JP5704552B2 - Radar equipment - Google Patents

Description

  The present invention relates to a technique for measuring the distance and speed to a target in a measuring device using electromagnetic waves or sound waves.

  Radar is one of the applications of measurement technology using electromagnetic waves.

  In recent years, research on collision prevention technology using millimeter-wave vehicle-mounted radar as a sensor in an intelligent road information system ITS (Intelligent Transport Systems) is one of the applications of radar technology. It is expected that collision avoidance and collision damage reduction will be realized by detecting danger at an early stage with millimeter-wave radar and controlling the car safely.

  Patent Document 1 describes an example of a technique applied to a conventional radar device.

Japanese Patent No. 3699450

  In order to accurately predict the movement of other vehicles, it is necessary to measure distance, angle, and speed correctly, while millimeter-wave automotive radar is low cost (ie, narrow receiver bandwidth, low signal processing rate) High range resolution is required. Currently, many millimeter-wave on-vehicle radars employ an FMCW (Frequency Modulated Continuous Wave) system that provides high range resolution with relatively low-speed signal processing. However, in the FMCW system, a pairing malfunction of the detection frequency of the up sweep and the down sweep occurs at the time of multiple targets, and it is expected that it becomes a problem when the millimeter-wave vehicle-mounted radar becomes widespread. On the other hand, the pulse compression method requires high-speed correlation processing, and in order to realize high distance resolution, a wideband receiving system, a high-speed analog / digital (A / D) converter, and signal processing are required. Currently available A / D converters have limitations on conversion speed, and cost reduction is indispensable for widespread use of millimeter wave on-vehicle radars. Therefore, a new radar system that solves these problems is demanded.

  The present invention has been made in view of these points, and an object thereof is to enable highly accurate radar detection at low cost.

The present invention is a pulse in the frequency spread transmitted pulse, N (N is an integer of 2 or more) is transmitted by step the frequency using the frequency of Rutotomoni, the frequency step within one observation time M When the target is detected by receiving the reflection from the target of the transmission pulse repeatedly (M is an integer of 2 or more), and by combining after the autocorrelation of the two complementary codes , A signal using two complementary codes capable of suppressing the distance side lobe is assigned as a set of transmission pulses transmitted continuously at the same frequency within a fixed time, and a transmission pulse by a signal using the two complementary codes Is frequency stepped.
For a set of transmission pulses transmitted continuously at the same frequency within a fixed time, the pulse repetition time for ensuring the time until the transmission pulse is reflected from the target and received is used for each complementary code. The pulse repetition time is obtained by the following formula (where T _PRI is the pulse repetition time, R _max is the maximum instrument distance required for the radar, and c is the speed of light).
T _PRI ≧ (2R _max / c)
The radar apparatus in which the value of M corresponding to the number of transmission pulses of the same code at the same frequency within one observation time is smaller than when no set of transmission pulses is assigned. The received pulse is subjected to pulse compression processing and pulse Doppler processing, and the phase modulation pulses of two complementary codes are shifted by the pulse repetition time with respect to the signals subjected to the pulse compression processing and pulse Doppler processing. Correction processing for correcting a phase error due to the above and a distance error generated when there is a relative speed with the target is performed, and band synthesis is performed after addition.

According to the present invention , two complementary signals that can suppress components other than the pulse signal due to the reflected wave from the target are assigned as transmission pulses that are continuous at the same frequency by combining two complementary functions. The target separation can be improved.
Therefore, according to the present invention , the radar apparatus has the effect of improving the target object detection accuracy and the resolution at the time of multiple targets.

It is explanatory drawing which shows a pulse radar ranging principle. It is explanatory drawing which shows the distance resolution of a pulse radar. It is a block diagram which shows the example of a pulse Doppler radar signal process. It is explanatory drawing which shows a pulse Doppler radar speed measurement principle. It is explanatory drawing which shows a pulse Doppler radar and a pulse compression radar. It is explanatory drawing which shows the kind of modulated wave used with a pulse compression radar. It is a block diagram which shows the example of a pulse compression radar signal process. It is a wave form diagram which shows the example of an LFM modulated wave. It is a wave form diagram which shows the example of a pulse compression process reference function. It is a wave form diagram which shows the example of a pulse compression output waveform. It is a wave form diagram which shows the example of a pulse compression output waveform. It is a wave form diagram which shows the example of a pulse compression output waveform after weight application. It is a comparison figure of a pulse compression radar and SSW. It is a transmission frequency sequence diagram of SSW. It is a block diagram which shows the signal processing example of SSW. It is a characteristic view which shows the example of a pulse compression result. It is explanatory drawing which shows a spectrum shift process state. It is explanatory drawing which shows the synthetic | combination zone | band equivalent to a transmission occupation bandwidth. It is explanatory drawing which shows the example of a range profile. It is explanatory drawing which shows the example of a range profile. It is a conceptual diagram of the nonlinear frequency step of NL-SWW. It is a transmission frequency sequence diagram of NL-SWW. It is explanatory drawing which shows the determination process of the cubic function by a nonlinear parameter. It is a transmission frequency sequence diagram of multi-frequency NL-SWW. It is explanatory drawing which shows the example of the conditions which can use a transmission frequency sequence. It is a block diagram which shows the process example of multifrequency NL-SWW. It is a characteristic view which shows the example of an output after Fourier-transforming the m direction sampling signal with respect to n. It is a characteristic view which shows the example of a nonlinear frequency step space | interval. It is explanatory drawing which shows (sigma) dependence of PSL. It is a characteristic view which shows the example of an output after an inverse Fourier transform. It is a characteristic view which shows the example of an output after an inverse Fourier transform. It is a transmission / reception sequence diagram of FM pulse compression radar. It is a wave form diagram which shows the example of the received wave in FM pulse compression. It is a wave form diagram which shows the example of the reference function in FM pulse compression. It is a wave form diagram which shows the example of an output of FM pulse compression. It is a characteristic view which shows the derivation | leading-out example of a complementary signal. It is a wave form diagram which shows the example after the pulse compression process of the phase modulation wave by a complementary code | symbol. It is a wave form diagram which shows the example of the addition waveform of a complementary signal. It is a transmission / reception sequence diagram of LFS. It is a transmission-and-reception sequence diagram of 2 frequency CW radar. It is a transmission / reception sequence diagram of a multi-frequency step ICW radar. It is a block diagram of the example of a process by a multifrequency step ICW system. It is explanatory drawing which shows the example of isolation | separation by a frequency axis. It is explanatory drawing which shows the example of isolation | separation (case 1) by a time-axis. It is explanatory drawing which shows the example of isolation | separation (time 2) by a time-axis. It is a transmission-and-reception sequence diagram by a complementary signal band synthesis method. It is a block diagram which shows the process example until it obtains a received wave pulse. It is a block diagram which shows the signal processing example after the pulse compression process by a hybrid CFS method. It is a block diagram which shows the example in the case of replacing IFFT with MUSIC in complementary code | symbol + SWW. It is a block diagram which shows the example in the case of detecting a speed component with the structure of FIG. (A) shows the detection state of the signal of Code1, (b) shows the detection state of the signal of Code2, and (c) is a waveform diagram showing the detection state of the signal of Code1 and the signal of Code2. It is a wave form diagram which showed the velocity component detected from the signal of Code1. It is the wave form diagram which showed the velocity component detected from the signal of Code1 and the signal of Code2. It is a conceptual diagram of the nonlinear frequency step of NL-SWW. It is the block diagram which showed the signal processing structural example of NL-SWW. It is explanatory drawing which showed the example of the constraint conditions with the waveform. It is explanatory drawing which showed the example of the constraint conditions. It is explanatory drawing which showed the band synthesis process output example of NL-SWW. It is explanatory drawing which showed the example of the distance waveform in NL-SWW.

Embodiments of the present invention will be described in the following order.
1. 1. Explanation of radar detection technology as a premise of the first embodiment 2. Description of the first embodiment 3. Explanation of radar detection technology as a premise of the second embodiment 4. Description of the second embodiment Modification of each embodiment

1. Description of Radar Detection Technology as a Premise of the First Embodiment First, the radar detection technology as a premise of the first embodiment will be described in order.
1.1 Pulse radar One of the radar ranging methods is pulse radar. This is a distance measurement method in which a short pulse corresponding to a desired distance resolution is transmitted and a target distance is calculated by measuring a time delay of a received pulse with respect to a transmission timing.
Hereinafter, the signal processing configuration and various principles of the pulse radar will be described.
1.2 Ranging Principle of Pulse Radar Since radio waves are attenuated when propagating in space, it is desirable to use pulses with high power, that is, high peak power, for measuring distances to long distances. Therefore, the pulse radar transmits a pulse having a large peak power. As shown in FIG. 1, a pulse is transmitted, and a time delay τ (round trip time) of the received pulse with respect to the transmission timing is measured. From this, the speed at which the radio wave travels in the space is the speed of light c, so that the target distance R can be calculated as shown in [Expression 1].

In addition, in order to prevent distance ambiguity (one or more before, one or more after, or the reflected wave of both transmission pulses entering), the radio wave reciprocates from one transmission pulse to the next transmission pulse. From this condition, a pulse repetition interval PRI (Pulse Repetition Interval) is determined, and the maximum instrument distance depends on this. Alternatively, the PRI is determined from the maximum instrument distance required by the radar.

1.3 Distance Resolution of Pulse Radar The distance resolution δR of the pulse radar that measures the distance by the time delay τ depends on the pulse width Tw (see FIG. 2) of the transmission pulse, as shown in [Formula 3]. Therefore, it is desirable for pulse radar to transmit a pulse having a large peak power and a narrow pulse width.

  Further, in order to receive the transmitted pulse, the receiver is required to have a band B corresponding to the reciprocal of the pulse width Tw. At this time, the distance resolution δR of the pulse radar can be expressed as follows, and in order to obtain a high distance resolution by the pulse radar, a wide band is required from the equation [4].

1.4 Signal Processing Configuration of Pulse Doppler Radar FIG. 3 is a signal processing configuration example of the pulse Doppler radar device. That is, the millimeter wave transmission wave generated by the transmission wave generation unit 11 is pulsed by the pulse processing unit 12 and transmitted by the transmission antenna 13. The signal transmitted and reflected by the target is received by the receiving antenna 14 and phase-detected by the phase detector 15. As a configuration for phase detection, the transmission wave is mixed by the mixer 16 and the mixed signal is detected by extracting only the low frequency by the low-pass filter 17. The signal detected by the phase detector 15 is digitally converted by an analog / digital converter 18, Fourier transformed by a plurality of stages of Fourier transforms 19 a, 19 b, 19 c... It supplies to the speed calculation part 20, and calculates the distance and speed to a target object.

  The detection by the phase detection unit 15 is a demodulation process, which is to remove a carrier wave component from a received wave and to extract a baseband signal. Thereafter, the analog / digital converter 18 converts the analog signal into a digital signal. The signal is sampled every distance bin (a constant time width such as a pulse width). As a result, digital signal processing such as Fourier transform described later can be performed. A radar that uses a ranging method similar to that of a pulse radar and detects a target relative velocity by Fourier transform is called a pulse Doppler radar.

1.5 Speed measurement principle of pulse Doppler radar When the target is moving, radio waves are affected by the Doppler effect as well as sound waves. At this time, the Doppler frequency fd is as follows.

v: Relative speed (positive in the direction away from the radar)
fRF: frequency of the radio wave radiated from the antenna

  In the pulse Doppler radar, signals with the same time delay are reflected waves from the same target, so the same time delay sampling data (distance bins) are collected as shown in FIG. An existing target Doppler frequency fd is obtained. As shown in [Expression 6], the target relative speed V can be calculated from the obtained Doppler frequency fd. At this time, the velocity resolution depends on the observation time Tc, as shown in [Formula 6].

1.6 Pulse compression radar In order to observe a long-distance target with high range resolution by pulse radar, a narrow pulse with a large peak power must be transmitted as shown in Fig. 5 (a). Don't be. However, if a high output cannot be obtained due to hardware restrictions, as shown in FIG. 5B, a transmission pulse having a long pulse width is modulated and transmitted with the peak power kept low, and then the target power is reduced. By performing autocorrelation processing when a reflected wave is received, the same effect as when a narrow pulse having a large peak power is transmitted can be obtained. This signal processing is called pulse compression processing. As shown in FIG. 6, the transmission pulse may be subjected to FM (Frequency Modulation) modulation or PM (Phase Modulation) modulation.

Hereinafter, the signal processing configuration of the pulse compression radar and the pulse compression radar of the LFM modulated wave will be described.
FIG. 7 is a diagram illustrating a signal processing configuration example of the pulse compression radar apparatus.
The difference from the pulse Doppler radar in FIG. 3 is the modulation process and the pulse compression process. Since the signal after the pulse compression processing is compressed into a pulse shape, the same processing as the pulse Doppler radar is performed thereafter. That is, the output of the analog / digital converter 18 is subjected to pulse compression by the pulse compression processing unit 21, and then subjected to Fourier transform by the Fourier transform units 19a, 19b, 19c,. At the time of pulse compression in the pulse compression processing unit 21, a reference function H (t) is obtained from the modulation pulse generation unit 11.

1.7 Pulse Compression Processing of LFM Modulated Wave A modulated wave (chirp wave) is a waveform whose frequency linearly changes with time t as shown in FIG. 8 and can be expressed as follows.

  Here, f is the carrier frequency (center frequency), and μ is the rate of change of the instantaneous frequency of s (t) with respect to time.

Assuming that the duration of s (t) is T, the variation B of the instantaneous frequency of s (t) is μT according to the equations [8] and [9].
In the pulse compression processing, after the transmitted modulated wave is received, the received signal r (t) is subjected to correlation processing using the following as a convolution integration response. Here, it is assumed that the target is stationary.

At this time, the reference function h (t) can be expressed as shown in Equation 11 and FIG.

  By performing the pulse compression processing, the LFM modulated wave is compressed in a pulse shape as shown in FIG. 10. This output waveform has an amplitude of √ (TB) times in amplitude, where T is the duration of the modulated wave and B is the bandwidth. (Power times TB), the pulse width is about 1 / TB times, and the same effect as transmitting a narrow pulse with a large peak power can be obtained.

1.8 Equivalence between the product of the convolution integral on the time axis and the frequency axis Since the pulse compression process is a convolution integral on the time axis,

If each term is Fourier transformed and substituted,

  Therefore, the following [Expression 15] is obtained.

  From [Expression 15], it can be seen that the convolution integral on the time axis is equivalent to the product of the frequency axis. Therefore, since the calculation amount is small, it is executed as multiplication on the frequency axis during the pulse compression processing.

1.9 Distance Sidelobe in Pulse Compression Pulse compression radar can achieve the same effect as transmitting narrow pulses with large peak power by performing pulse compression processing. It can be observed. However, a high distance side lobe is generated in the compressed signal as shown in FIG.

One method for lowering the distance side lobe level is to multiply the received signal by a weight in the frequency domain. There are various types of weights. Here, a typical Hamming window and a signal obtained by multiplying a signal by a Hamming window are shown in FIG.
As shown in FIG. 12, the distance side lobe can be reduced by multiplying the weight, but there are drawbacks in that a loss occurs in the received signal itself and the pulse width increases.

1.10 SWW method Realizing high range resolution in pulse compression radar requires a receiver bandwidth equivalent to the transmission frequency bandwidth. That is, since a broadband receiving system is required, the cost of the radar system is increased. On the other hand, the SWW (Synthetic Wideband Waveform) method realizes a high range resolution in a narrow band receiver bandwidth by frequency stepping a pulse compression wave having an intermediate band.
Hereinafter, frequency steps and signal processing configurations in the SWW method will be described.

1.11 Frequency Step of SWW Method As shown in FIG. 13, transmitting a wideband signal in a pulse compression radar requires an equivalent receiver bandwidth and a wideband receiving system capable of processing the same at the same time. Incurs high costs. Furthermore, in the case of higher distance resolution, there are restrictions on hardware such as the performance of the A / D converter. On the other hand, a frequency step is mentioned as a method of transmitting a signal in a time division manner. In the SWW method, an LFM modulated wave (sub-pulse) having a bandwidth matched to hardware constraints such as the performance of an A / D converter is frequency-stepped as shown in FIG. Specifically, the carrier bandwidth and the local signal inside the receiver are transmitted while switching each PRI (pulse repetition interval) and reception is performed between the PRIs. A bandwidth corresponding to the width is sufficient. Therefore, by performing the frequency step, it is possible to realize a narrow band receiver width (equivalent to the sub pulse bandwidth) as compared with the band of the signal radiated to the space (transmission occupied bandwidth).

1.12 SWW Method Transmission Frequency Sequence FIG. 14 shows a SWW method transmission frequency sequence. A continuous wave (CW wave) local signal f _n (n = 0, 1,..., N−1) coherent (constant in phase) within the pulse repetition time PRI is switched for each PRI, and sub-pulses using these as carrier waves ( LFM modulated wave having bandwidth b) is transmitted and received within the PRI.

1.13 Measurement signal model of SWW method In describing the measurement signal model in the SWW method, the time t _n = t−PRI · n, where the _n -th transmission wave transmission start time is 0, and the amplitude If 1, the transmitted wave is

It becomes. Note that μ = b / Tp is an LFM slope, and Φn is an arbitrary phase. The received wave for this transmitted wave is affected by time delay and Doppler shift,

(At this time, the speed v is positive in the direction of moving away). τ is time delay (round trip time to target)

Where R is the target distance, v is the relative speed to the target, and c is the speed of light. This received wave is mixed with the local signal fn (n = 0, 1,... N−1), and the sum signal of the frequencies is removed by an LPF (Low Pass Filter), whereby the following difference signal is obtained.

  In the formula [19], from 2v / c≈1,

It becomes. Next, when the time with T _n + τ as the time origin is set to t0,

Further, assuming that the time delay τ has n dependence, if τ _n is set,

Furthermore, from the transmission start frequency f ₀ , the local signal f _n (n = 0, 1,... N−1) is

It is.

Here, ignoring f ₀ > Δf and the constant phase term,

And the measurement signal

Is obtained.

1.14 SWW Signal Processing Configuration FIG. 15 is a diagram illustrating a signal processing configuration example according to the SWW method.
In the SWW method, pulse compression processing is performed on the measurement signal of [Expression 28]. Since the convolution integral on the time axis and the product on the frequency axis are equivalent, the pulse compression processing is processed as multiplication on the frequency axis as shown in FIG.

  That is, as shown in FIG. 15, the output of the sub-pulse generator 31 is supplied to the mixer 32, mixed with the output of the oscillator 33, supplied to the transmission / reception antenna 35 via the circulator 34, and transmitted. The signal received by the transmission / reception antenna 35 is mixed with the output of the oscillator 33 by the mixer 36, and the output is passed through the low-pass filter 37 and then digitally converted by the analog / digital converter 38. The converted signal is supplied to the Fourier transform unit 39 for conversion, and the converted signal is supplied to the mixing processing unit 40. In the mixing processing unit 40, a signal obtained by converting the reference signal from the reference signal generating unit 41 by the Fourier transform unit 42 is supplied, and a multiplication process is performed. The output of the mixing processing unit 40 is sequentially supplied to the spectrum shift unit 43, the combine unit 44, and the inverse Fourier transform unit 45.

  The pulse compression processing of [Formula 31] is output in a state where they all overlap on the frequency axis. Therefore, a spectrum shift process is performed on the pulse compression result at each frequency step. The shift width at this time depends on the frequency step width Δf, and spectrum shift is performed as shown in FIG.

  Synthetic band processing is performed on the result of pulse compression at each frequency step that has been expanded to a band equivalent to the transmission occupied bandwidth by spectrum shift processing. As a result, a combined band corresponding to the transmission occupation bandwidth as shown in FIG. 18 can be obtained.

  A range profile is obtained by performing IDFT processing on the obtained combined band.

  20 and 21, according to the SWW method, distance resolution corresponding to the transmission occupied band can be obtained, but there is a problem that distance side lobes and grating lobes are high.

1.15 NL-SWW (Nonlinear Synthetic Wideband Waveforms) Method The SWW method achieves high range resolution in a narrow receiver band by frequency stepping a pulse compression wave having a medium band. However, high distance peak side lobes (PSL) and grating lobes (GL) in the synthesis band are problems. To solve this problem, an NL-SWW method is proposed in which the effect of reducing the distance PSL is expected without loss of received power, unlike the case where the received signal is multiplied by a weight such as hamming by making the frequency step non-linear as shown in FIG. Has been.
Hereinafter, the non-linear frequency step and the signal processing configuration in NL-SWW will be described, but they are omitted because there are some points that overlap with SWW.

1.16 Nonlinear Frequency Step The frequency step in SWW is a linear frequency step with a constant frequency step width, whereas in NL-SWW, transmission is performed with frequency step widths that are not equally spaced, as shown in FIG. By setting the frequency step width to an unreasonable interval, the frequency density distribution within the transmission occupied bandwidth is changed, and the sampling interval becomes an unreasonable interval in the synthesis band (inverse Fourier transform). Thus, the frequency step is expected to reduce the grating lobe and the like.

  Signal processing after reception in the NL-SWW is the same as that in the SWW as shown in FIG.

2. Description of First Embodiment Next, an example of the first embodiment of the present invention will be described based on the above description.

2.1 Configuration of First Embodiment A high frequency resolution is realized in a narrow-band receiver bandwidth by stepping a non-linear frequency step of a pulse compression wave having a medium band, and combined band processing (IDFT processing) The NL-SWW method has been proposed in which the high distance PSL generated in (1) is reduced without signal loss unlike the case where the received signal is multiplied by a weight such as humming. However, the NL-SWW method uses an LFM modulated wave (chirp wave) as a sub-pulse, and requires a plurality of PRI (pulse repetition intervals) observation times depending on the frequency step. When there is a speed, there is a problem that it is greatly affected by the Doppler frequency and accurate distance measurement cannot be performed.

In the present embodiment, a configuration in which Doppler frequency estimation / correction processing using a pulse Doppler filter is combined with the problem.
Hereinafter, the transmission frequency sequence that is the basis of this configuration will be described, and the nonlinear frequency step and signal processing configuration in the first embodiment will be described.

2.2 Transmission Frequency Sequence In the first embodiment, as shown in FIG. 23, the start and end points of the frequency step are fixed, and the cubic function determined by the nonlinear parameter σ giving the third point therebetween is shown in FIG. A transmission frequency sequence having a non-linear frequency step interval dF _n of the subpulse (example of LFM modulated wave) shown in FIG.

As the cubic function corresponding to the nonlinear frequency step interval dF _n , for example, when N is an even number, it is possible to employ solving the following simultaneous equations.

Each coefficient of the cubic function is determined, and is determined by multiplying this by the maximum frequency step width f _max (frequency width from the start point to the end point of the frequency step). Thereby, the nonlinear frequency step interval dF _n can be controlled only by the nonlinear parameter σ.

  Since the multi-frequency NL-SWW method uses pulse radar as the basic principle, distance ambiguity does not occur. Therefore, radio waves reciprocate from one transmission subpulse to the next transmission subpulse, the pulse repetition time PRI (Pulse Repetition Interval). From the condition that

Need to be satisfied. Here, R _max is referred to as the maximum instrument distance required for the radar.
On the other hand, if the required speed resolution is σV, the required observation time Tc from the formula [6] is

It becomes. Here, λ is the wavelength of the transmission signal. If the total number of pulses of Tc within the observation time is N ₀ ,

It becomes. Next, let V _max be the velocity field of view required for the radar.

It is necessary to satisfy Here, M is the number of data samples within the observation time Tc necessary for obtaining the required speed field of view. At this time, N0> M can be satisfied depending on the maximum instrument distance, speed resolution, and speed field required for the radar. At this time, if Ts (≡Tc / M)> PRI and the ratio is an integer value N,

It can be. In this case, the transmission frequency sequence of FIG. 24 can be used. An in-vehicle radar or the like targeting a relatively short distance can satisfy these conditions, and therefore, multi-frequency transmission is possible only in one observation section Tc without reducing the speed resolution.
The upper limit that can be selected as an integer value N depends on the required maximum instrument distance R _max and the maximum velocity field of view V _max ,

It becomes.
For example, FIG. 25 shows the relationship of [Formula 41] when the number of frequency steps N = 8. In FIG. 25, the horizontal axis represents the maximum instrument distance R _max , and the vertical axis represents the maximum velocity field of view V _max . The solid line, dotted line, and broken line represent transmission frequencies of 100 GHz, 10 GHz, and 1 GHz, respectively. In FIG. 25, if the maximum instrument distance and maximum velocity field required for the radar are in a range below the respective lines, there is no ambiguity in both distance and speed.

2.3 Formulation of the first embodiment Formulation of the first embodiment is performed. Here, a measurement signal model of each subpulse will be described based on the formulation. When the time t _nm = t-PRI · n-PRI · N · m and the amplitude is 1 when the pulse repetition number mth and nonlinear frequency step nth transmission wave transmission start time is 0, the transmission wave is

It is written. Tp is a sub-pulse width, μ = b / Tp is an LFM slope, and Φnm is an arbitrary phase. The received wave for this transmitted wave is affected by time delay and Doppler shift,

(At this time, the speed v is positive in the direction of moving away). τ is a time delay (round trip time to the target) and is expressed as follows.

R is the target distance, v is the relative speed with respect to the target, and c is the speed of light. This received wave is mixed with the local signal f _n = f + dF _n (n = 0, 1,... N−1), and the following difference signal is obtained by removing the sum signal of the frequencies by the LPF.

From 2v / c-1≈1 in the formula 46, the following is obtained.

Next, if the time with t _nm + τ as the time origin is set to t ₀ ,

Further, assuming that the time delay τ is dependent on _{n and m,} and setting τ _{n, m} ,

Furthermore, from the transmission start frequency f, the local signal f _n = f + dF _n (n = 0, 1,... N−1) is

And it is as follows.

In this equation, ignoring f> dF _n and the constant phase term,

And the Doppler frequency fd (= 2vf / c)
Measurement signal of each sub pulse

Is obtained. When there are multiple targets, the measurement signal can be expressed as a linear sum of [Equation 56].
As can be seen from the equation 56, the target relative speed is obtained from the frequency of the m-direction sampling signal, and the frequency of the n-direction sampling signal is a function of the target distance and the relative speed. The first embodiment is characterized in that a frequency sequence number n is fixed and a transmission frequency sequence that can obtain a desired velocity resolution and maximum velocity field required for a radar by Fourier transform of an m-direction sampling signal is used.

2.4 Signal Processing Configuration of First Embodiment FIG. 26 shows a signal processing configuration of the first embodiment using the transmission frequency sequence shown in FIG. A cubic function determined by the nonlinear parameter σ is defined as a nonlinear frequency step interval dF _n, and the local signal f _n = f + dF _n (n = 0, 1,... N−1) is switched for each PRI, Sub-pulse (LFM modulated wave with bandwidth b) is transmitted.
That is, the local signal output from the oscillator 53 is mixed with the output of the LFM modulation unit 51, transmitted from the transmission / reception antenna 54, and the signal reflected by the target is received by the transmission / reception antenna 54. The received signal is supplied to the sub-pulse compression unit 55 via the transmission / reception separation unit 52, and then processed in the order of the Doppler frequency estimation processing unit 56, the Doppler frequency correction processing unit 57, and the synthesis band processing unit 58. .
Hereinafter, each process will be described.

2.5 Sub-pulse pulse compression Pulse compression is performed on the measurement signal of each sub-pulse obtained by the equation [56]. Since the convolution integral on the time axis and the product on the frequency axis are equivalent, the pulse compression is processed as a multiplication (Multiply) on the frequency axis and returned to the time axis by inverse Fourier transform (IFFT).

2.6 Doppler frequency estimation / correction processing by pulse Doppler filter Target sampling with desired velocity resolution and velocity field of view by Fourier transform of sampling signal in m direction for each n with respect to the signal after sub-pulse pulse compression Perform speed detection.
That is, as Doppler frequency estimation processing by pulse Doppler, m-direction Fourier transform shown in the following equation is performed on each n in Equation 60, which is a signal after pulse compression.

  Here, k (= 0, 1,... M−1) is a frequency channel number. The amplitude value | F (n, k) | after substituting the [Expression 60] expression into the [Expression 61] expression is the frequency channel number at each frequency step n.

In coherent integration, a peak is obtained.
In this way, the target Doppler frequency is obtained by detecting the frequency channel number k _peak where the output amplitude of the formula [61] reaches its peak. From the detected number k _peak , the target relative speed V is

Obtained from.
When there are a plurality of targets in the same distance gate, it is expressed by a linear sum of the formula [60], but fading occurs depending on the phase relationship. In order to alleviate this problem, for example, for each k, the sum of the absolute values of the Fourier transform output channels at each frequency step n is taken,

Is an input value for detection threshold processing.
As shown in [Formula 56], the frequency of the n-direction sampling signal is a function of the target distance and the relative speed, and correction processing is performed based on the obtained Doppler frequency.

  As a result, if only the target distance information is included in the n-direction sampling signal, it is expressed as follows.

2.7 Band Synthesis by Inverse Fourier Transform According to [Formula 66], inverse Fourier transform is performed in the n direction at Nr points equal to or greater than N on the signal with the Doppler frequency corrected. However, since the frequency step width is not constant and is given by dF _n , the ambiguity distance is indeterminate. Therefore, the inverse Fourier transform is performed on the entire range up to the maximum instrument distance _Rmax .

If the value of l at which the absolute value of P (l) takes the maximum value is l _peak , the distance R _cal closest to the target can be measured by obtaining l _peak as shown in the following equation.

  In the configuration of the first embodiment described above, there is a relative speed between the high distance resolution in the narrow receiver band, the low distance PSL in the synthesis band, and the target that has been a problem in the SWW and NL-SWW methods. Even in some cases, accurate distance measurement is possible. Furthermore, since the distance is obtained for each detected target relative speed, it is possible to avoid a malfunction of pairing of the detected frequencies of the up sweep and the down sweep, which is a problem in the FMCW method.

2.8 Simulation Example of the Present Embodiment By computer simulation, the Doppler frequency estimation / correction processing by the pulse Doppler filter and its correction error in the configuration of the first embodiment, the correction error, the distance resolution by the nonlinear synthetic broadband wave, the distance by the nonlinearization Each peak side lobe (PSL) reduction effect was evaluated.

The simulation conditions were assumed to be applied to millimeter-wave vehicle-mounted radars, particularly long-range radars. The following parameters that satisfy one of the required specifications, the maximum instrument distance of 200 m or more and the maximum speed field of view of ± 200 km / h or more, were adopted.
Transmission frequency f: 76.5 GHz (wavelength λ = c / f = 3.922 × 10 ⁻³ )
Pulse repetition period PRI: 2.0 μs (maximum instrument distance Rmax = 300 m)
-Number of frequency steps N: 8 (maximum velocity field of view Vmax = ± 220.588 km / h)
Subpulse bandwidth b: 80 MHz (distance resolution with the same bandwidth in the pulse compression radar system: δR = 1.875 m)
Sub pulse width Tp: 0.15 μs
・ Transmission frequency occupied bandwidth B: 360 MHz
-Number of same frequencies within observation time M: 256
-Total observation time Ts: 4.096 ms

On the other hand, the target number was set to 1.
・ Target distance R: 200m
・ Target speed V: 200 km / h (= 55.556 m / s)

2.9 Doppler Frequency Estimation / Correction Processing by Pulse Doppler Filter FIG. 27 shows the amplitude of each frequency channel of the Fourier transform output of the m-direction sampling signal for each n after pulse compression of each subpulse, and the output amplitude peaks The target Doppler frequency fd is obtained by detecting the frequency channel number. From this, when the relative speed was calculated, a detected value of 55.53 (m / sec), a set value of 55.556 (m / sec), and an error of -0.026 (m / sec) were obtained.

  As described above, in the Doppler frequency estimation process by the pulse Doppler filter in the configuration of the first embodiment, the detection error based on the velocity resolution δV = 1.723 km / h (= 0.48 m / s) is ± 0 at the maximum. .24 m / s is generated. However, since the speed resolution is increased by increasing the total observation time Ts, this detection error can be reduced.

2.10 σ Dependence of Distance Peak Sidelobe (PSL) Nonlinear Frequency Step Interval dFn in the Configuration of the First Embodiment, and Nonlinear Parameter σ that Determines the Third-Order Function that Gives it is −0.01 to −0. It was changed in the range of 07. This is aimed at collecting sub-pulses, that is, transmission power at the center of the band by non-linearization as shown in FIG. When σ = −0.01, the sub-pulses are gathered at the center, and from there, σ is changed to approach the linear frequency step (σ = −1 / 14).
The σ dependence of the distance PSL after the inverse Fourier transform (combined broadband) | P (l) | was evaluated, and the σ dependence result of the distance PSL is shown in FIG.

  When σ = −0.055 under the condition of the number of frequency steps N = 8 and the transmission frequency occupation bandwidth B = 360 MHz, the result that the PSL was the smallest was obtained. The pulse compression result of the sub-pulse at this time, the configuration of the first embodiment (σ = −0.055), frequency step N = 8, frequency step width Δf = 40 MHz, transmission frequency occupation bandwidth B = 360 MHz, etc. FIG. 30 shows an inverse Fourier transform output in the L-SWW having the same radar parameters as the configuration of the first embodiment.

  As shown in FIG. 30, in the configuration of the first embodiment, unlike the case where the received signal is multiplied by the weight, the effect of reducing the distance PSL to −22 dB at maximum is obtained without any signal loss. Also, this result is a distance PSL reduction effect of −9 dB compared to L-SWW (theoretical value—13.2 dB). Further, the grating lobe (GL) is −19 dB, and the frequency step interval is not constant. Therefore, although it cannot be strictly compared with L-SWW, a reduction effect of −3 dB is obtained.

2.11 Distance resolution by nonlinear synthetic broadband waves The distance resolution in radar depends on the transmission frequency occupation band. The pulse compression radar requires the same receiver bandwidth as the transmission frequency occupation bandwidth. However, in the configuration of the first embodiment, the receiver bandwidth needs only the sub-pulse bandwidth (b = 80 MHz) due to the nonlinear frequency step. As a result, the configuration of the first embodiment having a non-linear synthesis band (= transmission frequency occupation band B = 360 MHz) obtained by synthesizing N sub-pulse bands while having the same receiver bandwidth is as shown in FIG. Compared with sub-pulse pulse compression (bandwidth b = 80 MHz), a distance resolution of about 4.5 times was realized.

3. Description of Radar Detection Technology as Premise of Second Embodiment Next, radar detection technology as a premise of the second embodiment will be described in order. A part of the description overlaps with the description of the base technology in the first embodiment, but will be described again in order.

3.1 Detection Distance and Distance Resolution of Pulse Radar Pulse radar is applied also in the second embodiment. The representative of the radar performance is the detection distance and resolution, but the maximum detection distance of the pulse radar is derived from the following [Equation 69], and the distance resolution is derived from the following [Equation 70]. The However, the transmission output and _{P t.}

  Therefore, in order to design a pulse radar with good performance, how to obtain a large S / N ratio and a pulse having a small pulse width are important points. However, since the transmission / reception bandwidth B increases proportionally as given by the following [Equation 71] due to the reduction of the pulse width, the high range resolution of the radar by the simple reduction of the pulse width can be achieved by the transmitter / receiver. This leads to higher costs and is a foothold to be adopted as a product.

  Moreover, since the output that can be transmitted as electromagnetic waves is defined by laws and regulations, there is a limit to the improvement of the SN ratio due to the increase in output power.

3.2 Pulse compression radar In the pulse radar, there are two problems that the transmission power Pt must be increased to increase the detection distance and that the distance resolution must be improved within a limited frequency bandwidth. A technique called pulse compression used to solve this problem uses a transmission signal having a wide pulse width obtained by performing frequency modulation (FM) or phase code modulation (PM) in a transmission wave pulse. In this technique, demodulation is performed in the autocorrelation processing after reception to convert the pulse width into a large pulse width with a large power.

  The transmission / reception sequence of the pulse compression radar is shown in FIG. A reception pulse b is obtained with a delay of a predetermined time τ from the transmission of the transmission pulse a, and a short pulse c having a large electric power is obtained by compressing the reception pulse.

3.3 Pulse compression radar and autocorrelation processing By performing pulse compression by autocorrelation processing of the received wave with the transmission wave, it is possible to obtain a short pulse that is easy to detect and has high distance resolution with a large maximum detection distance. The autocorrelation processing performed here is G (t) shown in the following equation, and the received wave B (t) received with the transmission wave reflected by the target and attenuated in amplitude, and the time-reversal signal H of the transmission wave It is the output by the convolution integral by (t).

  Here, an example of FM pulse compression is shown. Let the received wave be B (t),

Assuming that (μ is an arbitrary constant), graphs of the time reversal signal H (t) of B (t) and B (t) are as shown in FIGS. FIG. 33 is a waveform diagram showing an example of a received wave in FM pulse compression, FIG. 34 is a waveform diagram showing an example of a reference function in FM pulse compression, and FIG. 35 shows an output example of FM pulse compression. It is a waveform diagram.

  Convolutional integration requires a very large amount of calculation. In actual signal processing, in [Expression 70], FFT is performed for each of B and H, and a product on the frequency axis is used before IFFT is used. Many. In addition, here, the transmission / reception modulation wave is shown as being performed by frequency modulation (FM), but the same applies to the case where phase code modulation (PM) is performed.

3.4 Pulse Compression Radar Using Complementary Code Complementary Code (CC) is a phase modulation pulse compression method that assigns a code to a phase. The feature of this code is that it is possible to completely suppress the distance side lobe, which is a component other than the pulse signal due to the target reflected wave, by combining two complementary functions.

A method for generating a complementary code will be described.
Consider two codes a _i and b _i of length K, and let their autocorrelation functions be A _j and B _j . At this time, what added these is shown by following Formula.

When this condition is satisfied, these are called complementary codes. For example, it can be easily confirmed that a _i = {1,1} and b _i = {1, −1} are complementary.

Further, the complementary code has an important property that a code having a length of 2K can be derived from a code having a length of K as follows. the concatenation of _{b i} after the a _i and _{c i,} the concatenation by inverting the sign of _{b i} after the _{a i} and _{d i.}

  At this time, the autocorrelation functions of these codes are

Thus, clearly, C _j and D _j constitute a complementary code having a length of 2K. Further, by repeating this operation, a complementary code having a length K = 2 ^m can be derived from a complementary code having K = 2. Further, when actually used for phase modulation, phases of 0 and π are assigned to values of 1 and −1, respectively.
FIG. 36 shows the result of using this result to derive the complementary code when K = 16.

  Using this code, the transmission wave of the carrier wave f is phase-modulated and transmitted / received to / from one target measurement object. Assuming that the received waves are B1 (t) and B2 (t), the following equations are obtained. However, since it is confused with a complex symbol, the complementary code array j is set to t.

  The complementary codes obtained by these equations are similarly subjected to the pulse compression processing described above, and FIG. 37 shows two signal waveforms.

Furthermore, when these two pulse-compressed waveforms are added to the j-axis, the waveform shown in FIG. 38 can be obtained.
As can be seen from FIG. 38, the complementary code can confirm that the distance side lobe is completely suppressed on the horizontal axis j other than the pulse peak by addition.

3.5 Phase modulation by complementary phase code (CPC)
Complementary Phase Code (CPC) is an extension of two-phase complementary code (Complementary Code) that has two phases. It uses two or more complementary functions to completely cancel out sidelobe generation. It is a function that can The features are based on the Complementary Code.

3.6 SWW (Synthetic Wideband Waveform) method In order to achieve high range resolution in pulse radar or pulse compression radar, a wide receiver bandwidth equivalent to the transmission frequency bandwidth is required from [Equation 71]. This leads to higher costs. Here, the Linear Frequency Stepping (LFS) method is one of the SWW methods for realizing high range resolution with a low receiver bandwidth by linearly stepping the pulse compression wave with respect to the frequency axis. There is. FIG. 39 shows an LFS transmission / reception sequence.

In an up-step wave (step FM wave) whose frequency increases stepwise, transmission is performed while switching between the carrier frequency and the local signal inside the receiver for each pulse repetition time (T _PRI : Pulse Repetition Interval). Also receive. That is, it is sufficient that the receiver bandwidth has a sub-pulse bandwidth of the transmission pulse.

  By mixing the received signal with a local signal, a signal having the number of frequency steps can be obtained. By performing synthetic band processing (IDFT, MUSIC, etc.) between each frequency step on this signal, fewer receivers can be obtained. Distance resolution can be obtained in the frequency bandwidth. Furthermore, by repeating this LFS sequence, the Doppler shift can be derived from the phase gradient of the pulse train with the same frequency, and the target relative velocity can be obtained.

  Here, modeling using mathematical expressions is considered. Assuming that a pulse signal transmitted as LFS is reflected by a target and received after a certain time delay τ after transmission, the received signal Rx (t) is defined by the following equation.

  However, for simplicity, the amplitude value is 1, the distance to the target is R, the target speed is v, and the variable n (n = 0, 1,... N−1) is the frequency step number. Here, when B is a transmission frequency bandwidth, the frequency step width of Δf and the time delay τ are

It is shown. By substituting these [Equation 80] and [Equation 81] into the [Equation 79], RX _n (t) becomes

It is shown. Next, an A / D converter having the same sampling interval as the transmission pulse width is used to represent complex I and Q signals divided into a plurality of sampling points for each PRI. Sampling in the step direction with a time delay fixed, with one frequency step interval as T _PRI (time t = nTT _PRI signal is

It becomes.

3.7 SWS Processing of LFS Method by Discrete Inverse Fourier Transform By performing SWW processing for the frequency step n direction with respect to Equation [83], the distance resolution can be improved. Since it is clear that n is a discrete value, here, when SWW processing by frequency step N as SWW processing and SWW processing by IDFT is performed, the following equation is obtained. However, the number of sample points for performing IDFT is Nr for generalization.

However, Nr ≧ N, k = 0, 1,... Nr−1
Since n ² exists in the term of Σ in the [Equation 84], if k = kp when the absolute value of IDFT (k) is the maximum is the target speed v = 0,

Can be expressed as It can also be seen that IDFT (k) is a sinc function having a peak at k = kP. Further, by using the discretization unit c / {2Nr · (B / N)}, the estimated target distance R _cal is

As given.
However, when the target speed v ≠ 0, that is, while the Doppler shift is present, the target distance R and the target speed v, which are unknowns, exist simultaneously in the term of Σ, so the absolute value of IDFT (k) is the maximum. If k = kP is obtained and R _cal is derived from the equation [85], it does not correspond to the correct target distance R. In order to solve this, it is necessary to measure the target speed v by Doppler shift and make it known before SWW by the LFS method. As described above, this can be derived by Fourier transforming a pulse train having the same frequency by repeating the LFS sequence. This speed detection is described in the item of multi-frequency step ICW radar.

  Based on the above, LFS has the feature that the required receiver frequency bandwidth is small, the point that distance ambiguity (grating lobe) occurs, the need to perform phase compensation to overcome the adverse effects of Doppler shift You have to overcome that.

3.8 Two-frequency CW radar The two-frequency CW radar uses a LFS transmission pulse as a continuous wave and a frequency step number N = 2, and can detect the target distance and speed in a very narrow frequency occupation band. Radar system. In the two-frequency CW system, as shown in FIG. 40, a CW wave having a frequency f2 that is slightly separated from the transmission frequency f1 is transmitted as a continuous wave in a time division manner during the time T _PRI (total observation time is 2). To do.

Section of the transmission frequency f ₁ in the receiving system is frequency f _1, section of the frequency f ₂ is mixed with the local signal of frequency f _2. The output signal after mixing has a small difference between the transmission frequencies f ₁ and f ₂ .

It becomes. That is, the received signals from the same target are observed as the same Doppler frequency in both sections of the transmission frequencies f ₁ and f ₂ . The relative speed v with the target at this time is

It becomes. The sampling data in each section of the frequencies f ₁ and f ₂ are Fourier transformed (FFT (Fast Fourier Transform) is generally used), and the Fourier transform output | F (n, k) | (the variable n is the frequency f ₁ , f ₂ , and the Doppler frequency (that is, the target relative speed) from the frequency (frequency channel number k) at which the value of k (= 0, 1,... M−1) represents the frequency channel number of the Fourier transform becomes a peak. v), and the target distance is obtained by using the phase difference Δψ = ψ2-ψ of the frequency component,

It is requested from. At this time, in order not to generate distance ambiguity,

Need to be satisfied.
Although the target distance and the relative velocity in a narrow frequency band occupied by the two-frequency CW method is obtained, a constant velocity ([number 87] where [Expression 88] where in f _D are the same) if there are several targets of the peak frequency The distance measuring method using the phase difference of the components ([Expression 90]) has a principle problem that malfunction occurs. Note that the speed resolution by Fourier transform in the 2-frequency CW method is

It is.

3.9 Multi-frequency step ICW radar This is a radar system that expands the dual-frequency CW system and enables distance measurement of the constant speed multiple targets, which is a problem, and has the following features in configuring the LFS transmission / reception sequence. Have.
(1) The transmission frequency sequence shown in FIG. 41 is used. Here, it is assumed that the transmission wave of each step frequency is coherent within the observation time Tc. On the other hand, the phase between each frequency is arbitrary. (2) Transmission is pulsed and reception is processed for each distance gate.
(3) The target distance is estimated by the one-dimensional super-resolution method using the output of the target speed detection process by FFT.
In the following, a transmission frequency sequence that is the basis of a multi-frequency step ICW radar will be described, and a constant-speed multi-target distance separation method that is a problem in the 2-frequency CW system will be described.

3.10 Multi-frequency step ICW transmission / reception sequence A multi-frequency step ICW transmission / reception sequence is shown in FIG.
Since no distance ambiguity occurs in the pulsed radar, from the condition that the radio wave reciprocates from one transmission pulse to the next transmission pulse (pulse repetition time T _PRI : Pulse Repetition Interval),

Need to be satisfied. Here, R _max is called the maximum instrument distance required for the radar. On the other hand, if the required speed resolution is δV, the required observation time T _c from the equation [92] is

It becomes. Here, λ is the wavelength of the transmission signal, and it is assumed that λ≡λ _n = c / f _n (n = 0, 1,..., N−1) as in the two-frequency CW system. If the total number of pulses within the observation time T _c is N ₀ ,

It becomes. Next, assuming that the speed field of view required for the radar is ± V _max ,

It is necessary to satisfy Here, M is the number of data samples within the observation time _Tc necessary to obtain the required velocity field of view. At this time, N0> M can be satisfied depending on the maximum instrument distance, speed resolution, and speed field required for the radar. At this time, when T _s (≡T _c / M)> T _PRI and the ratio is a positive value N,

It can be. In this case, the transmission frequency sequence of FIG. 41 can be employed. In-vehicle radars targeting relatively short distances can satisfy these conditions, so multi-frequency CW waves can be transmitted only in one observation section _Tc without reducing the speed resolution. is there. The upper limit that can be selected as a positive value N depends on the required maximum instrument distance R _max and the maximum speed field of view V _max ,

It becomes.

3.11 Constant Speed Multiple Target Distance Separation Method in Multifrequency Step ICW System A constant speed multiple target distance separation method using the transmission frequency sequence shown in FIG. 41 will be described. FIG. 42 shows a configuration block diagram.
The output of the oscillator 61 that outputs the step frequency is supplied to the up-converter 62 and transmitted from the transmission / reception antenna 65 via the high-frequency switch 63 and the circulator 64. A signal received by the transmission / reception antenna 65 is supplied to the down converter 66 via the circulator 64 and subjected to frequency conversion, and the frequency-converted signal is supplied to the detection unit 69 via the Fourier transform unit 67 and the frequency smoothing unit 68. Supply.
42, the oscillator 61 has a function of generating a CW wave f _n (n = 0, 1,..., N−1) that is coherent (the phase is constant within the observation time) within the observation time T _c . _These are sequentially switched and output for each T _{PRI at the} timing shown in FIG. In the high frequency switch 63, the transmission wave from the oscillator 61 is pulsed (pulse width _Tw ). The pulsed transmission wave is radiated from the transmitting / receiving antenna 65 to the space via the circulator 64.

  A measurement signal model in the multi-frequency step ICW method will be described. For simplicity, consider a transmission / reception signal in a non-pulsed situation. Similarly, if the amplitude value is 1 for simplicity, the transmission wave is

It is shown. φ _n is an arbitrary phase for each step frequency. The transmission wave reflected on the target enters the transmission / reception antenna as a reception wave after a time delay τ corresponding to the round trip time to the target. At this time, the received wave is

It becomes. Here, the amplitude is set to 1 for simplicity. Here, if λn (≡c / f _n ), f _dn (= 2v / λ _n ) is the Doppler frequency, c is the speed of light, and R is the target distance at time t = 0.
This received wave is mixed with the transmission signal from the oscillator 61 by the down converter via the circulator, and as an observation signal at the distance gate number (that is, the time delay τ) including the target,

Is obtained. Here, similarly to the two-frequency CW system, the difference Δf in frequency at each frequency step is sufficiently small with respect to the transmission frequency, and the Doppler frequency at each frequency step is equal.

Next, consider a measurement signal model when transmission is pulsed. If the pulse repetition number is m (= 0,..., M−1), the real time t _{m, n} of the distance gate corresponding to the time delay τ is

And the measurement signal model when the target is included in the distance gate from the equation [100] is

It is written. When there are a plurality of targets within the same distance gate, the measurement signal can be written as a linear sum of [Equation 104].

As can be seen from the equation (104), the target relative speed is obtained from the frequency of the m-direction sampling signal, and the frequency of the n-direction sampling signal is a function of the target distance and the relative speed. A super-resolution method such as two-dimensional MUSIC, two-dimensional ESPRIT / two-dimensional unitary ESPRIT, or Capon can be applied as a two-dimensional frequency analysis method to the measurement signal represented by the formula [104].
However, as described above, the multi-frequency step ICW method uses a transmission frequency sequence that provides a desired velocity resolution and maximum velocity field required by the radar by Fourier transform of the m-direction sampling signal with a fixed frequency step number n. It is a feature. As a result, in order to perform distance separation exceeding the resolution of Fourier transform, it is possible to reduce the amount of calculation compared to a method using the two-dimensional super-resolution method. Hereinafter, the signal processing configuration will be described.

1) Target relative velocity detection processing (Target velocity detection)
First, the target relative speed detection with a desired speed resolution and speed field of view is performed by Fourier-transforming the sampling signal in the m direction for each n.
In other words, in the target speed detection process, a Fourier transform process in the m direction shown in the following expression is performed on each n with a measurement signal (formula 104) for each distance gate.

Here, k (= 0, 1,..., M−1) is a frequency channel number. The amplitude value | F (n, k) | after substituting the [Equation 104] into the [Equation 105] is the frequency channel number at each frequency step n,

In coherent integration, a peak is obtained.
In this way, the target Doppler frequency is obtained by detecting the frequency channel number k _{peak at} which the output amplitude of the formula [105] reaches its peak. From the detected number _kpeak , the target relative speed V is

Obtained from. Also, the frequency channel output at k _peak is

It becomes.
When there are a plurality of targets within the same distance gate, it is expressed by a linear sum of the formula [104], but fading occurs depending on the phase relationship. In order to alleviate this problem, for example, for each k, the sum of the absolute values of the Fourier transform output channels at each frequency step n is taken,

Is an input value for detection threshold processing.
In the multi-frequency step ICW system having the configuration shown in FIG. 42, if there are a plurality of speed difference targets below the speed resolution by Fourier transform (Equation 105) in the same distance gate, based on information from the tracking filter or the like. When the determination is made, the target distance is obtained from the phase difference of the detection frequency channel k _peak in two consecutive frequency steps (n and n + 1) based on the normal two-frequency CW method. That is, the phase difference of the detection frequency channel depending on the time difference of each frequency step was corrected for F (n, k _peak ).

The phase ψ (n, k _peak ) of the phase and the phase difference or the average of the phase differences

The target distance can be obtained from

2) Target distance detection processing; Frequency averaging (Frequency smoothing)
In multi-frequency step ICW, target distance detection processing using super-resolution method is applied. In other words, when the frequency channel k _peak exceeds the threshold value and target detection occurs in the threshold processing of the formula [109] at a certain distance gate, the distance gate width and the detected relative speed (the formula 107) If there is any possibility that multiple targets exist from the prior information (tracking filter information, other sensor information, etc.) within the velocity resolution centered on, the Fourier transform output in the target relative velocity detection process shown below is The target distance detection process used is applied.

First, as a pre-processing using the super-resolution method in a multi-wave environment, it is necessary to perform frequency averaging in order to restore the correlation matrix rank. Since the speed difference between the target is arbitrary continuous quantity, the input of the frequency average, frequency number k _peak and its longitudinal ± 1 total channel 3 in this example without limiting the peak frequency number k _peak channels on the Fourier transform output A channel data vector is used. As a sub-matrix composed of Ns rows in the column direction (n direction) with respect to the matrix F having these data vectors as column vectors,

Define However, q = 0,..., N−Ns. Here, submatrix [X; n = a, b, k = c, d] represents a submatrix of the matrix X from a row to b row and from c column to d column. The frequency average is a process for obtaining the following correlation matrix R by the average process of the partial correlation matrix of the sub-matrix Fq.

Here, H is a complex transpose of a matrix, and <*> is an average operation with respect to q.

3) Target distance detection processing; Super resolution range estimation
A distance estimation method when the MUSIC (MUltiple SIgnal Classification) method is adopted as an example of the super-resolution method will be described. When the frequency step width Δf is limited to an equal interval, it is possible to employ ESPRIT with a smaller calculation amount.
In the MUSIC method, the eigen-expansion of the correlation matrix R (equation 113) after frequency averaging is performed, and the noise space E = [consisting of eigenvectors eα (α = 1,... Ns−L) corresponding to the eigenvalues of noise. e ₁ ,... e _Ns−L ]. Here, L is the number of signals, and is obtained from the number of eigenvalues larger than the eigenvalue of noise, for example.
Next, as a steering vector a (R) for searching for a target distance by the MUSIC method,

Is used. Here, k _peak is a frequency channel number obtained by the target speed detection process, and can be handled as a known amount in the equation [114]. Therefore, the unknown quantity included in the steering vector a (R) is only the distance R to be estimated. In the MUSIC method, using this steering vector a (R) and the noise space E,

Is the evaluation function, and the distance R at which the peak is obtained within the target distance gate is set as the target distance estimated value. When there are a plurality of constant velocity targets, a plurality of peaks are observed in the evaluation function of the formula [115]. The distance R at which these peaks are obtained is defined as a distance estimated value R.
With the multi-frequency step ICW method described above, it is possible to measure the target distance and speed in the constant-speed multi-target environment with the measurement time of only one observation time _Tc , and obtain the distance for each detected target relative speed. Therefore, it is possible to avoid a malfunction in pairing of the detection frequency.

4). 4. Description of Second Embodiment 4.1 Outline of Complementary Coding Band Synthesis Method In this embodiment, the pulse in the LFS method described as the existing technology is phase code modulated by the CPC code described above. This is a technique of pulse compression at In the present embodiment, this is referred to as a complementary coding band synthesis method.

As a characteristic of the CPC, since the side lobes can be completely canceled, it is possible to suppress the inflow of the signal level of the reflected wave outside the measurement target by pulse distance gating. In this case, what is performed by IFFT as processing in SWW is presented as Hybrid-CFS. Furthermore, it can be expected that the distance resolution capability is further enhanced by a complementary multi-frequency step SWW using MUSIC, which is one of the super-resolution methods used in SWW.
There are the following two complementary coding band synthesis methods of the present embodiment described here.
・ Complementary coding band law 1 Hybrid CFS (CPC + LFS + IFFT)
2 Complementary multi-frequency step CW (CPC + LFS + MUSIC)

In the present embodiment, because of the combination of the CPC and LFS methods, considering the arrangement of two different complementary codes, the following two pulse arrangement methods are assumed.
(1) Arrangement of different signs in the frequency axis direction (2) Arrangement of different signs in the time axis direction

  Furthermore, it must be taken into account that there is a need for phase correction, since each will have a pulse for a different carrier frequency and different reception time.

4.2 Transmission / reception sequence of complementary coding band synthesis method Considering transmission / reception sequence as complementary coding band synthesis method, the above-mentioned (1) and (2) can be broadly considered. As represented by code modulation mapping in wireless communication, various methods can be applied. Among them, in this embodiment, in the complementary coding band synthesis method, the on-vehicle radar in which the phase compensation for the influence of the time axis shift, the phase compensation for the influence of the Doppler shift, and the target position exists at 0 to about 200 m. Taking the application into consideration, an example using the following four transmission / reception sequences will be described.

(1) Method (1) of assigning different codes to the frequency axis Separation by frequency axis Case 1
The transmission / reception sequence shown in FIG. 43 is a method of transmitting / receiving CPC modulated waves having different codes that are alternately complementary to the frequency axis fn direction _where N steps exist as LFS. By adding f _2n and f _{2n + 1} (where n = 0, 1,... N / 2-1) as a pair, the side lobes can be canceled during CPC pulse compression.
ex) adds _{f 1} and _{f 2} as a pair, adds _{f 3} and _{f 4} as a pair (hereinafter the same)
Therefore, compared with the multi-frequency step ICW, the number N of pulses in the frequency axis direction becomes 1/2 after the addition of complementary codes.

(1) -2 Separation by frequency axis Case 2
The transmission / reception of different CPC modulated waves that are alternately complementary to the direction of the frequency axis f _n with N steps is the same as in case 1 of (1). However, the addition of f _n and f _{n + 1} (where n = 0, 1,... N−1) as a pair makes it possible to cancel side lobes during CPC pulse compression.
ex) Add f ₁ and f ₂ as a pair, add f ₂ and f ₃ as a pair (the same applies hereinafter)
Therefore, compared with the multi-frequency step ICW, the number N of pulses in the frequency axis direction becomes N−1 / N after the addition of complementary codes.
(2) Method of assigning different codes to the time axis (2) Separation by 1 time axis Case 1

As shown in FIG. 44, this is a method of transmitting and receiving different CPC modulated waves that are successively complementary after a certain time delay with respect to the time axis t.
After reception, pulse compression is performed on different complementary codes adjacent at the same frequency, phase compensation is performed, and addition is performed. If the observation time is the same as the multi-frequency step ICW, the number of pulses M on the time axis after the addition of the complementary code is ½ compared to the multi-frequency step ICW.

(2) -2 Separation by time axis Case 2
As shown in FIG. 45, N frequency steps are swept in the same code with respect to the time axis t, and a code-modulated wave is transmitted and received. Thereafter, different CPC modulated waves that are complementary are similarly frequency stepped and transmitted / received. After reception, pulse compression is performed on different complementary codes adjacent at the same frequency, phase compensation is performed, and addition is performed. If the observation time is the same for the multi-frequency step ICW, the number of pulses M on the time axis after the addition of the complementary code is ½ compared to the multi-frequency step ICW.

4.3 Transmission / Reception Sequence of Complementary Coding Band Synthesis Method Applied in Second Embodiment As long as two complementary codes are used, some performance is obtained with respect to the number N of pulses in the frequency step direction and the pulse M in the observation time direction. It must be allowed to decline.
Here, the effects on the number N of frequency steps and the pulse M in the observation time direction are summarized as follows.

-Reduction in the number N of frequency step direction pulses Since the distance estimated value by the IDFT depends on the [Equation 86], the distance resolution decreases. Furthermore, when using MUSIC in the multi-frequency step ICW, the number of same-speed multi-target separables depends on N-N _s (refer to [Equation 112]), so multi-target separation performance, that is, radar distance resolution is remarkably high. descend.

-Decrease in the number M of pulses in the observation time direction This directly leads to a decrease in velocity resolution (see [Equation 94]). Moreover the increase of T _s is the speed detected when performed after different signs added to the complementary, also decreases the maximum rate detection performance. However, by using absolute values between different codes before addition, these can be made to have the same waveform, so that different code pulses at the same frequency step are regarded as the same code pulses and FFT (1 in (2)). In case 1, the maximum speed detection performance is not sacrificed by performing speed detection by non-uniform sampling.

  Here, considering that it is used as an in-vehicle radar, and if a general 76.5 GHz band is used as an in-vehicle radar, a speed resolution of 1 km / h or more is obtained in an observation time of several msec to several tens of msec. Can be obtained. Based on this, it can be considered that the penalty for the increase in observation time has little effect on the operation as an on-vehicle radar. Furthermore, when operating as an on-vehicle radar, a decrease in the same speed multi-target separation performance due to a decrease in the number N of frequency step direction pulses is a significant performance decrease. Therefore, as a complementary coding band synthesis method, the method of (2) is used as a transmission / reception sequence design, and the method of preventing the decrease in the speed resolution by extending the observation time by reducing the number M of observation time direction pulses is the most in-vehicle radar. Is appropriate.

  Next, the difference between 1 in (2) and 2 in (2) will be described. The most different point between the two is the difference in time delay between signals to be added. Phase compensation is performed on each of the two signals before adding the two complementary signals. However, considering that the CPC side lobe cancellation is performed when complete phase compensation is performed, there is an error in phase compensation. Less is appropriate. Further, when considering multipath waves, clutter, etc. when operating as an actual in-vehicle radar, a signal with a small time delay is easy to detect a signal and a correction error is small. For this reason, the method (2) (2) with a small delay with respect to the time axis is a more appropriate transmission / reception sequence when using CPC.

4.4 Derivation of Target Speed / Target Distance by Complementary Coding Band Combining Method Here, FIG. 46 shows details of the transmission / reception sequence of the method of FIG.

In the complementary coding band synthesizing method according to the present embodiment, a complementary code pulse that is phase-modulated by using a complementary code is transmitted at equal frequency LFS at a frequency step, and a measurement target reflected pulse is received. The received wave is converted to baseband at each frequency step to perform complex IQ detection. Thereafter, A / D conversion is performed to obtain a sampling signal within 1 PRI. Using this sampling signal, complementary code pulse compression processing (autocorrelation processing) is performed. As a result, a signal is obtained. FIG. 47 shows a signal processing system diagram of the complementary multi-frequency step ICW using this signal. As shown in FIG. 46, the signal is described by four variables (ss, α, n, m).
Referring to FIG. 47, the reception signal obtained by the reception antenna 71 is detected by the LFS detection unit 72. The LFS detection unit 72 includes an oscillator 72a, a mixer 72b, and a bandpass filter 72c. The output of the band pass filter 72 c is supplied to the complex IQ detection unit 73. The complex IQ detection unit 73 includes an oscillator 73a, a mixer 73b, a phase shifter 73c, a mixer 73d, a low-pass filter 73e, and a 73f low-pass filter to obtain two received signal components, and a correlation processing unit 74. To supply. In the correlation processing unit 74, digital conversion is performed by the analog / digital converters 74 a and 74 c, the correlation is detected by the matched filters 74 b and 74 d, and the detection signal is supplied to the received wave pulse processing unit 75. The received wave pulse processing unit 75 performs pulse compression.
The four variables are as follows:

M: Number of pulses “of each sign” on the time axis. However, if complementary codes are used, there are two types of pulses, so the total number of pulses for one frequency is 2M.
N: Number of frequency steps.
α: A value that identifies a complementary code. Since there are two types of complementary codes, 0 or 1 is taken.
SS: ^Number of sampling in ^TPRI .

The variables N and M are the same as those in the multi-frequency step ICW without complementary encoding, and α and SS are newly introduced this time. In particular, in order to achieve complete sidelobe cancellation by adding a code complementary to the phase encoding by CPC, the variable due to sampling within 1 ^TPRI must also perform phase compensation for the sidelobe in pulse compression. Since it is essential that it does not become, it was set. If a pulse compression wave is used only for one code such as a standard multi-frequency step ICW or P4 code when performing phase modulation, a certain threshold is set for the gain of the pulse wave. However, it is sufficient to perform phase compensation only on the pulse peak.
Next, a measurement signal model in the multi-frequency step ICW method will be described. The basic configuration conforms to the multi-frequency step ICW.

4.5 Derivation of Received Wave Pulse in Complementary Coded Band Synthesis Method Analyzing the received wave pulse in the complementary coded band synthesis method using mathematical formulas.
First, consider the transmitted wave. For the sake of simplicity, considering the entire transmission carrier wave with an amplitude of 1, as in [Equation 90] in the multi-frequency step ICW,

Can be written. φ _n is any phase that is equal in the variable n.
The transmission wave reflected on the target enters the transmission / reception antenna as a reception wave after a time delay τ corresponding to the round trip time to the target. Similarly, the received wave at this time is derived from [Equation 100].

It is shown. Here, the amplitude is set to 1 for simplicity. Here, if λ _n (≡c / f _n ), f _{d, n} (2V / λ _n ) is the Doppler frequency, c is the speed of light, and R is the target distance at time t = 0.
This received wave is down-converted and further mixed with a transmission wave f _n (n = 0, 1,... N−1) from the oscillator and observed at a distance gate number (ie, time delay τ) including the target. As a signal

I get the signal. R is a target distance. A pulse signal for each distance gating is obtained by performing A / D conversion on this signal and then performing pulse compression by autocorrelation processing with the transmission signal.

  Incidentally, it goes without saying that the time axis of this signal becomes a discrete value for each sampling period by A / D conversion, but here again, using the four parameters M, N, α, and SS described above, The time t of the received signal by distance gating is expressed by the following equation.

Where n = 0, 1,... N-1 m = 0, 1,... M-1 α = 0, 1 ss = 0, 1,.

T _chip represents the sampling interval time of the A / D converter. Therefore, if the [Equation 119] equation is substituted into the [Equation 118] and the phase-modulated transmission pulse is PMcode (t), then the [Equation 118] is

Can be expressed as However, the difference Δf of the frequency fn (= f0 + n · Δf) (n = 0, 1,... N−1) at each frequency step with respect to the transmission frequency is sufficiently small, and the Doppler frequency at each frequency step is equal. Because

As it has been the single value of the Doppler frequency at each frequency step f _d.
Further, if there are a plurality of Γ targets, γ (γ = 0, 1,..., Γ in [Equation 120] having different R (distance component) · f _d (velocity component) for each target. -1) is observed as a linear sum, the final observed waveform model equation is

Will be given in

4.6 Target Speed Estimation and Target Distance Estimation in Hybrid-CFS Method Next, the pulse processing obtained by the equation [Formula 122] is processed by SWW using IFFT (in fact, it is a discrete point, IDFT). A procedure for processing by the CFS method and deriving a target speed and a target estimated distance is shown in a block diagram.

  Hereinafter, with reference to FIG. 48, signal processing by the Hybrid-CFS method in the transmission / reception sequence of the present embodiment will be described. The signal processing system includes a Fourier transform unit 81, a phase compensation unit 82, an addition processing unit 83, and an inverse Fourier transform unit 84.

(I) FFT processing The FFT processing in the Fourier transform unit 81 will be described.
Similar to the multi-frequency step ICW, FFT is performed in the m direction as target relative speed detection processing to detect the speed. In the complementary coding band synthesis method of this example, the ss-m plane is considered for each frequency step and code, that is, for each variable n and α, and FFT is performed in the m direction to perform pulse Doppler processing. The difference from the multi-frequency step ICW is that not only the pulse peak but also the side lobe generated by the pulse compression is detected and the phase is corrected, and the FFT is performed on all sampling points ss. become. Although speed detection by non-uniform sampling between different codes is possible, it is preferable to perform FFT in consideration of problems such as calculation load and rounding of the pulse compression output. When expressed as an equation, as in [Equation 96],

It becomes. Here, k (= 0, 1,... M−1) is a frequency channel number. Again, when there are a plurality of targets within the same distance gate, it is represented by a linear sum of the formula [120], but fading occurs depending on the phase relationship, so that the Fourier transform output channel of each frequency step n Consider taking the sum of absolute values. Furthermore, since the amplitude value is the same in different complementary codes α,

Is an input value for detection threshold processing.
Further, it should be noted that different complementary codes are alternately transmitted from the [Equation 100], and the amplitude value | F (K, n, α, ss) | is a frequency channel number at each frequency step n.

In coherent integration, a peak is obtained. Further, the target Doppler frequency can be obtained by detecting the frequency channel number k _{peak at} which the output amplitude of the formula [122] reaches its peak. From the detected number kpeak, the detection target relative speed detectV is

Obtained from.

(Ii) Phase compensation Next, the phase compensation unit 82 adds complementary codes to cancel side lobes. At this time, the phase change due to the Doppler shift is corrected using the target speed detectV (ss) obtained previously. Here, for the sake of simplicity, when the detection speed is converted to the Doppler shift correction term, the following equation is obtained.

  Phase correction is performed using this equation. Since the phase term in which the Doppler shift term of the equation [120] exists is corrected, it is expressed by the following equation.

  Furthermore, in order to add α = 1 to the sign of α = 0, it is necessary to perform phase correction with respect to the time axis. This phase correction term is

  Represented as:

(Iii) ADD processing After performing correction using [Formula 127] and [Formula 128] of the phase correction term derived in the phase compensation of (ii), addition is performed between different complementary codes. Expressed in terms of equations, from [Equation 123], [Equation 127], and [Equation 128],

It becomes. However, the function F (m, n, α, ss) is replaced because the arrays of k and m are equivalent.

From this [Equation 130], the pulse interval is 2T _PRI , and further, an equation equivalent to the LFS method in which the target speed v = 0 is obtained from the phase correction or the multi-frequency step ICW is obtained. However, since the pulse peak originally appears in the last bit after the CPC pulse correlation process with the number of bits K, the pulse peak is further converted from the time delay τ corresponding to the target distance to the bit length K encoded as CPC and the sampling time T _chip. It appears at a point further delayed by the time multiplied by. Therefore, when the signal at the pulse peak point is expressed by an equation, the phase term f _d = 0 is obtained.

It becomes.

(Iv) -1 IFFT Processing In the Hybrid-CFS method, an inverse Fourier transform (hereinafter referred to as IFFT) is used for SWS of LFS for deriving a target estimated distance. Actually, since it is a discrete value for each sampling, discrete inverse Fourier transform (hereinafter referred to as IDFT) is performed. In (iii), the pulse of the side lobe 0 is obtained by the equation [Equation 130]. Here, in the equation [Equation 82] by the LFS method, it is considered that the phase compensation is performed, and the target velocity term in the phase term is set. If it is 0,

Since m is a function of only n that does not depend on the sampling time, and B / N = Δf, it can be seen that the correspondence with [Equation 131] is established. Therefore, in practice, the sampling point ss that is a pulse peak is detected as detectSS in the equation [131], and is extracted as an array of N rows and 1 column in the frequency step direction related to detectSS. If the pulse given by the equation [Equation 131] is subjected to IDFT processing, the equation [Equation 83] is obtained.

However, Nr = N, k = 0, 1,... Nr−1
Since this equation is an equation in which phase compensation is performed with the target speed v = 0, the peak point is obtained from the equation [84].

Satisfies to represent as Therefore, the estimated target distance R _cal is calculated from the equation [85].

As given. If there are multiple targets, the target estimated distance can be derived from the phase of the points by detecting the peak points in order from the highest and performing the same operation.

4.7 Target Speed Estimation and Target Distance Estimation in Complementary Multifrequency Step ICW Next, a procedure for deriving the target speed and target distance in the complementary multifrequency step ICW is shown in FIG.
FIG. 49 is complementary to FIG. 48 in which the IFFT processing ((iv) -1) in the Hybrid-CFS method is replaced with the super resolution method by the MUSIC (MUltiple SIgnal Classification) processing in the MUSIC processing unit 85. This is a multi-frequency step ICW. Therefore, (i), (ii), and (iii) will be omitted because they perform the same processing, and here, the SWW processing by the MUSIC method in (iv) -2 at the final stage will be described.
(Iv) -2 MUSIC processing The complementary multi-frequency step ICW is obtained from (Equation 131) from [Equation 131], so that an equation equivalent to the multi-frequency step ICW is obtained. The SWW processing by MUSIC processing, which is one, may be performed in the same manner as the multi-frequency step ICW in which the pulse interval given by the equation (3.14) is 2 TPRI.
As a result, target distance estimation for each speed range

Result can be obtained. Note that even if a plurality of targets are closer to the distance resolution c / 2Nr · (B / N) by the LFS IDFT at the same speed, the MUSIC process can perform target separation.

According to the present embodiment, the complementary function capable of completely canceling the distance side lobe of the pulse in the presence of multiple adjacent targets has no influence on other distance components. A very useful function is obtained for signal separation for. In particular, this is remarkable when focusing on signal separation of a plurality of targets in Hybrid-CFS using IFFT in which the target number cannot be estimated by SWW processing. On the other hand, since the SWW processing result in the MUSIC processing can detect a plurality of targets even for very weak distance sidelobe components, the target detection can be performed even for a minute pulse corresponding to the gradient from the peak point in the FFT processing. It can be seen from confirmation of the distance component of detectSS. It can be seen that it is important to set a signal separation threshold for multiple targets.
This is a transmission / reception sequence in the complementary coding band synthesis method that is expected to be applied to millimeter-wave in-vehicle radar by having both high resolution and low cost due to narrow receiver bandwidth. . As a result, IFFT used in Hybrid-CFS in SWW processing and MUSIC used in multi-frequency step ICW can be applied. Furthermore, in parameter setting considered at the time of application as an on-vehicle radar, target interference with high accuracy can be achieved with a small receiver bandwidth and less interference of distance side lobes.

5. Next, a modification applied to each embodiment will be described.

5.1 Randomization of Transmission Frequency Both the transmission sequence shown in FIG. 25 and the like in the first embodiment and the transmission sequence shown in FIG. 46 and the like in the second embodiment increase the transmission frequency in order. When a certain transmission frequency is reached, the process of reducing it to the original frequency is repeated, but it is not always necessary to increase or decrease the transmission frequency.

  That is, if there is an average period of time for each step during the predetermined time, the transmission frequency may be apparently changed randomly. For example, it is assumed that there are five steps in which transmission frequencies f0, f1, f2, f3, and f4 and five types of transmission frequencies are transmitted during the period Tx. At this time, during the period Tx, in addition to the process of increasing the transmission frequencies of the five transmission frequencies f0, f1, f2, f3, and f4 in order, the transmission frequency is decreased step by step or in a completely different order. Thus, the five transmission frequencies may be changed so as to change randomly in appearance. However, it is necessary to set so that five transmission frequencies exist the same number of times within a certain period Tx. In the second embodiment shown in FIG. 46, when the complementary code is transmitted continuously at two timings, the transmission frequency must be the same at the two consecutive transmission timings. .

In addition, when the transmission frequency is changed in an apparently random manner as described above, the apparently changed pattern may be different in a plurality of radar apparatuses. By preparing multiple patterns that change the frequency at random, and setting one of the multiple patterns to each of multiple radar devices, the possibility of interference with other radar devices is reduced. Can do.
That is, for example, in a situation where there are many automobiles equipped with the radar apparatus having the configuration described in the embodiment in the vicinity, if the transmission frequency change pattern of each radar apparatus is the same, by chance, a plurality of neighboring radars If transmission is performed from the apparatus at the same transmission frequency, the transmission pattern changes in the same manner, and thus the state of the same transmission frequency continues.
On the other hand, by making the pattern that appears to change randomly appear to be different patterns, even if the transmission frequencies overlap in multiple neighboring radar devices, they overlap temporarily and cause mutual interference. Defects can be minimized.

5.2 Example of Speed Detection when Stepping the Transmission Frequency Stepwise In the configuration example shown in FIG. 48, the transmitted signal is received and the distance component is detected using two complementary codes of Code1 and Code2. The configuration was shown. Here, as a configuration for detecting the velocity component, when simply considered, as shown in FIG. 51A, the velocity is detected from a signal obtained by collecting received signals at the same frequency position of Code1, and FIG. As shown in (b), it can be considered that the speed is detected from signals obtained by collecting received signals at the same frequency position of Code2. In FIG. 51, the horizontal axis represents time, and the vertical axis represents the reception level.

When the velocity component is detected only from the signals of the respective codes shown in FIGS. 51A and 51B, there may be a case where sufficient detection accuracy cannot be obtained.
The following equation 137 is an example in which velocity detection is performed from all sample positions, but in the case of the configuration of FIG. 48, such velocity detection is impossible in principle.

  The following Expression 138 is an expression for performing speed detection from Code1, and Expression 139 is an expression for performing speed detection from Code2.

Here, the detection accuracy can be improved by detecting the velocity component from the received signal of Code 1 and the received signal of Code 2 at the same frequency position.
FIG. 51 (c) is a diagram showing a state detected from the received signal of Code 1 and the received signal of Code 2 at the same frequency position, and the following equation 140 is an equation for performing speed detection from Code 1 and Code 2. It is.

  Formula 141 shows the conditions of Formula 137 to Formula 140.

FIG. 52 and FIG. 53 show the speed detection state when the speed detection is performed as described above. 52 and 53, the horizontal axis represents the velocity component. FIG. 52 shows an example in which speed detection is performed only from the received signal of Code 1 at the same frequency position. Even when the speed detection is performed only from the received signal of Code 2 at the same frequency position, it is the same as FIG. FIG. 53 shows an example in which speed detection is performed from the received signal of Code 1 and the received signal of Code 2 at the same frequency position.
As can be seen by comparing the two, when speed detection is performed only from the received signal of Code 1 at the same frequency position shown in FIG. 52, many peaks of the same level appear and the true peak can be discriminated. The speed detection may be detected later.
On the other hand, when speed detection is performed from the received signal of Code 1 and the received signal of Code 2 shown in FIG. 53, the peak position having the highest level (f0 in the figure) is one place. Therefore, it is possible to detect an accurate speed from the peak position. That is, an integer value that provides the largest output during one transmission cycle (fTs cycle) is set as a true value.

FIG. 50 shows an example in which a block for speed detection is added in the configuration shown in FIG. As shown in FIG. 50, the signal subjected to the FFT processing by the Fourier transform unit 81 is supplied to the speed detection unit 86, and the received signal of Code1 and Code2 at the same frequency position are calculated from the calculation formula shown by Formula 140. The received signal is used to perform speed detection processing to obtain a speed component.
The speed detection accuracy can be improved by detecting the speed component in this manner.

As shown in FIG. 22, in the first embodiment to which the multi-frequency NL-SWW method is applied, by applying [Equation 35] and [Equation 36], the nonlinear frequency step interval dF _n is It was stated that it can be controlled only by the nonlinear parameter σ.
On the other hand, a certain distance of the output distance waveform may be constrained to determine a cubic function parameter so that a good waveform can be obtained. The restriction of a certain distance in the output distance waveform is to determine the amplitude level ε at a certain distance.

Hereinafter, an example of restraining at a certain distance of the output distance waveform will be described.
Here, FIG. 54 shows a transmission frequency sequence in the multi-frequency NL-SWW method. The transmission frequency sequence in FIG. 54 is basically the same as the transmission frequency sequence shown in FIG. FIG. 54 (a) shows that the frequency increase of the sub-pulse is repeated every period m = 0, m = 1,..., M = M−1, and FIG. The non-linear frequency step is shown. _{B N} shown in FIG. 54 (b) is a full bandwidth by frequency steps.

  FIG. 55 shows a reception configuration based on the multi-frequency NL-SWW method. A signal received by the reception antenna 101 is a mixer 102, and local frequencies f0, f1,..., FN−1 from the local signal generation unit 103 are shown. Are multiplied. Further, the modulated wave from the LFM modulated wave generator 104 is also multiplied. The output of the mixer 102 is subjected to sub-pulse compression processing by the sub-pulse compression unit 105, and the output is supplied to the Doppler frequency estimation processing unit 106 to perform m-direction Fourier transform processing. The speed detection unit 108 detects the speed based on the output of the Doppler frequency estimation processing unit 106, and the Doppler frequency correction processing unit 107 performs correction processing. Further, the band synthesis processing unit 109 performs band synthesis processing (IDFT) on the output of the Doppler frequency correction processing unit 107. This band synthesis process is an n-direction inverse Fourier transform process. A distance waveform is obtained as an output of the band synthesis processing unit 109.

  By processing with the configuration shown in FIG. 55, speed can be obtained by Doppler frequency estimation processing, and distance error can be compensated after Doppler frequency correction processing. The distance waveform which is the final output is obtained by multiplying the output subjected to the band synthesis process by the band synthesis processing unit 109 by the pulse compression output of the subpulse.

  The constraint method applied to the multi-frequency NL-SWW method will be described. This constraint method constrains the relative amplitude value of a specific distance of the band synthesized output in the multi-frequency NL-SWW, and satisfies the distance waveform. The nonlinear function parameters are determined by the nonlinear least square method so that is obtained. Here, the non-linear function is assumed to be the following third-order polynomial that is an odd function targeting the center of the frequency step.

  n is the frequency of each subpulse. From the conditions of the start point, end point, and intermediate point of the frequency step, the conditions shown in the following [Expression 143], [Expression 144], [Expression 145], and [Expression 146] are satisfied.

At this time, the total bandwidth by the frequency step is set to B _N (see FIG. 54).
The band synthesis output in the multi-frequency NL-SWW method is expressed as follows.

At this time, X (n) is a signal after being processed by the Doppler frequency correction processing unit 107 shown in FIG. 55, and R is a directivity distance of band synthesis processing (IDFT). When the relative amplitude value at a position (constraint distance) separated from the distance R ₀ where the directivity distance R and the target distance R _{d of the} band synthesis processing (IDFT) coincide with each other by a distance Δr is constrained as ε,

It becomes.
Substituting the equation 142, the constraint condition equation is expressed as follows.

At this time, X is the number of constraint conditions. From the formula [150], it is clear that the frequency step non-linearization method based on the constraint condition is established even if the target distance _Rd is unknown. Here, if the number of constraint conditions is 3 (x = ₀ , ₁ , ₂ ), the four unknowns (P ₀ , P ₁ , P ₂ , p ₃ ) are determined from the formula [Equation 144], and [ [Formula 144] Formula, [Formula 145] Formula and Constraint Condition Formula [Formula 150] Formula A total of five formulas, and the coefficients (P ₀ , P ₁ , P ₂ , p ₃ ).

In this example, the distance side lobe constraint and the pulse compression null constraint as shown in FIGS. 56 and 57 are adopted for the distance Δr giving the constraint condition and the relative amplitude value ε in consideration of the following. The constraint distance Δr of the constraint conditions 1, 2, and 3 in FIG. 57 is the position of the constraint conditions 1, 2, and 3 in FIG.
The constraint conditions in FIG. 57 are the following two conditions.

(1) The relative amplitude value of the band synthesis output is constrained to a small value within the main lobe of the pulse compression output.
(2) The relative amplitude value of the band composite output at a distance where the pulse compression output becomes null (referred to as pulse compression null distance) is increased.

  As described above, the distance side lobe can be reduced in the distance waveform that is the product of the pulse compression output and the band synthesis output.

Here, an example of simulation under this constraint condition will be described.
Assuming millimeter-wave vehicle-mounted radar, the following radar parameters were adopted.
-Transmission frequency f: 76.5 GHz
・ Pulse repetition period TPRI: 2μs (maximum instrument distance: 300m)
-Subpulse bandwidth b: 80 MHz (distance resolution: 1.875 m)
・ Number of frequency steps N: 8 (maximum speed field of view: ± 220.588 km / h)
Reference frequency step width Δf: 70 MHz (distance resolution of proposed method: 0.263 m)
Occupied bandwidth B: 570 MHz (= b + B _N )
-Total observation time Ts: 4.096 ms (speed resolution: 1.723 km / h)
・ Number of targets: 1 (target distance _Rd : 200 m, target speed V: 200 km / h)
Constraint conditions (Δr, ε): Condition 1 (0.368 m, 0.001)
Condition 2 (0.632m, 0.001),
Condition 3 (2.025m, 0.1)

Under this condition, the band synthesis processing output when the nonlinear function derived based on the constraint conditions shown in FIGS. 56 and 57 is used is shown in FIG. 58, and the distance waveform (band synthesis processing output and sub-pulse compression) as the final output is shown in FIG. 59) is shown in FIG.
As can be seen from FIG. 59, in the processing in this example shown by the solid line, a distance resolution of about 7 times (−3 dB) is obtained as compared with the pulse compression (broken line) of the subpulse, and the distance side lobe is generated in all distance ranges. It was found to reduce to -20 dB or less.

In the description so far, the coefficient of the nonlinear function is obtained so that the relative amplitude value of a specific distance of the band synthesis (IDFT in the frequency step direction) output in the multi-frequency NL-SWW is constrained and a distance waveform satisfying the constraint is obtained. P _0, for a plurality of unknowns is _{P 1, P 2, P 3} , an example of performing estimation using the least squares method.
On the other hand, an example in which coefficients (unknown numbers) P ₀ , P ₁ , P ₂ , and P ₃ of the nonlinear function are expressed by one parameter P ₀ and estimation is performed from the constraint equation by the least square method will be described below. To do.

・ Frequency step nonlinearization method by constraint condition in multi-frequency NL-SWW (Estimation of nonlinear function by one parameter)
In this example, in the multi-frequency NL-SWW, the relative amplitude value of a specific distance of the band synthesis output is constrained, and the nonlinear function parameter is determined by the nonlinear least square method so that a distance waveform that satisfies the constraint is obtained. Here, the nonlinear function is the following cubic polynomial that is an odd function with the center of the frequency step as a target.

From the conditions of the start point, end point and intermediate point of the frequency step, B _N is the total bandwidth in frequency steps.

  Further, in order to express the coefficient of the nonlinear function with one parameter, from the formulas [154] and [155], ½ {F (N−1)} − F {(N−1) / 2} It becomes as follows.

  Further, from 1/4 {F (N-1)}-F {(N-1) / 2}, the following is obtained.

Therefore, from [Equation 151], [Equation 152], [Equation 157], and [Equation 159], the nonlinear function is expressed by the following equation, with only the coefficient P ₀ .

  Further, the band synthesis output in the multi-frequency NL-SWW is expressed as follows.

At this time, X (n) is a signal after Doppler correction processing is performed with the configuration shown in FIG. 55, and R is the directivity distance of the IDFT. From this Figure 56, the relative amplitude values of the IDFT-oriented distance R and the target distance R _d is separated from the distance R ₀ that matched the distance Δr position of (synthetic band treatment) (constraint length) as shown in FIG. 57 epsilon When restrained,

When the [Equation 160] equation is substituted, the constraint condition equation is expressed by the following equation. At this time, X is the number of constraint conditions.

It is clear from the formula [164] that the frequency step non-linearization method based on the constraint condition is established even if the target distance _Rd is unknown. Here, when the number of constraint conditions is 3 (x = 0, 1, 2), the unknown P _{0 is} obtained by the nonlinear least square method from the three constraint conditions (Equation 164) for the unknown P ₀ . . [Expression 153] _P 3 = 0 is determined from the equation [Expression 157] from equation _{P 0,} the coefficients of the nonlinear function by using the number 159] Formula _{(P 0, P 1, P} 2, P 3) Ask.

  In each of the examples described so far, the example in which the constraint is performed by the cubic function has been described. However, the constraint may be performed by a higher-order function (odd function) such as a quintic function.

In addition, the processing configuration example of the radar apparatus described in each of the above-described embodiments is a preferable example, and is not limited to the illustrated configuration.
Further, the processing described as the first embodiment and the processing described in the second embodiment may be combined. That is, the processing configuration in which the step interval for stepping the transmission frequency described in the first embodiment is a non-linear step that is a cubic function with the upper limit and lower limit of the frequency fixed is described in the second embodiment. It may be combined with the processing configuration that assigns the two complementary signals described in the above as continuous transmission pulses at the same frequency. By combining in this way, it is possible to obtain a radar apparatus that can perform highly accurate detection having both effects.

DESCRIPTION OF SYMBOLS 11 ... Transmission wave generation part, 12 ... Pulse conversion process part, 13 ... Transmission antenna, 14 ... Reception antenna, 15 ... Phase detection part, 16 ... Mixer, 17 ... Low pass filter, 18 ... Analog / digital converter, 19a-19c ... Fourier transform unit, 20 ... target distance / velocity calculation unit, 21 ... pulse compression processing unit, 22 ... modulation pulse generation unit, 31 ... sub-pulse generation unit, 32 ... mixer, 33 ... oscillator, 34 ... circulator, 35 ... transmission / reception antenna , 36 ... mixer, 37 ... low pass filter, 38 ... analog / digital converter, 39 ... Fourier transform unit, 40 ... mixing processing unit, 41 ... reference signal generation unit, 42 ... Fourier transform unit, 43 ... spectrum shift unit, 44 ... combine unit, 45 ... inverse Fourier transform unit, 51 ... LFM modulation unit, 52 ... circulator, 53 ... oscillator, 54 ... transmission / reception Antenna, 55 ... sub-pulse compression unit, 56 ... Doppler frequency estimation processing unit, 57 ... Doppler frequency correction processing unit, 58 ... synthesis band processing unit, 61 ... oscillator, 62 ... up converter, 63 ... high frequency switch, 64 ... circulator, 65 Transmission / reception antenna, 66 ... Down converter, 67 ... Fourier transform unit, 68 ... Frequency smoothing unit, 69 ... Detection unit, 71 ... Reception antenna, 72 ... LFS detection unit, 72a ... Oscillator, 72b ... Mixer, 72c ... Band pass filter 73 ... Complex IQ detector, 73a ... Oscillator, 73b ... Mixer, 73c ... Phase shifter, 73d ... Mixer, 73e ... Low pass filter, 73f ... Low pass filter, 74 ... Correlation processor, 74a ... Analog / digital converter, 74b ... Matched filter, 74c ... Analog / digital converter 74d ... matched filter, 81 ... Fourier transform unit, 82 ... phase compensation unit, 83 ... addition processing unit, 84 ... inverse Fourier transform unit, 85 ... MUSIC processing unit, 101 ... receiving antenna, 102 ... mixer, 103 ... local signal generation , 104... LFM modulated wave generator, 105... Sub-pulse compressor, 106. Doppler frequency estimation processor, 107. Doppler frequency correction processor, 108... Speed detector, 109.

Claims (2)

Pulses within the frequency spread transmitted pulse, N (N is an integer of 2 or more) is transmitted by step the frequency using the frequency of Rutotomoni, the frequency step M times within one observation time (M is In a radar apparatus that repeatedly transmits an integer of 2 or more and receives the reflection of the transmission pulse from the target,
A set of transmissions in which signals using two complementary codes capable of suppressing the distance side lobe are continuously transmitted at the same frequency within a fixed time by combining two complementary codes after addition after autocorrelation Assigned as a pulse,
A radar device for frequency stepping a set of transmission pulses by a signal using the two complementary codes ;
A set of transmission pulses that are continuously transmitted at the same frequency within a certain period of time is a signal using a complementary code for a pulse repetition time that secures a time until the transmission pulse is reflected from the target and received. Are secured and sent in order,
The pulse repetition time is given by the following formula (where T _PRI is the pulse repetition time, R _max is the maximum instrument distance required for the radar, and c is the speed of light).
T _PRI ≧ (2R _max / c)
By satisfying the above, a radar in which the value of M corresponding to the number of transmission pulses of the same code at the same frequency within the one observation time becomes smaller than that in the case where one set of transmission pulses is not assigned. Device,
The received pulse is subjected to pulse compression processing and pulse Doppler processing, and the phase modulation pulses of the two complementary codes are shifted and transmitted / received with respect to the signals subjected to the pulse compression processing and pulse Doppler processing. A radar apparatus characterized by performing a correction process for correcting a phase error caused by a difference and a distance error that occurs when there is a relative speed with a target and performing band synthesis after addition . The radar apparatus according to claim 1 , wherein
It has a velocity field correction unit that performs a velocity field correction process for correcting the ambiguity of the velocity field from the phase gradient in the detection signal after pulse compression of the signal using two consecutive complementary codes received. Radar device.

Images