JP2013178263A - Radar of direct sequence spectrum spreading system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To quickly accurately detect a Doppler frequency shift of a target by a radar of a direct sequence spectrum spreading (DSSS) system.SOLUTION: A radar of a DSSS system includes: means which transmits a transmission signal including a prescribed code sequence to one or more targets; means which receives signals reflected by the one or more targets; means which constitutes vectors v=(v(t) to v(t))of sample values of reception signals at mutually different times tto tand calculates correlation vectors rand rfrom the vectors and the reception signals at mutually different times; means which combines the correlation vectors to constitute a pseudo spatial average covariance matrix R; and means which uses (RR)to calculate a Doppler frequency shift from a pseudo spectrum or an algebraic equation.

Description

  The disclosed embodiment relates to a direct sequence spectrum spreading (DSSS) radar or the like.

  In general, a radar is also referred to as a radio wave detector, which radiates a radio wave to a target body (target) and receives and analyzes a reflected wave from the target body to detect the position and speed of the target body. In the following, an explanation will be given taking the case of in-vehicle use as an example. The radar is classified into long range radar (LRR) and short range radar (SRR) when classified by the detection distance. The former specifies a target object over a wide range, for example, 100 meters, and mainly uses a frequency modulation (FM) method. The latter specifies a target object in a narrow range such as several tens of meters, and uses FM method or SS (spread spectrum method). In particular, the latter is expected for a wide range of uses such as controlling an airbag.

Fig. 1 shows a system block diagram (simulation model) of DSSS radar. The radar first generates a PN code having an appropriate length and timing by the code generator 102 and uses it as a system reference signal (baseband) that characterizes the signal transmitted from the radar. If the radar device number is DN, the PN code length is N, the pulse waveform carrying the code is p (t), and the chip duration of the pulse is _Tc , this reference signal is expressed as follows: The

リファレンス信号v_TXは、高周波発生器(発振器、ＲＦ−ＯＳＣ)１０４で生成されHYB１０５で分波されたキャリア信号によりミキサ１０６でアップコンバートされる。アップコンバートされた送信信号は、高出力増幅器又はハイパワーアンプ(HPA: High Power Amplifier)１０８で増幅され、送信アンテナA_Tからプローブ信号として送信される（即ち、キャリア信号をシステムリファレンス信号で変調したものがプローブ信号である）。送信信号は例えばミリ波で送信される。図示の例ではプローブ信号と目標体との相互作用は、伝搬遅延時間τ、ドップラ周波数偏移f_d及び加法性白色雑音(AWGN: Additive White Gaussian Noise)で表現される。即ち、目標体がレーダーの視野(FOV: Field of View)内に存在すれば、プローブ信号は視線方向相対距離（以下、距離と記す）に応じた時間遅延τと、目標の視線方向相対速度(以下、速度と記す)に応じたドップラ周波数偏移f_dと、雑音の影響とを受けた後に、レーダーで受信される。

The reference signal v _TX is up-converted by the mixer 106 by the carrier signal generated by the high frequency generator (oscillator, RF-OSC) 104 and demultiplexed by the HYB 105. The up-converted transmission signal is amplified by a high power amplifier or high power amplifier (HPA) 108 and transmitted as a probe signal from the transmission antenna _AT (that is, the carrier signal is modulated with the system reference signal). Is the probe signal). The transmission signal is transmitted by millimeter waves, for example. In the illustrated example, the interaction between the probe signal and the target body is expressed by propagation delay time τ, Doppler frequency shift f _d, and additive white Gaussian noise (AWGN). That is, if the target body is in the field of view (FOV) of the radar, the probe signal has a time delay τ corresponding to the relative distance in the line-of-sight direction (hereinafter referred to as distance) and the target line-of-sight relative velocity ( Hereinafter, it is received by the radar after receiving the Doppler frequency shift f _d corresponding to the speed) and the influence of noise.

Reception signal received by the receiving antenna A _R is a low noise amplifier: is amplified by (LNA Low Noise Amplifier). The amplified received signal is down-converted by the mixer 124 to which the carrier signal (carrier wave) demultiplexed from the carrier distributor 105 is input, and is input to the mixer 132 of the baseband demodulator 126. The mixer 132 multiplies the down-converted signal by the PN code (demodulation SS code) delayed by the delay unit 130 for a certain period τ _n . The output signal of the mixer is integrated by the integrator 134 over the PN code unit length _Tf , and a correlation value is output. Therefore, the baseband signal v _RX (t) representing the output signal of the mixer 132 is

で表現される。ここで、k(1,…,K)はK個の目標体の内のｋ番目の目標体を特定する目標番号、a_kは信号振幅、ω_ｄ ^ｋはドップラ角周波数偏移、τ_kは遅延時間、n(t)は加法性ガウシンアンノイズ(AWGN)を表す。

It is expressed by Here, k (1,..., K) is a target number that identifies the kth target body among the K target bodies, a _k is the signal amplitude, ω _d ^k is the Doppler angular frequency shift, and τ _k is The delay time, n (t), represents additive Gaussian noise (AWGN).

FIG. 2 schematically shows the relationship between the SS code, the code length (T _f ), and the chip period (T _c ). In FIG. 2, the data (DATA) multiplied by the SS code is generally drawn so as to become a transmission signal, but it is not essential to multiply the data.

FIG. 3 shows an example of auto-correlation output. Specifically, the delay unit 130 sets the delay amount τ _n to various values, and plots the output of the integrator 134. It corresponds to that. The two sharp peaks mean that there are at least two target objects with different distances, and their delay periods τ are about 320 chips and 340 chips, respectively. The detector 136 in FIG. 1 detects the position of such a peak, and the detection result is given to the delay unit 130 and the counter 138 as a synchronization timing for the reflected signal from each target. In order to scan such a sharp correlation peak over a predetermined period, the delay unit 130 shifts the SS code, for example, by one chip (generally, a constant multiple of one chip period-for example, 1/3-) to mix the mixer. Input is given to 132. A PN code having a certain delay amount (mT _c ) is multiplied by the received signal and integrated over the code length T _f . For example, when the received code and the demodulated code from the kth target are synchronized including the delay, the summation corresponding to the kth target in the latter half of equation (2) is 1, and the received signal related to this target is It can be expressed as follows (the component out of synchronization is -1 / N and is substantially 0).

v _RX (t) = a _k exp [jω _d ^k (t−τ _k / 2)] + n (t) (2) ′
When a peak corresponding to the target body is detected, the delay unit 130 fixes the delay amount to a value that gives the peak (detection of synchronization timing), and synchronization for the target body is ensured. When there are a plurality of target bodies (k = 1,..., K), the synchronization timing is detected for each target body, and the same processing as described above is performed in the demodulator 126 for each synchronization timing. In the figure, “× K” for the demodulating unit 126 indicates that K processing blocks such as the demodulating unit 126 are used in parallel in order to easily understand the operation (therefore, accordingly). This does not prevent the implementation of using one demodulator in time division).

Next, the speed or distance of the target body is detected based on the received signal in which synchronization is ensured. Of these, the distance d _k of the _kth target is calculated by the following equation, where c is the speed of light:
d _k = (cτ _k ) / 2 ≒ (cmT _c ) / 2
Therefore, the distance resolution is proportional to T _c .

By the way, the received signal has a Doppler frequency corresponding to the speed of the target body. Radar is a measuring device, can not know the value of the Doppler frequency f _d in advance. That is, the speed of the target body cannot be known with the output signal of the integrator 134 as it is. Therefore, in the conventional method, the received signal is sampled over a period 1 / f _d (= 2π / ω _d = T _d ) or more derived from the smallest predicted Doppler frequency, and then Fourier transformed by the FFT 140. The Doppler frequency f _d was estimated. Therefore, this period becomes longer as the Doppler frequency is smaller.

  For example, Non-Patent Document 1 describes a system that estimates Doppler frequency using FFT with a DSSS radar.

Masahiro WATANABE, et al., “A 60.5GHz Millimeter Wave Spread Spectrum Radar and the Test Data in Several Situations”, Proc. On IEEE Intelligent Vehicle Symposium, 2002pp.87-91

When the Doppler frequency f _d is a relatively small value, the period T _d = 1 / f _d is relatively long. Since the speed of the target body can vary between large and small, conventionally, for example, the received signal is sampled over a period corresponding to the smallest relative speed and is processed by FFT to estimate the Doppler frequency f _d .

FIG. 4 is a diagram in which the received samples after baseband demodulation are overlaid along the time axis for each of the _two target bodies (T ₁ , T ₂ ). Each data sample is obtained every integration time (T _f ) of the integrator 134, and in the example shown, T _f = 0.0002 ms. One target (T ₁ ) is located 10 meters away and is moving at 6 km / h. The other target (T ₂ ) is located 11 meters away and is moving at 12 km / h. In this case, it can be seen that the received sample value fluctuates with a period of about 0.7 ms for the target body (T ₂ ). Therefore, if sampling is performed over a period of about 0.7 ms and the result is Fourier-transformed, the Doppler frequency for the target body T _{2 and} its speed can be found. The target body T ₁ needs to be sampled for a period longer than 0.7 ms.

In particular, in a vehicle-mounted radar or the like, generally, a target speed at a short distance must be estimated with high speed and high accuracy from the viewpoint of safety. However, the method using FFT is the time required to detect the speed of low-speed moving targets, especially in scenes where targets moving at various speeds are mixed (images of crowded highways where the distance between vehicles is extremely short). However, since the time required for speed detection of other high-speed moving targets is limited, there is a big problem.
Therefore, from the viewpoint of realizing high-speed and high-accuracy measurement, it is not preferable to employ a Doppler frequency measurement method based on FFT that requires long-time data collection.

Theoretically, since the received signal component is obtained for each integration time NT _c (= T _f ) necessary for demodulating the PN code, another conventional method for estimating the Doppler frequency focuses on this point. Specifically, for the composite signal of equation (2), after synchronizing with the component for the kth target, (v _RX (t) = a _k exp [jω _d ^k (t−τ _k / 2)] + n (t)), the Doppler frequency f _d (= ω _d / 2π) is derived from the m-th value among the sample values obtained every T _f according to the following equation.

ここで、n_R(t)、n_I(t)はn(t)の実部と虚部である。

Here, n _R (t) and n _I (t) are a real part and an imaginary part of n (t).

In the above method, when the SN ratio is sufficiently large, the Doppler angular frequency ω _d can be quickly derived. However, the above formula greatly depends on the noise n (t) = n _R (t) + jn _I (t) at one time point (t = mT _f ), so that the SN ratio at the measurement target time point is low However, the accuracy may be greatly deteriorated. Also, for example, | mT _f −ω _k / 2 | → π / 2ω _d ^k , 3π / 2ω _d ^k ,. . . In such cases, the accuracy is significantly degraded due to the fact that the cosine function of the denominator that makes up the argument of the arc tangent function approaches zero, or that the tangent function of the numerator diverges to positive or negative infinity. There is a risk. Furthermore, there is a concern that accuracy may deteriorate when | mT _f −ω _k / 2 | → 0.

  The problem of the embodiment is to quickly and accurately estimate the Doppler frequency shift of the target with a direct sequence spread spectrum radar.

The direct sequence spread spectrum (DSSS) radar according to the embodiment is
Means for transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Means for receiving signals reflected by one or more target objects;
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from received signals at different times;
Means for combining the correlation vectors to form the pseudo-space average covariance matrix R;
And a means for calculating the Doppler frequency shift from a pseudo spectrum or an algebraic equation using (RR ^H ) ⁻¹ .

  According to the embodiment, it is possible to quickly and accurately estimate the Doppler frequency shift of the target body with the DSSS radar.

A detection and ranging apparatus used in an embodiment of the present invention includes one or more transmitters that emit a probe signal modulated in an appropriate manner from a sensor toward an object to be measured (hereinafter, a target); One or more receivers that receive and demodulate the probe signal reflected from the target to obtain the necessary baseband signal, compared to the time required for the target viewing angle and speed to change A means for sampling the baseband signal at a sufficiently short time interval and converting it into a discrete signal sequence, and a phase shift of two or more sample values at different times in the sample signal sequence, to a target relative line-of-sight speed. and means for calculating a Doppler frequency shift w _d the originating.

In one embodiment of the present invention, sum and difference signals are generated for baseband signals in two or more different time slots sampled over an appropriate time interval δt. The velocity of the target body is calculated by obtaining the Doppler frequency from the ratio of the sum signal and the difference signal using an inverse trigonometric function. Since only sample values obtained at several different times are required to obtain the Doppler angular frequency ω _d (= 2πf _d ), the Doppler frequency can be obtained very quickly. In addition, since this method uses sample values at a plurality of points in time, it is not greatly affected by noise at a single point in time.

  One embodiment of the present invention prepares a vector having a plurality of received samples obtained at different points in time as components, and uses the concept of spatial averaging and the concept of a propagator method in array signal processing. To do. This makes it possible to estimate the speed of each target object with high accuracy even when target objects having different speeds exist at the same distance.

  For convenience of explanation, the present invention is described in several embodiments. However, the division of each embodiment is not essential to the present invention, and two or more embodiments may be used as necessary.

  FIG. 5 is a block diagram showing a radar according to an embodiment of the present invention. FIG. 5 shows a code generation unit 502, an oscillator 504, a carrier distribution unit 506, a mixer 508, a high power amplifier (HPA) 510, a low noise amplifier (LNA) 512, a mixer 514, a demodulation unit 516, a delay adjustment unit 520, and a mixer 522. , An integrator 524, a peak detector 526, a counter 528, a Doppler frequency estimator 530 and a control and signal processor 532 are depicted.

The code generation unit 502 generates a predetermined code sequence constituting the system reference signal. Any appropriate sequence that can be used for spread spectrum may be used as the code sequence. For example, a PN sequence, an M sequence, a Gold code sequence, or the like may be used. The code length may be set as appropriate according to the application. In general, it is preferable to set a long code length (T _f ) from the viewpoint of suppressing interference between codes and increasing measurement accuracy. From the viewpoint of improving the accuracy of distance measurement, it is preferable to shorten the chip period _Tc .

  The oscillator 504 prepares a carrier frequency for carrying the probe signal. Although the present invention is not limited to a specific frequency, as an example, various appropriate values may be used for the carrier signal depending on the application of the radar, for example, values such as 76 GHz, 38 GHz, and 24 GHz may be used. Good.

  The carrier distribution unit 506 gives the carrier frequency prepared by the oscillator 504 to the transmission side and the reception side.

  The mixer 508 multiplies the code sequence generated by the code generation unit 502 and the carrier frequency signal (modulates the carrier signal with the system reference signal), and prepares a transmission signal (probe signal).

A high power amplifier (HPA) 510 amplifies the power of the transmission signal, and the amplified signal is transmitted from the transmission antenna _AT .

Although not explicitly shown for simplification of illustration, in performing the simulation, a functional block (element for introducing propagation delay τ, Doppler frequency shift f _d is introduced as shown on the left side of FIG. Elements that introduce noise components such as additive Gaussian noise).

  The low noise amplifier (LNA) 512 appropriately amplifies the power of the received signal.

  The mixer 514 converts the high frequency received signal into an appropriate baseband signal.

  The demodulation unit 516 includes a delay adjustment unit 520, a mixer 522, an integrator 524, and a peak detector 526. In the figure, regarding the demodulator 516, “× K” indicates that there may be K processing blocks such as the demodulator 526 (K = 1, 2,...). As described above, K is the total number of target bodies to be detected. Each demodulator 516 uses the synchronization timing for each target body. Note that it is not essential to the present invention that K demodulation units 516 are prepared in parallel when there are a plurality of target bodies. This is because the same result can be obtained by using one demodulator 516 K times while changing the synchronization timing.

The delay adjusting unit 520 of the demodulating unit 516 gives an appropriate delay to the system reference signal according to an instruction from the counter 528 (for example, a constant multiple of the chip period _Tc ), and a demodulation code sequence (SS code). Generate.

  The mixer 522 calculates the product of the received signal and the demodulation code sequence.

The integrator 524 integrates the output from the mixer 522 over a predetermined period (typically, a period of 1T _f that is one unit of the code sequence). The integral value obtained here is the correlation between the demodulated code sequence generated by delaying the reference signal and the baseband component of the received signal (including the system reference signal that has received a delay corresponding to the distance to the target). Represents a value.

  The peak detector 526 detects the peak of the correlation value exceeding a predetermined threshold value, and obtains the time position of the delay corresponding to the peak. Each peak generally corresponds to each of the targets. As described above, by taking into account the delay time corresponding to the peak position of the correlation value generated by a certain target body and establishing demodulation of the received signal, demodulation is performed from among the reflected signals from other targets. Only the reflected signal from the target can be separated and properly received.

  The counter 528 performs control so that the delay amount set by the delay adjustment unit 520 changes over a predetermined period. At the same time, the counter 528 sends information on the delay amount set at each time point to the signal processing unit 532, and the peak of the correlation value. Is detected, the signal processing unit uses it for calculating the distance of the corresponding target.

Doppler frequency estimation unit 530 based on the received signal in synchronization with the individual target body, and calculates the Doppler frequency shift f _d in the manner described below.

  The signal processing unit 532 calculates the distance to the target body from the delay information to the reference signal that gives a peak to the correlation value, and calculates the speed of the target body based on the Doppler frequency calculated by the Doppler frequency estimation unit 530. To do.

  Next, the operation for estimating the Doppler frequency by the Doppler frequency estimation unit 530 will be described. In order to simplify the explanation, there is only one target object, for example, a vehicle, in front of the radar (that is, the DOA: Direction Of Arrival (DOA) of the probe signal reflected by the target is 0 degree). It is assumed that the received signal strength is normalized to 1. In a more complicated situation, such as when multiple targets are present at the same distance, a more accurate estimation can be realized using the second embodiment of the present invention. Since the present invention is not constrained by the number of reception antennas, the number of demodulation devices, and the like, for example, one antenna may be shared by the transmitter and the receiver, or separate antennas may be prepared.

The carrier wave generated by the high frequency generator 504 is modulated by the mixer 508 by the system reference signal prepared by the code generation unit 502, and transmitted as a probe signal via the high power amplifier 510 and the transmission antenna _AT . The probe signal is reflected by the target object, it reaches the receiving antenna A _R. Probe signal reaches the receiving antenna A _R has a propagation delay τ with respect to the probe signal before transmission, has a frequency deviated by the Doppler frequency shift f _d, is further additive white noise n (t) It has been added. The received signal is down-converted by the carrier wave by the mixer 514 and converted into a baseband signal, and the mixer 522 multiplies the reference signal from the code generation unit 502 with a demodulation code generated by giving a certain delay to the delay unit 520. Are integrated by an integrator 524. By calculating the multiplication in the mixer 522 and the integration in the integrator 524 while changing the delay amount τ _n over a predetermined period, the delay amount up to the target body and the delay amount set in the delay unit 520 coincided. When the correlation value output shows a peak, code synchronization is established.
At this time, the peak of the correlation value is detected by the detector 526, and the delay amount set in the delay unit 520 when the peak is given is given to the signal processing unit 532 through the counter 528. Thereafter, the reflected wave from the target object to be detected is demodulated in a state where code synchronization is established, and the received signal (basically, a sine wave having a Doppler frequency corresponding to the target speed) is obtained. 530 is input.

The Doppler frequency estimator 530 receives two sample values separated by an appropriate time interval δt from among samples representing the received signal. These are expressed by the following equations when the sampling time is t ₁ .

v (t ₁ ) = exp (jω _d t ₁ ) + n (t ₁ ) (4)
v (t ₂ ) = exp (jω _d t ₂ ) + n (t ₂ ) (5)
However, t ₂ = t ₁ + δt. δt may be set to any appropriate value, but as an example, may be set to an integral multiple of the period of one code length (T _f ). Also, Δt may be defined as a time interval that is sufficiently shorter than the time during which the target line-of-sight angle or line-of-sight speed changes.

Next, the Doppler frequency estimation unit 530 calculates the ratio of the sum and difference of v (t ₁ ) and v (t ₂ ).

但し、Δn及びΣnは次式で定義される量である。

However, Δn and Σn are quantities defined by the following equations.

である。よって、|δt|がπ/2ω_d ^kの奇整数倍から十分に離れた値になる様に選んでおけば（cos(ω_dδt/2)が０にならないように、また、tan(ω_dδt/2)が発散しないようにしておけば）、SN比が十分に大きい場合、目標体の速度に由来するドップラ角周波数は、次式により得られる。

It is. Therefore, if | δt | is chosen to be a value sufficiently away from an odd integer multiple of π / 2ω _d ^k (so that cos (ω _d δt / 2) does not become zero and tan (ω _{If d} δt / 2) does not diverge, the Doppler angular frequency derived from the velocity of the target body is obtained by the following equation when the SN ratio is sufficiently large.

一般に、逆正接関数tan^-1(x)は|x|>π/2で精度が劣化するので、|ω_dδt/2|をモニタし、例えば、|ω_dδt/2|＜π/4の場合には (8)式が利用され、|ω_dδt/2|>π/4の場合には次の(9)式が利用されてもよい。

In general, since the accuracy of the arctangent function tan ^-1 (x) deteriorates at | x |> π / 2, | ω _d δt / 2 | is monitored, for example, | ω _d δt / 2 | <π / 4 In the case of (8), the equation (8) may be used, and in the case of | ω _d δt / 2 |> π / 4, the following equation (9) may be used.

但し、ω_dは未知なので、使用に相応しい関数の切替ポイントが目標に応じて変わるかもしれないことに留意を要する。そのような観点からは、
(10)式に示す線形結合が使用されてもよい。

However, since ω _d is unknown, it should be noted that the switching point of the function suitable for use may change depending on the target. From that perspective,
The linear combination shown in the equation (10) may be used.

但し、パラメータαは結合係数であり、固定値でもよいし、可変値でもよい。或いは、γを(6)式の逆正接関数の引き数、即ちγ≡imag((v(t₂)-v(t₁))/(v(t₂)+v(t₁)))とし、これに対して２つの閾値th₁及びth₂を用意し、γ＜th₁ならば(8)式が使用され、th₂＜γなら(9)式が使用され、そしてth₁≦γ≦th₂なら(10)式が使用されるように切り替えられてもよい。

However, the parameter α is a coupling coefficient and may be a fixed value or a variable value. Alternatively, γ is an argument of the arc tangent function of equation (6), that is, γ≡imag ((v (t ₂ ) -v (t ₁ )) / (v (t ₂ ) + v (t ₁ ))) For this, two threshold values th ₁ and th ₂ are prepared. If γ <th ₁ , equation (8) is used, if th ₂ <γ, equation (9) is used, and th ₁ ≦ γ ≦ If it is th _2, it may be switched so that equation (10) is used.

In the above description, only two received samples (instantaneous values) at certain time points t1 and t2 are used. From the viewpoint of improving the S / N ratio, different times t ₁ and t as shown in the following equation: The average values of two signal sequences acquired independently from ₂ to Δt may be used as v (t ₁ ) and v (t ₂ ). The number of samples used for averaging may be ensured as long as a stable value is obtained according to the quality of the received signal (situation such as SNR), and may be several hundred, for example.

上記の表記では、ドップラ角周波数ω_dは逆正接関数又は逆余接関数から導出されているが、このことは本発明に必須ではなく、三角関数の関係式に従って他の関数から導出されてもよい。例えば、ドップラ角周波数ω_dは逆正弦関数から導出されてもよい。

In the above notation, the Doppler angular frequency ω _d is derived from the inverse tangent function or inverse cotangent function, but this is not essential to the present invention, and may be derived from other functions according to the relational expression of the trigonometric function. Good. For example, the Doppler angular frequency ω _d may be derived from an inverse sine function.

ここで、βは速度の推定値の精度が良くなる様に、事前に最小二乗法等を用いて決定された補正係数を表す。

Here, β represents a correction coefficient determined in advance using a least square method or the like so that the accuracy of the estimated value of speed is improved.

  When calculating the above inverse trigonometric function, the calculation may be performed with an approximate expression based on Taylor expansion. Alternatively, from the viewpoint of performing calculation at higher speed and higher accuracy, the calculation may be performed using an approximate expression based on Pade expansion.

  When there are a plurality of target bodies, the Doppler frequency shift of each target body may be determined according to a certain priority order. For example, means for estimating the distance between the own device (detection and ranging device) and the target object may be prepared, and the priority order may be determined from the positional relationship between the own device and the target object.

According to the present embodiment, since only the sample values obtained at least at two different times are required to obtain the Doppler angular frequency ω _d (= 2πf _d ), the Doppler frequency can be obtained very quickly. In addition, since this method uses sample values at two time points, it is not greatly affected by noise at only one time point. The argument of the cosine function of the denominator that appears in Δn and Σn is ω _d δt / 2, so by selecting δt so that it is not equal to an odd integer multiple of π / 2, the cosine function of the denominator approaches 0 and the calculation accuracy The risk of deterioration can be easily avoided. Therefore, according to this embodiment, the Doppler angular frequency ω _d can be obtained quickly, accurately and stably.

(Numerical example)
In order to confirm the effect of the present embodiment, a simulation was performed with the following specifications.

1 chip period (T _c ) = 3.9 x 10 ^-10 (seconds)
Code length = 9 bits 1 frame period ( _Tf ) = 2.0 x ^10-7 (seconds)
Distance to _first target (T ₁ ) = 10m
Speed of the first target (T ₁ ) = 6km / h
Distance to _second target (T ₂ ) = 11m
Speed of _second target (T ₂ ) = 12km / h
The correlation output of the signal reflected by each target object is obtained as shown in FIG. 3, and the received signal after synchronization acquisition is obtained as shown in FIG. For such transmission signals and reception signals, the by Fourier transform as in the prior art find Doppler frequency f _d, 3431 frames were required for the second target member (T _2). This is a period as long as 3431 × 2.0 × 10 ⁻⁷ = 0.68 × 10 ⁻³ (seconds). Since the first target (T ₁ ) is slower, a longer period is required. On the other hand, according to the present embodiment, only 300 frames are required to obtain the target body speed with high accuracy such as 5.77 km / h (T ₁ ) and 12.03 km / h (T ₂ ). I'm done. This corresponds to the fact that only 300 × 2.0 × 10 ⁻⁷ = 0.06 × 10 ⁻³ (seconds) is required, and the calculation time can be reduced by 10 times or more. In principle, the Doppler frequency can be obtained in a short time with a smaller number of frames.

  In the DSSS method of the first embodiment, since the speed is estimated after identifying the distance between the radar and the target, if there are targets with different speeds at the same distance, the speed is estimated while distinguishing them. It's not easy to do. For example, it may be possible to use an FFT, but this is not preferable from the viewpoint of quick measurement. In the second embodiment of the present invention, a technique for appropriately finding the Doppler frequency is considered even in such a case. The method of deriving the Doppler frequency according to this method may be supplementary to the method of the first embodiment by the Doppler frequency estimation unit 530 of FIG.

First, the sample value acquired in chronological order at different N _A number of times the appropriate time as a starting point is provided. At this time, if each sample is assigned a number n (= 1,..., N _A ) according to the order in which it is acquired and expressed as v (n), this can be expressed as follows.

但し、a_mはm番目の目標体からの反射信号の振幅であり、位相φ_n,mは、

Where a _m is the amplitude of the reflected signal from the m-th target and the phase φ _{n, m} is

と表現できるので、N_A個の受信サンプルは、N_A次元のベクトルとして表現できる。

Therefore, N _A received samples can be expressed as N _A dimensional vectors.

但し、行列A、信号振幅ベクトルx及びノイズベクトルnは次式で定義されるものとする。

However, the matrix A, the signal amplitude vector x, and the noise vector n are defined by the following equations.

レーダーの受信信号はコヒーレンシー(coherency)が高ので、その性質を勘案して擬共分散行列を算出する。擬共分散行列を利用する手法については、例えば、IEEE Trans. on Signal Processing, Vol.52, No.4, 2004, pp.876-893 等で説明されている。

Since the received signal of the radar has high coherency, the pseudo covariance matrix is calculated in consideration of its nature. A technique using a pseudo-covariance matrix is described in, for example, IEEE Trans. On Signal Processing, Vol. 52, No. 4, 2004, pp. 876-893.

First, baseband signal correlation vectors r _v1 and r _v2 are calculated. For simplification, let N _P = N _S.

ここで、Eは期待値（平均値）をとることを意味する。「＊」は複素共役をとることを意味する（対象がベクトルや行列の場合は要素毎に複素共役をとる）。

Here, E means taking an expected value (average value). “*” Means taking a complex conjugate (when the target is a vector or a matrix, a complex conjugate is taken for each element).

Then, by rearranging the components of the correlation vector r _v1, r _v2, constitute a matrix R _f1, R _f2, constituting the matrix R _b1, R _b2 from further matrix R _f1, R _f2.

但し、Jは反対角要素が1であってそれ以外の要素が0である行列を示す。上付き文字の「Ｔ」は転置を表す。ここで、

Here, J represents a matrix whose opposite angle element is 1 and other elements are 0. The superscript “T” represents transposition. here,

等に注意して多少の計算を行えば、例えば、R_f1＝AXと書けることが分かる。そこで、R_f1〜R_b2までの4つを次のように並べることで行列Rを定義すると、行列Rは空間平均を適用した共分散行列と同じ情報を有することが知られている（便宜上、このプロセスをm-fbssと呼ぶ。）。

If some calculations are performed with attention paid to the above, for example, it can be seen that R _f1 = AX can be written. Therefore, when the matrix R is defined by arranging four of R _{f1 to} R _b2 as follows, the matrix R is known to have the same information as the covariance matrix to which the spatial average is applied (for convenience, This process is called m-fbss.)

R = (R _f1 R _f2 R _b1 R _b2 ) (23)
In the second embodiment of the present invention, the matrix R thus obtained is divided into an N _P × N _P matrix R ₁ and an (N _A −N _P ) × N _P matrix R _2, and (RR ^H ) ⁻¹ is calculated. The superscript “H” represents a conjugate transpose. Note that (RR ^H ) ⁻¹ is A = R ₁ R ₁ ^H (of course, different from A defined in equation (16)), B = R ₁ R ₂ ^H , And if D = R ₂ R ₂ ^H , it can be rewritten as shown in equation (24).

但し、S（説明の便宜上、スケール行列と呼ぶ）とQ（射影行列とも呼ばれる）は次式で定義される行列である。

However, S (referred to as a scale matrix for convenience) and Q (also referred to as a projection matrix) are matrices defined by the following equations.

このようにして導出された（RR^H）⁻¹を用いて擬スペクトラムP (ω)が定義される。、

Pseudospectrum P using (RR ^H ) ⁻¹ derived in this way (ω) is defined. ,

a^H(ω)は（16）式に登場するような位相ベクトルであり、ωは角周波数である。位相ベクトルa（ω）の値を所定の数値範囲にわたって変えることで（走査することで）、擬スペクトラムの値も変化する。位相角ωがドップラ角周波数に近い場合、擬スペクトラムにピークが出現する。従って、位相ベクトルを走査しながら擬スペクトラムの特性変化をモニタすることでドップラ角周波数を見出すことができる。或いは、(27)式右辺の分母をゼロと置くことで導出される代数方程式を解くことで、ドップラ角周波数が求められてもよい。

a ^H (ω) is a phase vector that appears in equation (16), and ω is an angular frequency. By changing the value of the phase vector a (ω) over a predetermined numerical range (by scanning), the value of the pseudospectrum also changes. When the phase angle ω is close to the Doppler angular frequency, a peak appears in the pseudo spectrum. Therefore, the Doppler angular frequency can be found by monitoring the characteristic change of the pseudo spectrum while scanning the phase vector. Alternatively, the Doppler angular frequency may be obtained by solving an algebraic equation derived by setting the denominator on the right side of Equation (27) to zero.

  By the way, the first term of the equation (24) includes the inverse matrix of the matrix A corresponding to the signal subspace itself, but the reason why the peak appears in the pseudo spectrum is between the signal subspace and the noise subspace. Since the orthogonal relationship is established, there may be a concern about a slight deterioration in accuracy.

In this regard, the remainder obtained by subtracting the matrix including the submatrix A ⁻¹ and the zero matrix from the left side of the equation (24), that is, QS ⁻¹ Q ^H is used, and instead of the equation (27), (29 It is conceivable to achieve high accuracy by using the expression (). This can also be regarded as a method in which the propagator method for performing a peak search for a pseudo spectrum defined by replacing (RR ^H ) ⁻¹ in Equation (27) with QQ ^H is improved by the introduction of the scale matrix S. The propagator method is described in, for example, IEEE Trans. On Signal Processing, Vol. 39, No. 3, 1991, pp. 746-749.
Specifically, instead of (RR ^H ) ^{−1 in} equation (27), using the second term QS ⁻¹ Q ^H appearing on the rightmost side of equation (24), a new pseudospectrum (29) equation is obtained. Define this and perform a peak search in the same manner as described above. Equation (28) is a singular value decomposition expression of S- ¹ , where σ _i is the reciprocal of the signal power and e _i is the corresponding eigenvector.

さて、ドップラ周波数近辺でない周波数位置では対応する信号成分が存在しないので、R₁、R₂≒0であるから、位相情報に関係なくP≒0であり、S→特異、となる。即ち、a^H（ω）Qe_i≒const. になる。しかも信号が存在しない以上、σ_iは非常に大きいので、P（ω）は小さくなる。

Now, since there is no corresponding signal component at a frequency position that is not in the vicinity of the Doppler frequency, since R ₁ and R ₂ ≈0, P≈0 regardless of the phase information, and S → singular. That is, a ^H (ω) Qe _i ≈const. Moreover, since there is no signal, σ _i is very large, so P (ω) is small.

In the vicinity of the Doppler frequency, R ₂ ≈ (A ⁻¹ B) ^H R ₁ holds, so S becomes a singular matrix, but at the same time a ^H (ω) Q is also approximated to 0 (PROPAGATOR), and the signal is further As long as it exists, σ _i also decreases, and as a result, a peak appears in P (ω). That is, a pseudo peak hardly appears other than the true Doppler frequency. Moreover, the peak corresponding to the corresponding Doppler appears very sharply due to the scaling effect of S- ¹ . As a result, the Doppler frequency can be obtained with high accuracy.

Note that exactly divide the R needs to know exactly N _P, but to this value can be used a method including the singular value decomposition such as AIC or MDL, norm of RR ^H such, the matrix R It can also be inferred by examining the properties of the matrix associated with.

The means taught by the embodiments will be listed below as an example.
(Appendix 1)
Direct sequence spread spectrum (DSSS) radar,
Means for transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Means for receiving signals reflected by one or more target objects;
Means for calculating the sum and difference of the received signals obtained at different time points, and determining the Doppler frequency shift of the target body from the phase difference between the sum signal and the difference signal;
DSSS radar, characterized by having
(Appendix 2)
The radar according to appendix 1, wherein the different time points are separated by a period corresponding to one code length of the predetermined code sequence.
(Appendix 3)
The radar according to claim 1, wherein the Doppler frequency shift is derived by using an imaginary part of a number obtained by dividing the difference signal by the sum signal as an argument of an arc tangent function.
(Appendix 4)
The radar according to claim 1, wherein the Doppler frequency shift is derived by using an imaginary part of a number obtained by dividing the sum signal by the difference signal as an argument of an inverse cotangent function.
(Appendix 5)
The Doppler frequency shift is an arctangent function using an imaginary part of the number obtained by dividing the difference signal by the sum signal as an argument, and an inverse remainder using the imaginary part of the number obtained by dividing the sum signal by the difference signal as an argument. The radar according to appendix 1, wherein the radar is derived by linear combination with a tangent function.
(Appendix 6)
The radar according to claim 1, wherein each of the received signals obtained at the different time points is derived by averaging a plurality of received samples.
(Appendix 7)
The radar according to appendix 1, further comprising means for calculating a correlation between a received signal and the predetermined code sequence and deriving a synchronization timing for each of the one or more target bodies.
(Appendix 8)
The sum and difference of received signals obtained at different points in time for each of one or more target bodies are calculated, and a Doppler frequency shift is derived for each target body from the phase difference between the sum signal and the difference signal. The radar according to appendix 7.
(Appendix 9)
The radar according to any one of appendices 3 to 4, wherein the inverse trigonometric function is calculated by Pade expansion.
(Appendix 10)
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from the signal;
Means for combining the correlation vectors to form the pseudo-space average covariance matrix R;
Means for calculating the Doppler frequency shift from a pseudospectrum or algebraic equation using (RR ^H ) ⁻¹ ;
The radar according to appendix 1, characterized by comprising:
(Appendix 11)
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from the signal;
Means for combining the correlation vectors to form the pseudo-space average covariance matrix R;
It means for estimating the number of incoming signals N _S,
Means for constructing a projection matrix Q from R according to N _S ;
Means for calculating a scale matrix S from a submatrix of R;
2. A matrix generated by appropriately combining Q and S, for example, means for calculating a Doppler frequency shift from a pseudo spectrum or an algebraic equation using QS ⁻¹ Q ^H radar.
(Appendix 12)
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from the signal;
Means for combining the correlation vectors to form the pseudo-space average covariance matrix R;
It means for constituting a projection matrix Q from the R in accordance with the number of incoming signals N _S set in advance,
Means for calculating a scale matrix S from a submatrix of R;
Means for calculating a Doppler frequency shift from a pseudo spectrum or an algebraic equation using a matrix generated by appropriately combining the Q and S, for example, QS ⁻¹ Q ^H ;
The radar according to appendix 1, characterized by comprising:
(Appendix 13)
Means for estimating the distance to the target body;
Means for determining priority of speed estimation from the relative relationship between the target body and the device, and calculating a Doppler frequency shift based on the priority;
The radar according to appendix 1, further comprising:
(Appendix 14)
In radar according to any one of Appendices 10 to 13, the QS ^-1 Q ^H, is composed of partial matrix A ^-1 and a zero matrix of elements corresponding to the signal subspace from the (RR ^{H) -1} A detection and ranging apparatus characterized in that the rough estimation and the detailed estimation of the Doppler frequency shift are switched in the same apparatus by subtracting the matrix.
(Appendix 15)
In the radar according to any one of appendices 10 to 14, means for calculating a Doppler frequency shift from a pseudo spectrum or an algebraic equation using QQ ^H instead of (RR ^H ) ⁻¹ or QS ⁻¹ Q ^H Radar further comprising:
(Appendix 16)
In the radar according to any one of appendices 10 to 14, D ⁻¹ , (B ^H A ⁻¹ B) ⁻¹ , (B ^H B) ⁻¹ , and (D−B ^H B B) are used instead of S ^−1. ) ^-1, or radar, characterized in that it comprises means for calculating a Doppler frequency shift from the pseudo-spectrum or the algebraic equation using the ^{S -1 (Q H Q) -1} .
(Appendix 17)
A method used in direct sequence spread spectrum (DSSS) radar,
Transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Receiving signals reflected by one or more target objects;
Calculating the sum and difference of the received signals obtained at different time points, and determining the Doppler frequency shift of the target body from the phase difference between the sum signal and the difference signal;
A method characterized by comprising:
(Appendix 18)
Transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Receiving signals reflected by one or more target objects;
Calculating the sum and difference of the received signals obtained at different time points, and determining the Doppler frequency shift of the target body from the phase difference between the sum signal and the difference signal;
Is a computer program that directly executes a sequence spread spectrum (DSSS) radar.
(Appendix 19)
A method used in direct sequence spread spectrum (DSSS) radar,
Transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Receiving signals reflected by one or more target objects;
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Calculating correlation vectors r _v1 and r _v2 from the signal;
Combining the correlation vectors to form a pseudo-space average covariance matrix R;
Calculating the Doppler frequency shift from a pseudospectrum or algebraic equation using (RR ^H ) ⁻¹ ;
Is a computer program that directly executes a sequence spread spectrum (DSSS) radar.
(Appendix 20)
A method used in direct sequence spread spectrum (DSSS) radar,
Transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Receiving signals reflected by one or more target objects;
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Calculating correlation vectors r _v1 and r _v2 from the signal;
Combining the correlation vectors to form a pseudo-space average covariance matrix R;
Estimating a number of incoming signals N _S,
Constructing a projection matrix Q from R according to N _S ;
Calculating a scale matrix S from a submatrix of R;
Calculating a Doppler frequency shift from a pseudo spectrum or an algebraic equation using a matrix generated by appropriately combining Q and S, for example, QS ⁻¹ Q ^H ;
Is a computer program that directly executes a sequence spread spectrum (DSSS) radar.
(Appendix 21)
A method used in direct sequence spread spectrum (DSSS) radar,
Transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Receiving signals reflected by one or more target objects;
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Calculating correlation vectors r _v1 and r _v2 from the signal;
Combining the correlation vectors to form a pseudo-space average covariance matrix R;
A step of configuring a projection matrix Q from the R in accordance with the number of incoming signals N _S set in advance,
Calculating a scale matrix S from a submatrix of R;
Calculating a Doppler frequency shift from a pseudo spectrum or an algebraic equation using a matrix generated by appropriately combining Q and S, for example, QS ⁻¹ Q ^H ;
Is a computer program that directly executes a sequence spread spectrum (DSSS) radar.

The system block diagram of DSSS radar is shown. It is a figure which shows typically the relationship between SS code _| cord _| chord, code length ( _Tf ), and chip _| tip period ( _Tc ). It is a figure which shows an example of a correlation output. It is a figure which shows a mode that the received sample is obtained continuously for every _Tf about each of two target bodies. It is a block diagram which shows the radar by one Example of this invention.

102 Code Generator 104 Oscillator 105 Carrier Distribution Unit 106 Mixer 108 High Power Amplifier (HPA)
110, 116 Delay unit 112 Doppler frequency setting unit 114 Mixer 118, 120 Additive Gaussian noise introduction unit 122 Low noise amplifier (LNA)
124 mixer 126 demodulator 130 delay adjustment unit 132 mixer 134 integrator 136 peak detector 138 counter 140 fast Fourier transform unit (FFT)
502 Code Generator 504 Oscillator 506 Carrier Distribution Unit 508 Mixer 510 High Power Amplifier (HPA)
512 Low noise amplifier (LNA)
514 Mixer 516 Demodulation unit 520 Delay adjustment unit 522 Mixer 524 Integrator 526 Peak detector 528 Counter 530 Doppler frequency estimation unit 532 Signal processing unit

Claims (2)

Direct sequence spread spectrum (DSSS) radar,
Means for transmitting a transmission signal including a predetermined code sequence to one or more target bodies;
Means for receiving signals reflected by one or more target objects;
A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from received signals at different times;
Means for combining the correlation vectors to form the pseudo-space average covariance matrix R;
And a means for calculating the Doppler frequency shift from a pseudo spectrum or an algebraic equation using (RR ^H ) ⁻¹ . A vector v = (v (t ₁ ),..., V (t _N )) ^T is constructed from sample values of received signals at different times t ₁ ,..., T _N. Means for calculating correlation vectors r _v1 and r _v2 from the signal;
It means for configuring the擬空between average covariance matrix R in combination the correlation vector, a means for estimating the number of incoming signals N _S, the means constituting the projection matrix Q from the R in response to N _S,
Calculates the Doppler frequency shift from the pseudo spectrum or algebraic equation using the means to calculate the scale matrix S from the submatrix of R and a matrix generated by appropriately combining the above Q and S, for example, QS ⁻¹ Q ^H The radar according to claim 1, further comprising:

Images