JP2019143978A - Object detection device - Google Patents

Abstract

To provide an object detection device capable of estimating a Doppler shift frequency without increasing the circuit scale and with a relatively small amount of calculation.SOLUTION: A Doppler shift frequency estimation unit 74 is provided in a distance measuring circuit that measures a distance to an object on the basis of a reception signal obtained from a reflected wave and a Doppler shift frequency of the reflected wave, generates a plurality of sample data by complex multiplication of the reception signal and a reference, performs extension processing for extending the number of data by performing zero padding on the sample data, and thinning processing for thinning out the expanded number of sample data, generates sample data with a predetermined frequency resolution, performs FFT processing on the sample data after thinning, and estimates the Doppler shift frequency.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an object detection device that detects a distance to an object based on a reflected wave from a relatively moving object.

  2. Description of the Related Art Conventionally, an object detection apparatus that transmits ultrasonic waves, receives the reflected waves, and measures the distance to the reflecting object is known. For example, as shown in FIG. 8, the object detection apparatus 10 includes a sensor 11 that transmits and receives ultrasonic waves, and a distance measurement circuit 12. For example, the sensor 11 transmits an ultrasonic wave having a frequency of several tens of kHz as a transmission wave Tx in a burst shape with a length of several to several tens of waves, and receives a reflected wave Rx from the object W. That is, the transmission wave Tx and the reflected wave Rx have the same frequency. The distance measuring circuit 12 detects the relative delay amount between the transmitted wave Tx and the reflected wave Rx by performing synchronous quadrature detection on the reflected wave Rx, and the distance D from the delay amount and the sound velocity to the object W. Is calculated.

  When such an object detection device 10 is mounted on a moving body such as an automobile, the object W is detected while moving, so that the frequency of the reflected wave Rx is the same as that of the transmission wave Tx due to the Doppler effect due to the relative velocity with the object W. Will be different. For example, when an object detection device 10 mounted on an automobile detects an obstacle when entering a garage, the object W that is an obstacle is stopped, but the object detection device 1 is moving, so the Doppler effect Therefore, a difference occurs between the transmitted wave Tx and the reflected wave Rx.

For example, when an automobile equipped with the object detection device 10 is approaching the object W at a speed of 11.8 km / h (3.29 m / s), the frequency (Doppler frequency) f ′ of the reflected wave Rx changed by the Doppler effect is , Can be calculated from Equation 1 below. Note that f is the frequency of the transmission wave Tx, V is the speed of sound in the atmosphere (340 m / s), v _o is the moving speed of the observer, and v _s is the moving speed of the sound source. Further, in the object detection apparatus 10, since the sound source and the observer are the same and the reflected wave Rx from the object W is observed, the sign of the moving speed _vo on the observer side is reversed.

  For example, when the frequency f of the transmission wave Tx is 50 kHz, the Doppler frequency f ′ is 50.777 kHz, and is subjected to a frequency shift of +977 Hz compared to the original frequency 50 kHz. That is, in the above Doppler environment, when the transmission frequency from the sensor 11 is 50 kHz, when the reflected ultrasonic wave is received by the sensor 11, an electric signal of 50.777 kHz is obtained. Hereinafter, this frequency shift is referred to as a Doppler shift frequency Δf.

  9 (a) to 9 (p) show a transmission wave, a reception wave, or a reference signal similar thereto after performing orthogonal detection at various frequencies on the reflected wave Rx obtained in the experiment under the above Doppler environment. And a peak waveform used for distance measurement obtained by this is shown. The rectangular frame in each figure shows the frequency used for the quadrature detection, and “DC” in the figure (a) is a case where the reception frequency is 50 kHz and the Doppler shift is not generated with respect to the transmission frequency 50 kHz. Represents the output of the correlation processing unit. Using such a waveform, subtracting the transmission time of the transmission wave and the processing delay of the circuit from the time of the peak of the peak waveform where the amplitude is maximum changes according to the distance D between the sensor 11 and the object W. Delay time can be measured.

  However, under the Doppler environment as described above, when the Doppler shift frequency Δf is unknown and orthogonal detection is performed at the original 50 kHz, the output obtained by the correlation processing has a waveform as shown in FIG. Since a peak waveform effective for distance measurement cannot be obtained, the quality of distance measurement deteriorates. That is, if a correct Doppler shift frequency Δf (for example, +977 Hz) can be estimated by any method, quadrature detection is performed at a frequency of “transmission frequency f (50 kHz) + Doppler shift frequency Δf (977 Hz)” or orthogonal at the transmission frequency f. After detection, by performing signal processing to shift the frequency by the Doppler shift frequency Δf, and performing correlation processing, as shown in the waveform orthogonally detected at +977 Hz in FIG. Can be obtained.

  2. Description of the Related Art Conventionally, a digital pulse compression device used in a pulse compression radar is provided with a plurality of Doppler correction circuits having different correction amounts with respect to a Doppler shift in order to correspond to an input signal whose Doppler shift frequency Δf is unknown. After the Doppler correction of the input signal by these Doppler correction circuits, each output signal is subjected to pulse compression, and the one having the maximum compression amplitude is detected and output every time such as a range bin (for example, patents) Reference 1). According to this, even when the Doppler shift frequency is unknown, stable compression performance can be obtained.

  Further, using a plurality of correlators, a correlation is obtained between the received signal and a plurality of correlation input signals reflecting a relative speed (for example, ± 4 km / h) at a predetermined interval, and a plurality of correlation outputs obtained are obtained. There is also known an apparatus that removes the influence of Doppler shift by integrating signals and performing distance measurement by selecting a correlation output signal from which a peak waveform of a target echo is obtained from the integration result (for example, Patent Document 2). reference).

JP 05-066268 A JP-T-2017-508133

  However, when a plurality of Doppler correction circuits are arranged in parallel in the distance measuring circuit as in Patent Document 1, the circuit scale is increased, which hinders downsizing of the object detection apparatus. Further, since the output signals are compared and output after correction by each Doppler correction circuit, there is a possibility that the amount of calculation increases and the processing time is delayed. In addition, in the configuration using a plurality of correlators as in Patent Document 2, since all correlators must ensure the amount of calculation necessary to ensure the accuracy required for ranging, the entire circuit scale is required. Or the amount of calculation becomes large. Furthermore, when the device of Patent Document 2 is realized by an electric circuit, a plurality of correlation outputs must be stored in a memory until a single correlator is selected, so a memory having a relatively large capacity is required. It becomes. However, since the memory occupies a large proportion of the chip area, it becomes an obstacle to cost reduction by reducing the chip area when making an IC for mass production.

  Therefore, an object of the present invention is to provide an object detection apparatus capable of estimating the Doppler shift frequency without increasing the circuit scale and with a relatively small amount of calculation.

  In order to solve the above-mentioned problem, the invention described in claim 1 is characterized in that a transmission wave is transmitted based on a transmission signal and a reflected wave from a relatively moving object is received, and a reception obtained from the reflected wave. A distance measuring means for measuring a distance to the object based on a signal and a Doppler shift frequency of the reflected wave, wherein the object detection device includes a complex multiplication of the received signal and a predetermined reference signal. A plurality of sample data is generated, an extension process for expanding the number of sample data to the sample data, and a thinning process for thinning out the expanded number of sample data are performed to generate sample data with a predetermined frequency resolution. Doppler shift frequency estimation means for performing fast Fourier transform on the sample data to estimate the Doppler shift frequency, and the estimated Doppler Based on the shift frequency, the reflected wave is received by a Doppler shift canceling means for canceling a Doppler shift included in the received signal, a correlation operation between the received signal from which the Doppler shift has been canceled, and a predetermined reference signal. Correlation processing means for extracting timing, and the distance measuring means calculates a distance to the object based on the correlation calculation result.

  The object detection apparatus according to claim 2, further comprising holding means for holding the received signal for a time required for estimating the Doppler shift frequency and outputting the held signal to the Doppler shift frequency estimating means.

  According to the first aspect of the present invention, a plurality of sample data is generated by complex multiplication of the received signal and a predetermined reference signal, and an extension process for extending the number of sample data is performed on the sample data. It is possible to generate sample data having a higher frequency resolution than the sample data. In addition, since the expanded sample data is subjected to thinning processing and then subjected to fast Fourier transform to estimate the Doppler shift frequency, the amount of calculation can be reduced by the thinning-out amount. As a result, the Doppler shift frequency can be estimated without increasing the circuit scale of the distance measuring means and with a relatively small amount of calculation. Further, a reflected wave is received by a correlation operation between the Doppler shift canceling means for canceling the Doppler shift using the estimated Doppler shift frequency, the received signal, and a predetermined reference signal after the Doppler shift frequency estimating means. By providing the correlation processing means for extracting the timing, the correlation processing without deterioration due to thinning can be performed again, so that the object distance calculation accuracy is improved.

  Furthermore, since the correlator for distance measurement and the Doppler shift frequency estimation means are provided separately, the amount of FFT computation for performing the fast Fourier transform is reduced to the minimum amount of computation required for estimating the Doppler shift frequency. In addition, the number of correlators that ensure distance measurement accuracy can be reduced to one. Therefore, when the apparatus of the present invention is realized by an electric circuit, the entire circuit scale (manufacturing cost) can be reduced.

  According to the second aspect of the present invention, since the holding unit that holds the received signal for a time necessary for the estimation of the Doppler shift frequency and outputs the received signal to the Doppler shift frequency estimation unit is provided, Processing time for executing the correlation processing after canceling the frequency shift using the estimated Doppler shift frequency value for the received signal can be obtained. Further, only the storage capacity of one correlator is required for the purpose of holding the received signal until the Doppler shift estimation is completed. Therefore, when the apparatus of the present invention is realized by an electric circuit, only that much is required. The chip area can be reduced, and the cost can be easily reduced.

It is a block diagram which shows the structure of the object detection apparatus which concerns on embodiment of this invention. It is explanatory drawing which shows the processing content of the Doppler shift frequency estimation part of the object detection apparatus of FIG. It is a table | surface which shows operation | movement of the shift register of the object detection apparatus of FIG. It is a table | surface which shows the content of the reference utilized in the Doppler shift frequency estimation part of FIG. It is a table | surface which shows the content of the expansion process of a sample number by the Doppler shift frequency estimation part of FIG. 2, and the thinning-out process. It is a table | surface which shows the frequency allocated to FFT in the Doppler shift frequency estimation part of FIG. It is explanatory drawing which shows the processing content of the correlation process part of the object detection apparatus of FIG. It is a block diagram which shows schematic structure of the conventional object detection apparatus. It is a graph which shows the waveform obtained by performing the orthogonal detection and the correlation process of a reference signal with various frequencies with respect to the reflected wave obtained in the Doppler environment.

  The present invention will be described below based on the illustrated embodiments.

  1 to 7 show an embodiment of the present invention, and FIG. 1 is a block diagram showing a configuration of a distance measuring circuit of an object detection apparatus 1 according to this embodiment. The object detection device 1 includes a sensor (transmission / reception means) 2, a clock generation circuit 3, a sensor drive circuit 4, an analog reception circuit 5, an AD conversion circuit 6, and a digital signal processing circuit 7. The sensor 2 transmits an ultrasonic wave for object detection, receives the ultrasonic wave reflected from the object, and converts it into an electric signal. The clock generation circuit 3, the sensor drive circuit 4, the analog reception circuit 5, the AD conversion circuit 6, and the digital signal processing circuit 7 constitute a distance measurement circuit (range measurement means).

  The clock generation circuit 3 supplies a timing signal to each transmission / reception circuit in order to maintain synchronization between transmission and reception. The sensor driving circuit 4 converts the pulse signal from the digital signal processing circuit 7 into an appropriate voltage so that ultrasonic waves for object detection are transmitted from the sensor 2. The analog reception circuit 5 amplifies the reflected wave to a signal level suitable for converting the reflected wave into a digital signal by the AD conversion circuit 6. At that time, unnecessary waves and noise are suppressed by a filter or the like in order to improve the signal-to-noise ratio of the signal. The AD conversion circuit 6 converts the reflected wave, which is an analog signal, into a digital reception signal.

  The digital signal processing circuit 7 performs signal processing on the reflected wave converted into the digital signal, and extracts information such as the presence / absence of an object, the distance between the sensor 2 and the object, and a delay time that changes according to the distance. The digital signal processing circuit 7 includes a transmission signal generation unit 71, a quadrature detection unit 72, a shift register (reception signal holding unit) 73, a Doppler shift frequency estimation unit (Doppler shift frequency estimation unit) 74, and a frequency shift unit ( A Doppler shift canceling unit) 75, a correlation processing unit (correlation processing unit) 76, a threshold generation unit 77, and a distance estimation unit 78 are provided.

  The transmission signal generation unit 71 outputs a transmission signal for detecting an object to the sensor driving circuit 4. Usually, the transmission signal is a burst signal using a rectangular wave, and the length of the transmission time varies depending on the distance to be measured, but is appropriately set so that the transmission wave and the reflected wave do not overlap. The set value may be held as a constant at the time of shipment from the factory, or may be set in a timely manner by host control such as an ECU (Engine Control Unit) when mounted in an automobile or the like.

  The quadrature detection unit 72 performs quadrature detection at a frequency synchronized with the transmission signal, and converts the received signal into a baseband signal. However, when the reflected wave has undergone Doppler shift, the frequency is shifted from the transmission frequency by the Doppler shift frequency Δf, so the Doppler shift frequency estimation value described later is received, and Doppler shift cancellation is performed simultaneously with quadrature detection. May be. In that case, even if the received signal has undergone Doppler shift, the output after quadrature detection is a baseband signal without Doppler shift.

  The shift register 73 holds and delays the received signal by an appropriate amount. The amount of delay is determined by the time required for calculating the Doppler shift frequency estimated value. By holding and delaying the received signal by the shift register 73, the subsequent correlation processing unit 76 cancels the frequency shift using the Doppler shift frequency estimated value calculated in advance for the received signal used for calculating the Doppler shifted frequency estimated value. It is possible to execute correlation processing after that.

  The Doppler shift frequency estimation unit 74 estimates and holds the Doppler shift frequency Δf. Details of the operation of the Doppler shift frequency estimation unit 74 will be described later. The frequency shift unit 75 cancels the Doppler shift included in the quadrature detection output signal based on the estimated Doppler shift frequency Δf. The correlation processing unit 76 extracts the timing at which the reflected wave is received by continuing to correlate the transmission signal, the reception signal, or a reference signal similar to them (predetermined reference signal) and the reception signal.

  The threshold value generation unit 77 generates a threshold value for determining the presence or absence of an object from the received signal and correlation output used for the correlation processing. If the correlation value is less than or equal to the threshold value, it is determined that it has not been detected. The distance estimation unit 78 calculates the time from when an ultrasonic wave is emitted until it is reflected and returned based on the peak timing when the correlation value exceeds the threshold, the transmission time, and the known circuit delay. To do. Further, the distance from the sensor 2 to the object is calculated by dividing 1/2 of the calculated time by the speed of sound. Information such as the calculated time and distance is transmitted to an upper control unit such as an ECU via an interface circuit (not shown).

  FIG. 2 shows an operation explanatory diagram of the Doppler shift frequency estimation unit 74. The received signal is a complex number expressed by a fixed-point discrete value (for example, 16 bits), and is input to the shift register 73 at a rate of one sample per ultrasonic wave. For example, when the frequency of the transmitted ultrasonic wave is 50 kHz, the sample rate is 50 ksps and the sample interval is 20 μs. Here, a case where a burst signal having a length of 128 waves using a 32-bit 4-bit Barker code (-1, -1, 1, -1) per bit will be described as an example.

  The length of the shift register 73 coincides with the length of the transmitted ultrasonic burst signal. In this example, it is 128 stages long and is composed of two 16-bit registers per stage because it is a complex number. In addition, when using FPGA, ASIC, etc., you may comprise by the ring buffer etc. which used RAM. The operation of the shift register 73 is performed every time a reception signal is input, as shown in FIG. That is, the shift register 73 holds the input received signal by sequentially substituting the received signal with 128-stage registers S1reg [0] to S1reg [127].

  The Doppler shift frequency estimation unit 74 performs complex multiplication of the contents of the shift register 73 and the contents of the reference (reference signal) by the same index number. A waveform similar to the transmission signal (ref [0], ref [1],... Ref [127]) is used for the reference. Here, the code pattern output from the transmission signal generator 71 during transmission is used as a reference. If a 4-bit Barker code (−1, −1, 1, −1) of 32 waves per bit is used as a transmission signal, the content of the reference is a 4-bit Barker code of 32 waves per bit. Is the value shown in the table of FIG.

  The result of complex multiplication of the contents of the shift register 73 and the contents of the reference by the same index numbers is a vector having 128 elements (S1reg [0] × ref [0], S1reg [1] × ref [1 ], S1reg [2] × ref [2], /... S1reg [127] × ref [127]).

  In this example, the content of the reference is a real number and is ± 1, so in fact, instead of complex multiplication, the content of the shift register 73 is unchanged if the corresponding reference is 1, and the sign is inverted if it is -1. It is only Since each element of the vector is the product of the received signal and the reference time series, the modulation by the 4-bit Barker code is performed at the timing when the transmission signal reflected back into the shift register 73 is properly accommodated. Canceled. Since this vector is input to an FFT (Fast Fourier Transform unit) at the subsequent stage, the transmission signal reflected back into the shift register 73 is appropriately accommodated at some frequency output of the FFT output. A peak waveform is formed. Note that the frequency output at which the frequency at which the peak waveform is formed is determined by the frequency component included in the received signal.

  When a discrete time waveform vector (data number M, sample rate R) is input, FFT becomes a discrete frequency component vector (frequency resolution: R / M, frequency range: number of elements × R / M). In order to realize the desired frequency resolution and frequency range after FFT, 0 is added to the input data to increase the number M of data (thinning), or the number of data is reduced, and the vector elements It is possible to reduce the amount of computation by reducing the number of. As shown in the table of FIG. 5, the Doppler shift frequency estimation unit 74 performs the extension of the number of samples by zero padding to the input data and the thinning out of the number of samples. Here, the initial values of the variables are M = 128 data, R = 50 kHz, 384 samples are added after 128 samples, the number of samples is 512, and 512 samples are thinned out to 32 samples. Change the frequency range from 50 kHz to 3.125 kHz. Although the SNR deteriorates by 12 dB due to the aliasing, the calculation amount can be reduced by 96.5%.

  By operating the Doppler shift frequency estimation unit 74 as described above, the number of FFT elements is 32 and the frequency resolution is 97.7 Hz. The frequency components assigned to 32 FFT outputs (denoted as 32 FFT in FIG. 2) are shown in the table of FIG. When thinning 512 samples to 16 samples, first every 4 samples are thinned to 128 samples, and then every 4 samples are averaged to 16 samples to reduce SNR degradation due to aliasing. Also good.

  The FFT operation is repeated every time one sample of the received signal is input. Each time, 32 sets of FFT output vectors are output from the FFT. The Doppler shift frequency estimation unit 74 performs maximum value selection for selecting the one with the maximum amplitude (fftmax) from the 32 sets of FFT outputs. Then, two vectors (fftmax, fftnum) are output for each sample, each element having a pair of fftnum indicating the place where the maximum value is output.

  Further, the Doppler shift frequency estimation unit 74 compares the output maximum values and performs a max hold for storing the champion values (fftmax_max, fftnum). When a fftmax value having an amplitude larger than the stored champion value appears, the stored champion value is updated. fftmax becomes maximum at the synchronization point, and is maintained without being updated thereafter. At this time, the frequency component indicated by fftnum held together is output as an estimated value of the Doppler shift frequency Δf. Note that the champion value (fftmax_max, fftnum) that has been held is cleared when the measurement is completed. Further, during the measurement, for example, if it is not updated once for 256 samples, a plurality of objects having different relative velocities may be detected by resetting the max hold or the like.

  As shown in FIG. 7, the correlation processing unit 76 is an FIR filter (finite impulse response filter) having a reference as a tap coefficient, and a transmission waveform, a reception waveform, or a reference signal similar to them is used as the tap coefficient of the FIR filter. (Reference: ref [0], ref [1],... Ref [127]) are prepared, and correlation outputs with the received signals stored in the registers S1reg [0] to S1reg [127] of the shift register 76a are prepared. Ask for.

  The threshold generation unit 77 generates the threshold described above from the received signal used for the correlation processing and the correlation value of the correlation processing unit 76. The distance estimation unit 78 compares the correlation value with the threshold value, and determines that the object has not been detected when the correlation value is equal to or less than the threshold value. On the contrary, when the correlation value is larger than the threshold value, the distance estimation unit 78 calculates the peak timing when the correlation value exceeds the threshold value, the transmission time, and the known circuit delay. Based on the above, the time from when the ultrasonic wave is emitted until it is reflected and returned is calculated, and further, the distance from the sensor 2 to the object is calculated by dividing ½ of the calculated time by the speed of sound.

  As described above, according to the object detection device 1 of this embodiment, the Doppler shift frequency estimation unit 74 uses the received signal (S1reg [0] to S1reg [127]) obtained from the reflected wave and the reference ( A plurality of sample data (S1reg [0] × ref [0], S1reg [1] × ref [1], S1reg [2] × by complex multiplication with a predetermined reference signal: ref [0] to ref [127]) ref [2],... S1reg [127] × ref [127]) and the expansion processing for expanding the number of sample data is performed on the sample data, so that the frequency resolution is higher than the sample data generated first. High sample data can be generated. In addition, since the expanded sample data is subjected to thinning processing and then subjected to fast Fourier transform to estimate the Doppler shift frequency, the amount of calculation can be reduced by the thinning-out amount. As a result, the Doppler shift frequency can be estimated without increasing the circuit scale of the distance measuring means and with a relatively small amount of calculation.

  Although the embodiment of the present invention has been described above, the specific configuration is not limited to the above embodiment, and even if there is a design change or the like without departing from the gist of the present invention, Included in the invention.

DESCRIPTION OF SYMBOLS 1 Object detection apparatus 2 Sensor (transmission / reception means)
DESCRIPTION OF SYMBOLS 3 Clock generation circuit 4 Sensor drive circuit 5 Analog reception circuit 6 AD conversion circuit 7 Digital signal processing circuit 71 Transmission signal generation part 72 Quadrature detection part 73 Shift register (reception signal holding means)
74 Doppler shift frequency estimation section (Doppler shift frequency estimation means)
75 Frequency shift unit (Doppler shift canceling means)
76 Correlation processing unit (correlation processing means)
77 Threshold generation unit 78 Distance estimation unit

Claims (2)

Based on the transmission / reception means for transmitting the transmission wave based on the transmission signal and receiving the reflected wave from the relatively moving object, the received signal obtained from the reflected wave, and the Doppler shift frequency of the reflected wave, the object A distance measuring means for measuring the distance to the object detecting device,
A plurality of sample data is generated by complex multiplication of the received signal and a predetermined reference signal, an expansion process for expanding the number of sample data to the sample data, and a thinning process for thinning out the expanded sample data number are performed, Generating Doppler shift frequency estimating means for generating sample data of frequency resolution and performing fast Fourier transform on the sample data after thinning to estimate the Doppler shift frequency;
Doppler shift canceling means for canceling the Doppler shift included in the received signal based on the estimated Doppler shift frequency;
Correlation processing means for extracting a timing at which the reflected wave is received by a correlation calculation between the received signal in which the Doppler shift is canceled and a predetermined reference signal;
The distance measuring means calculates a distance to the object based on the correlation calculation result.
An object detection device comprising: Receiving signal holding means for holding the received signal for a time required for estimation of the Doppler shift frequency and outputting it to the Doppler shift frequency estimating means,
The object detection apparatus according to claim 1.

Images