive threshold value and equal to or more than a negative threshold value. A signal when the number of times of continuous (deviating) is less than or equal to a predetermined number is determined as a spike caused by dripping or a water film, and the spike signal is replaced with a signal having a value within the above threshold range, for example, a zero signal. Reduce spikes.

In the configuration of claim 3, the positive and negative continuous spikes [FIG. 3C] are removed, and in the configuration of claim 4, the signal including disconnection [FIG. 4D] and the signal including reverse spikes [FIG. (E)] is removed.
In the case of the configuration of claim 6 above, when the data that exceeds the threshold range is included in n data by (n + 1) / 2 times or more by obtaining the median value of n data before the present The median is a value that exceeds the threshold range, and when data that exceeds the threshold range by (n + 1) / 2 times or less is included, the median is within the threshold range. This is because the processing of claim 1 in which spike removal is performed by determining the number of times the threshold range has been continuously exceeded, or the reverse of (n-1) / 2 times or less when the threshold range is continuously exceeded. An equivalent result is obtained when interpolation processing is performed to replace the signal with a predetermined value when there is a direction spike. The processing is simplified by setting n to an odd number, and the processing becomes simple. However, it is also possible to obtain the median value of two points by setting n to an even number, and to determine the median by determining the average or size.
In the configurations of the fifth and seventh aspects, since the signal after replacing the spike signal with a signal having a value within the threshold range or the signal after replacing the median value is again subjected to a moving average calculation, Reduction is ensured.

  According to the microwave Doppler detection device of the present invention, it is possible to remove sporadic short-time spikes by determining whether or not the signal obtained by the short-time moving average exceeds the threshold range. It is possible to reduce the influence of dripping on the detection window and the like, and to reduce erroneous detection of the moving object.

Compared with the method of claim 1, the invention of claim 6 is capable of removing spikes by the median replacement process, and determines whether or not the threshold range is exceeded and counts the number of consecutive times exceeding the threshold range. There is an advantage that the processing becomes simple.
According to the fifth and seventh aspects of the present invention, by further performing a moving average calculation on the detection signal from which the spike has been removed, random noise can be reliably reduced, and a moving object such as a person can be detected satisfactorily. . In various apparatuses such as vending machines, it is possible to save energy by performing necessary operations such as turning on a display lamp and an electric lamp when a purchaser, a user, or the like approaches.

The processing circuit block of the microwave Doppler detection apparatus which concerns on the Example of this invention is shown, FIG. (B) is a figure which shows the structure in the microcomputer of FIG. (A). It is a side view which shows the structural example of the Doppler detection apparatus of an Example. It is explanatory drawing which shows the first half of the various patterns of the spike removal in a Doppler detection apparatus in an Example. It is explanatory drawing which shows the second half of the various patterns of the spike removal in a Doppler detection apparatus in an Example. The signal processed by the principal part of the Doppler detection apparatus of an Example is shown, A figure (A) is an output signal waveform figure from a Doppler sensor circuit part, A figure (B) is a signal waveform figure after a short-time moving average calculation, a figure (C) is a signal waveform diagram after removing spikes and performing a moving average calculation again, and FIG. (D) is a signal waveform diagram after detection determination. It is a figure for demonstrating the other processing method (median value extraction and replacement process) of spike determination and removal of the microwave Doppler detection apparatus of an Example.

  1A and 1B show a configuration of a microwave Doppler detection device according to an embodiment of the present invention, and FIG. 2 shows a configuration example of a device attached to a vending machine or the like. Yes. As shown in FIG. 2, in a vending machine or the like, a flat window (protective cover) 10 made of plastic or the like that transmits radio waves is provided on the front surface thereof, and a Doppler sensor 12 is attached to the inside of the window 10. The antenna surface (detection surface) 12A on the front surface of the Doppler sensor 12 is arranged so as to be parallel to the planar window 10, for example.

  1A includes the Doppler sensor 12 and includes a quadrature detection (mixer) circuit or the like, so that an I signal and a Q signal that are 90 degrees out of phase from the signal received by the sensor 12 are obtained. Output. High-pass filters (HPF) 15a and 15b for attenuating low frequencies corresponding to the movement of raindrops (water film), raindrops, etc. are provided in the subsequent stage of the sensor circuit unit 14 in response to two IQ signal outputs. In the high-pass filters 15a and 15b of the embodiment, the low-frequency cut-off frequency is set to about 50 Hz, for example. Amplifiers (amplifiers) 16a and 16b, AD (analog / digital) converters 17a and 17b, and a microcomputer 18 are connected to the high-pass filters 15a and 15b.

  That is, the amplifiers 16a and 16b amplify the IQ signal from which the low frequency range corresponding to the movement such as raindrop is removed, and the AD converters 17a and 17b convert the amplified IQ signal to, for example, 1 Doppler signal center frequency. Sampling is performed at a sampling time of / 8 period or less.

FIG. 1B shows an internal processing block (circuit) of the microcomputer 18, and the microcomputer 18 includes a memory 20a, A / A for storing the I signal output from the AD converter 17a. A memory 20b for storing the Q signal output from the D converter 17b is provided. These memories 20a and 20b each have approximately a quarter cycle of the Doppler signal center frequency _fd0 for each sampled IQ signal. Are stored as a data string for a predetermined time. In the embodiment, as described above, sampling is performed at a time interval of 1/8 period or less of the center frequency of the Doppler signal, and data sampled approximately ¼ period before is used as one of the I and Q signals. In order to take the product with the above data, the above-described operation of shifting the ¼ period can be easily realized even with a microcomputer having a small memory.

Further, as a product / moving average processing circuit, an accumulating unit 21, a short-time moving average unit 22, a spike determination / removal unit 23, and a second moving average unit 24 are provided in the subsequent stage of the memories 20a and 20b. The moving average processing circuit executes IQ signal product / moving average processing based on each data sequence of the IQ signal. The accumulating unit 21 accumulates data (for example, I signal) shifted by a time corresponding to approximately ¼ period of the Doppler signal center frequency f _d0 of the IQ signal and data (for example, Q signal) that has not been shifted, for the short time. The moving average unit 22 calculates an average value of data in a short time interval (from the current time to a certain time before), for example, several milliseconds (msec) to several tens of milliseconds, and uses these as a data array for a certain time. The spike determination / removal unit 23 spikes data (signal) in which the number of times that the threshold value range below the predetermined positive threshold value and the negative threshold value is continuously exceeded is not more than the predetermined number in the array data. And a continuous signal of a predetermined number of times or less is replaced with a value in the threshold range, for example, 0, and spikes are removed. The predetermined number of times is set within a range of 100 to 500 msec. For example, if the short-time moving average calculation is 30 msec, the spike determination is executed with the predetermined number of times being about ten.

  The second moving average unit 24 averages data for a signal (data string) output from the spike determination and removal unit 23 from the current time to a certain time before (interval longer than the short time interval). Calculate the value. A detection determination circuit 25 that finally determines a moving body such as a person is connected to the subsequent stage of the second moving average unit 24.

  The embodiment has the above configuration. First, in the embodiment, the antenna surface 12A of the Doppler sensor 12 is arranged in parallel to the flat window 10 so that raindrops flowing on the surface of the window 10 can be detected with respect to the sensor antenna surface 12A. Since it flows substantially in parallel, the flow such as raindrop is in a direction closer to the sensor 12 (to the sensor 12) and in a direction away from the sensor 12.

In addition, the high-pass filters 15a and 15b of the embodiment remove the low-frequency components of the Doppler signal, thereby greatly attenuating noise components due to raindrops and the like, and preventing the subsequent amplifiers 16a and 16b from being saturated. That is, the Doppler output frequency fd of the sensor 12 is calculated from fd = 2 · f ₀ (v / C ₀ ) · cos α [C ₀ : speed of light, f ₀ : from the velocity v of the moving object and the angle α of the moving object with respect to the sensor 12. As described above, when water flows parallel to the antenna surface 12A, the frequency component of the Doppler output is concentrated on the lower frequency side because the angle α is close to 90 degrees. . Therefore, in the embodiment, noise components such as raindrops are attenuated to some extent by removing low-frequency components with the high-pass filters 15a and 15b. The high-pass filters 15a and 15b can be band-pass filters to remove high-frequency noise components.

Next, each of the I and Q signal outputs from the amplifiers 16a and 16b is input to the AD converters 17a and 17b, and is sampled with a sampling time of, for example, 1/8 period or less of the Doppler center frequency (f _d0 ). Since the Doppler frequency of human movement can be limited to a certain range, the sampling time is preferably the center frequency of the Doppler frequency to be acquired, and an interval shorter than 1/8 period.

The data output from the A / D converters 17a and 17b is stored in the memories 20a and 20b in the microcomputer 18 for a predetermined time, and the integration unit 21 in the next stage starts Doppler from the current point of one of the IQ signals. The value of the product of the data (for example, I data) acquired before the time corresponding to ¼ period of the center frequency (f _d0 ) and the current data (for example, Q data) of the other signal is time-series. This data string (a plurality of product data) is stored in the memory.

  The short-time moving average unit 22 in the microcomputer 18 calculates an average value (short-time moving average value) of data from the current time to a short time of several milliseconds to several tens of milliseconds in the data string. Then, in the spike determination and removal unit 23 in the next stage, the data in which the number of times that the threshold range that is equal to or less than the predetermined positive threshold and is equal to or greater than the negative threshold is continuously equal to or less than the predetermined number is extracted as a spike This spike signal is removed.

  FIG. 3 shows various spike removal patterns of the embodiment. In the case of the positive spike in FIG. 3A, the positive threshold is +1, the negative threshold is -1, and the predetermined number of times is 10 times. Then, the signals of S1 and S2 in the figure are signals of level +2 if the original raw detection, but since the number of times exceeding the positive threshold continuously is less than the predetermined number, for example 0 It is replaced with (value within the threshold range). In the case of the negative spike shown in FIG. 3B, the signals S3 and S4 are level-2 values in the case of raw detection, but the number of times below the negative threshold (-1) is less than the predetermined number. Because there is, it is replaced with 0.

  The spike-removed signal (data string) is supplied to the second moving average unit 24, where the moving average calculation is performed again. For example, the average value of data from the current time to a certain time before is calculated at an interval longer than the short-time moving average interval, and the detection determination unit 25 uses the output from this moving average unit 24 to The moving body is finally determined.

  FIG. 5 shows a signal waveform obtained in the above main part. As shown in the waveform of FIG. 5A, the output from the Doppler sensor circuit unit 14 is a waveform of a moving body such as a person. Although noise such as raindrops is superimposed, short-time spikes are extracted as shown in FIG. 5B when moving average is performed for a short time after passing through the high-pass filters 15a and 15b. When the spike is removed and the moving average is performed again, a signal of a moving body such as a person is extracted as shown in FIG. 5C, and finally, the detection determination unit 25 receives the signal from FIG. The detection output of a moving body such as a person is obtained. Accordingly, in the embodiment, random noise having no correlation is attenuated by the high-pass filters 15a and 15b, and spikes generated at a relatively high level in a short time due to strong rain or the like are well removed, so that a moving object such as a human being can be obtained. Is reliably detected.

  In the embodiment, various spikes shown in FIGS. 3C and 4D to 4F are also removed. FIG. 3C is an example of positive and negative continuous spikes, and in the output of the short-time moving average unit 22, a continuous signal is output from a state exceeding the positive threshold to a state below the negative threshold. Thus, when each of the number of times exceeding the positive threshold and the number of times falling below the negative threshold is equal to or less than the predetermined number, this positive / negative continuous signal is determined as a spike. As a result, the signal of S5 in FIG. 3C is a level-2 value in the raw state, but is replaced with a value of 0 (or any other value within the threshold range). On the contrary, even when a continuous signal is output from a state below the negative threshold to a state above the positive threshold, if each signal is equal to or less than the predetermined number of times, it is determined as a spike and replacement with 0 is performed.

  FIG. 4D is an example of a signal including a break. After the output signal of the short-time moving average unit 22 becomes equal to or greater than a positive threshold, the break returns to the positive threshold or less (within the threshold range) due to the break. If the number of times that the threshold is exceeded from the beginning again is less than the predetermined number of times and the number of times of returning to the threshold range is less than or equal to the predetermined number of times when it becomes the positive threshold value or more again, it is determined that the detection has been interrupted. The signal at the time of disconnection of S6 in FIG. 4D is a level 0 value as raw data, but is replaced with the previous and next detected data value +2, and is interpolated. The same applies when the detection (waveform) is reversed. After the negative threshold value is reached, the value returns to the negative threshold value or more (within the threshold range) due to disconnection. When the number of times below the threshold is equal to or smaller than the predetermined number and the number of times of returning to the threshold range is equal to or smaller than the predetermined number, the data at this time is similarly interpolated with the preceding and subsequent detection data.

  FIG. 4E shows an example of a signal including a reverse spike. In the output signal of the short-time moving average unit 22, the signal becomes equal to or higher than the positive threshold and then becomes equal to or lower than the negative threshold. When the number of times that the threshold value was exceeded from the beginning was less than the predetermined number, and the number of times the negative threshold value was not exceeded (which is also less than the positive threshold value) was less than the predetermined number, the reverse spike was superimposed. The state is determined. The S7 signal in FIG. 4E is a level-2 value in the raw state, but is replaced with the preceding and following detected data value +2, and is interpolated. The same is true when the detection (waveform in the figure) is reversed, and the previous and subsequent detection data values are replaced.

  FIG. 4F is an example of a positive and negative continuous spike signal including a break. In this example, the signal breaks in S8, the signal immediately before is not less than the positive threshold value, and the immediately following signal is not more than the negative threshold value. Therefore, the 0 value, which is the intermediate value between the preceding and subsequent detection data, is interpolated as the value of the S8 signal.

  FIG. 6 shows another method of spike determination and removal. In this example, the spike determination and removal unit 23 replaces the median value of a plurality of previous signal data with the current data and spikes. Determine. In the embodiment, for example, the data to be used is 5 times (arbitrary number), and when signals S11 to S15 are obtained as shown in FIG. -> S12-> S15-> S14, the median value becomes the value of S12, so this S12 value is replaced and stored as the current data. That is, four times are within the threshold range, and one time exceeds the threshold value, and the signal S13 exceeding the threshold value is removed from detection as a spike.

  As shown in FIG. 6B, when the signals S21 to S25 are obtained, the order is S21 → S23 → S24 → S25 → S22, and the median is the value of S24. Store as current data. In this case, three times exceed the positive threshold value, and one time in the meantime is equal to or less than the negative threshold value. By using the median signal S24 as the current data, S22 that is below the negative threshold value is in the reverse direction. Are removed from detection as spikes.

  According to such replacement of the median value, a method of extracting the median value of a plurality of signals and replacing it with the current signal, determining whether the threshold value range is exceeded each time, and counting the number of consecutive times Compared to, the process becomes simple.

  By performing each of the above processes, in the embodiment, a short-time spike-like detection signal generated when raindrops or the like pass immediately before the sensor 12 is reliably removed, and a rain detection signal that is a random noise component Can be reduced by moving average, and even when it is raining, malfunctions can be greatly reduced and only human movement can be detected.

10 ... Window, 12 ... Doppler sensor,
14 ... sensor circuit part, 15a, 15b ... high-pass filter,
17a, 17b ... AD converter,
18 ... Microcomputer, 20a, 20b ... Memory,
21 ... Integration unit, 22 ... Short moving average unit,
23 ... Spike determination and removal unit,
24 ... 2nd moving average part 25 ... Detection determination part.

Claims (7)

In a microwave Doppler detection device comprising a Doppler sensor and a quadrature detection circuit that generates two I and Q signals having a phase difference of 90 degrees from the output of the Doppler sensor,
The I and Q signals output from the quadrature detection circuit are sampled at predetermined intervals, and either the I signal or the Q signal is shifted by approximately ¼ period at the center frequency of the Doppler signal and the other signal And a product / moving average processing circuit that calculates a moving average of the product,
The moving average calculation of this product / moving average processing circuit is performed in a short time, and the number of times that the value of this short-time moving average continuously exceeds the threshold range that is equal to or less than the predetermined positive threshold and equal to or greater than the negative threshold is equal to or less than the predetermined number of times. A microwave Doppler detection apparatus characterized in that a signal at the time is determined as a spike, and the spike signal is replaced with a signal having a value within the threshold range.   2. The microwave Doppler detection device according to claim 1, wherein the predetermined number of times for determining the spike is set between 100 and 500 milliseconds.   In the above short-time moving average calculation, a signal exceeding the positive threshold and a signal falling below the negative threshold or vice versa are continuously output, and the number of times the positive threshold is continuously exceeded and the negative threshold is continuously output. 3. The micro of claim 1, wherein when the number of times exceeds the predetermined number of times, the positive / negative continuous signal is determined as a spike, and the spike signal is replaced with a signal having a value within the threshold range. Wave Doppler detector. After the signal obtained by the above-mentioned short-time moving average calculation is continuously equal to or greater than the predetermined threshold number of times, the signal returns to the positive threshold value or less, and the positive threshold value is continuously equal to or less than the predetermined number of times. When it becomes below the positive threshold again after becoming below, the data below the above positive threshold is interpolated with the data above and below the positive threshold,
Alternatively, after the number of times is not more than the predetermined number of times, the value returns to the negative threshold value or more, and after the number of times equal to or less than the number of times has been continuously set to the value of the negative threshold value or less, the value is again less than the negative threshold value. 3. The microwave Doppler detection device according to claim 1, wherein when the value becomes equal to or greater than the negative threshold value, the data is interpolated by data before and after the positive threshold value.   5. The microwave Doppler detection device according to claim 1, wherein the moving average calculation is performed again on the detection signal after the spike signal is replaced with a signal having a value within the threshold range. In a microwave Doppler detection device comprising a Doppler sensor and a quadrature detection circuit that generates two I and Q signals having a phase difference of 90 degrees from the output of the Doppler sensor,
The I and Q signals output from the quadrature detection circuit are sampled at predetermined intervals, and either the I signal or the Q signal is shifted by approximately ¼ period at the center frequency of the Doppler signal and the other signal And a product / moving average processing circuit that calculates a moving average of the product,
The moving average calculation of this product / moving average processing circuit is performed in a short time, and the median value of a plurality of previous data obtained by this short-time moving average calculation is extracted, and this median value is used as the current data value. A microwave Doppler detection device characterized by replacing.   The microwave Doppler detection apparatus according to claim 6, wherein the moving average calculation is performed again on the detection signal after the median replacement.

Images