WO2011085629A1 - Method for direction finding - Google Patents

Abstract

A method for analyzing and processing signals of multiple antenna units, a method for direction finding and a system for analyzing and processing signals of multiple antenna units and for direction finding are disclosed in the present invention which relates to the field of signal analyzing and processing and the field of direction finding of wireless signal. The method for analyzing and processing signals and the method for direction finding include: adopting two antennae with different directional patterns; performing short-time statistics and analyzing over a time window having a certain length on the signals received synchronously by the two antennae respectively to obtain a ratio of short-time statistics and analyzing values of signals of the two antennae at an analyzing point in a corresponding analyzing domain; determining the components of the signal and/or the interference intensity according to the variation characteristic of the ratio; or obtaining the data graph of the ratio vs. azimuth, and then analyzing and processing the data graph to derive parameters such as arrival direction of the signal. Especially, a single station location method is provided in the present invention. The present invention provides reliable basis for signal processing and signal direction finding by analyzing component and structure of signal, and improves the sufficient understanding on electromagnetic circumstance by means of the exact and believable direction finding and locating of signal resource.

Description

 - - Direction finding method

Technical field

 The present invention relates to the field of signal analysis processing and wireless signal direction finding, and more particularly to a method for combining spatial signal analysis processing and signal source direction finding using a multi-antenna unit including a directional antenna, and the analysis processing method and/or Or a system of signal source direction finding methods. Background technique

 At present, it is mainly to use a single omnidirectional antenna to measure the amplitude, field strength, level and other parameters of the signal at each frequency point to realize the signal analysis of the electromagnetic environment, and the measured value is the result of superposition of each signal, and cannot be distinguished. The composition and structure of the signal makes it difficult to determine whether there are co-channel interference and multipath interference, as well as the frequency range of the interference and the strength of the interference.

 The current direction finding method differs greatly depending on the system. Some of the direction finding methods can only give a single degree of orientation, such as phase direction finding, large sound point direction finding, etc. Other methods can give multiple degrees of orientation, such as single antenna rotation direction finding and spatial spectrum estimation. However, the electromagnetic space signal has great complexity. The existence of the same frequency signal, multipath signal, and the spectrum change and propagation fading of the signal itself make the existing direction finding technology system face a huge dilemma. Its direction finding accuracy and reliability The complexity of the electromagnetic environment has dropped dramatically. Although the method of spatial spectrum estimation in the existing direction finding technology system can solve these problems to some extent, its sensitivity to the model and its processing ability are limited by the number of antenna elements, and its application in the more complex electromagnetic environment is also very The large limit, and the spatial spectrum estimation of the antenna array is large, which is not conducive to the realization of portable direction finding.

 In the amplitude direction finding, there is a two-channel ratio amplitude direction finding method. The ratio of the received potentials is determined by the ratio of the two orthogonal Edeck antennas, and the channel equalization technique is used to eliminate the influence of the channel gain inconsistency. However, this way of implementation is not desirable in terms of both cost of implementation and actual results. First of all, the realization of multi-channel equalization will incur a high cost. In addition, there are many defects in the technical effect. On the one hand, the ratio of the receiving potential is easily unstable due to the influence of noise and multipath signals. Aspects, using this ratio to estimate the orientation will inherently deviate due to the presence of multipath signals. Summary of the invention

An object of the present invention is to solve the technical problem that the existing direction finding method has inaccurate direction finding without prior analysis processing, and to provide an analysis and processing method capable of analyzing the composition and structure of the signal. The technical solution adopted by the present invention is as follows: A multi-antenna unit signal analysis processing method includes the following steps: Step 1: Synchronously receiving signals by using first and second antennas, at least a first antenna is a directional antenna, and two antennas have Different directionality;

 Step 2: Perform short-term statistical analysis on the signals received by the directional antenna and the other antenna simultaneously in a certain length of time window, and obtain a ratio of the short-term statistical analysis values of the two at an analysis point of the corresponding analysis domain ( For example, when the statistical analysis method adopts the short-time Fourier transform, the amplitude spectrum ratio at a certain frequency point in the transformation result may be taken, wherein the length of the time window is adjustable, and is greater than all signals (herein all the main signals) , the longest multipath delay that does not consider the extremely weak signal that is reflected multiple times);

 Step 3: Determine a composition of the signal and/or an interference strength according to the variation characteristic of the ratio.

Preferably, in step 3, the intensity of the interference to the signal is determined according to the statistical value σ ² , a calculation method of σ ²

Μ is: a ² = y(ln(D(S, t., T))-Df , where 5=^ ~~ ; , r) is calculated at the analysis point according to step 2 in the first time window The ratio is; for natural numbers, i ≤ j ≤ M, M is a natural number greater than ί; when the statistic is greater than a certain threshold ,, the signal can be considered to have the same frequency heterogeneous signal. Here, if the frequency domain is used for analysis, it is J'iy. The logarithm of the formula is used to equalize the contribution of the numerator and denominator of the ratio. Since the interference signal is much higher than the original signal, the change of the ratio will also slow down. Therefore, the statistical value σ ² indicates the degree of mixing of the signal to a certain extent. When the magnitudes of the two same-frequency heterologous signals are equivalent, the statistical value σ ² Most likely to get the maximum. When calculating σ ² , the data valid in D(^, t., r) is selected, and the validity of /) t., r) is determined by the magnitude of the short-term statistical analysis values of the first and second antennas.

A second object of the present invention is to provide a simple and accurate direction finding method which can utilize signal composition information determined by the above analysis processing method. The technical solution adopted by the present invention is: A multi-antenna unit signal direction finding method, comprising the following steps: Step 1: Synchronously receiving signals by using first and second antennas, wherein the first antenna is a rotatable directional antenna, The orientation of the first antenna can be relatively independently adjusted. The second antenna is an omnidirectional antenna or a fixed directional antenna. When receiving a signal, the first antenna is rotated. If the second antenna is a directional antenna, it is rotated to an appropriate orientation. Post-fixed; Step 2: Perform short-term statistical analysis on the signals synchronously received by the first antenna and the second antenna in a certain length of time window, and obtain an amplitude spectrum ratio of the short-term statistical analysis values of the two in a frequency domain. The length of the time window is greater than the longest multipath delay of all signals (here, all major signals, regardless of the extremely weak signal that is infinitely reflected); Step 3, determining the correspondence between the amplitude spectrum ratio and the azimuth angle Relationship, and according to this, the data curve of the amplitude spectrum ratio with respect to the azimuth is obtained. Preferably, the method further includes a step 4 of estimating a possible number of incoming waves according to the data curve, and each of the incoming wave direction signal parameters, and using the determined values as initial values, using a directivity function of the first antenna. The weighted superposition model solves the measured data curve by the numerical optimization method, and obtains the optimal value of the incoming wave direction and the signal parameters of each incoming wave direction.

 A third object of the present invention is to provide an analysis processing and direction finding system which can realize the above analysis processing method and/or direction finding method.

 The technical solution adopted by the present invention is: a multi-antenna unit signal analysis processing and direction finding system, comprising an antenna unit, a receiving unit, a data processing unit and a display unit, the antenna unit comprising at least two antennas, at least one of all antennas The pair is a directional antenna, and the orientation of each directional antenna can be relatively independently adjusted; the signals respectively measured by the antennas are input to the data processing unit through the receiving unit for synchronous analysis processing, and the data processing unit pairs the directional antenna and Synchronizing the received signals of the other antenna to perform short-term statistical analysis within a certain length of time window, and obtaining an amplitude spectrum ratio of the short-term statistical analysis values of the two at an analysis point of the corresponding analysis domain, wherein the time The length of the window is greater than the longest multipath delay of all signals (here, all major signals, regardless of the extremely weak signal that is reflected multiple times); the result of the data processing unit analysis processing is input to the display unit for output display.

 Preferably, the receiving unit further includes a mixing circuit and a phase locked loop controlled by the data processing unit, and each signal input to the receiving unit is first input to the mixing circuit for down-converting step by step, forming and Each of the corresponding baseband signals is input to the data processing unit for short-term statistical analysis; the data processing unit sends a local oscillator signal phase lock signal to the receiving unit when starting short-term statistical analysis, stopping The phase-locked loop tracks the phase of the signal frequency to keep the frequency and phase of the local oscillator signal down-converted for each signal. The beneficial effects of the invention are: by analyzing the composition and structure of the signal, providing a reliable basis for signal processing and signal direction finding; and promoting a better understanding of the electromagnetic environment by accurately and reliably determining the direction of the signal source.

DRAWINGS - Figure 1 is a block diagram of the two-channel receiving mode of the analysis processing and direction finding system of the present invention; Figure 2 is a block diagram of the single-channel fast switching receiving mode of the analysis processing and direction finding system of the present invention; A schematic structural view of the antenna unit shown in FIG. 1;

 Figure 4 shows primary signal analysis data of the analysis processing method according to the present invention;

 Figure 5 shows another signal analysis data of the analysis processing method according to the present invention;

 6 shows a primary source direction finding data according to the direction finding method according to the present invention, wherein the secondary signal direction uses an antenna with a weak directivity;

 Fig. 7 shows another signal source direction finding data according to the direction finding method of the present invention, wherein the secondary signal direction finding employs a more directional antenna.

BEST MODE FOR CARRYING OUT THE INVENTION The analysis processing method and the direction finding method of the present invention are mainly implemented based on a directional antenna pattern and a combined use of multiple antennas. The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

 As shown in FIGS. 1 and 2, the analysis processing and direction finding system of the present invention includes an antenna unit 1, a receiving unit 2, a data processing unit 3, and a display unit 4. The antenna unit 1 includes at least two antennas, respectively The antenna 11 and the second antenna 12, at least one of the antennas is a directional antenna. In this embodiment, at least the first antenna 11 is a directional antenna, and the orientation of the directional antenna can be relatively independently adjusted, that is, it can be pointed Fixed, it can also be rotated (change its pointing at a certain speed); the data processing unit 3 performs synchronous analysis processing on the signal data measured by the two antenna combinations of the antenna unit including at least one directional antenna; Unit 4 is used to display the results of the data processing unit analysis process.

 The antenna unit 1 is typically configured as follows: (1) an omnidirectional antenna and a pair of directional antennas, the combination of which can be used for static applications such as signal analysis, that is, no rotation is required in a single analysis. The application of the antenna can also be used for the dynamic application of the signal direction finding, that is, the application that needs to rotate the antenna; at the same time, the configuration does not excessively increase the size and weight of the antenna unit 1, so that the portable application form can be used, of course It can be used in the form of vehicles and fixed stations; (2) Two rotatable directional antennas, one of which is a main rotating antenna for rotating various angular velocities and the other is an auxiliary rotating antenna for Auxiliary measurement, this combination can also add an omnidirectional antenna, mainly for on-board or fixed stations.

In order to better implement the signal analysis processing and the direction finding function, the orientation of each directional antenna in the antenna unit 1 should be relatively independently adjustable, and the data processor unit 3 can calculate the optimal orientation of the directional antenna. When the directional antenna rotates, it should be carried out around the center of the antenna as much as possible, so as to reduce the phase shift of the signal caused by the rotation, thereby avoiding Bring model bias to signal analysis and direction finding. Since the general antenna shape has symmetry, the center of the antenna generally refers to the symmetrical center line of the antenna.

 As shown in FIG. 1 , the receiving unit 2 generally has two or more receiving channels, and includes a synchronous control circuit, so that the signal parameters of each antenna unit can be measured synchronously. In this embodiment, two antennas are used, which can be configured correspondingly. The receiving channels are respectively divided into a first receiving channel 21 corresponding to the first antenna 11 and a second receiving channel 22 corresponding to the second antenna 12. Of course, a single channel as shown in FIG. 2 can also be used, which is fast through a single channel. Switching, the approximate synchronous reception of the two antenna signals is realized under the control of the synchronization circuit. The receiving unit includes a mixing circuit and a phase locked loop controlled by the data processing unit 3, and the data processing unit 3 sends a local oscillator signal phase lock signal to the phase locked loop at the beginning of the analysis, so that after the data analysis starts, Keep the frequency and phase of the local oscillator signal unchanged. Here, since the local oscillator signal is used to down-convert the sampled data, if the phase-locked loop continues to work during signal acquisition, the frequency and phase of the local oscillator signal may drift, resulting in signal reception characteristics changing with time and affecting signal analysis. The stability and the numerical model of the source direction finding.

 The following describes the configuration of two sets of rotatable directional antennas. Before using the above analysis processing and direction finding system to perform signal direction finding, first perform signal processing analysis on the corresponding frequency band signals, and initially determine the presence and the approximate number of the same frequency heterogeneous signals. The process of analyzing and processing the signals of the present invention is as follows:

 First, the first antenna 11 is rotated, and the electric field intensity sampling value at a frequency point is recorded, and the possible directions of the incoming waves are initially determined, and the direction of the incoming waves of the two possible co-channel heterogeneous signals is screened therefrom; The first and second antennas face the direction of the incoming waves of the two signals, and try to avoid mutual influence. Finally, the data processing unit 3 analyzes and processes the two antenna signals to determine whether there are multipath signals and co-channel heterogeneous signals. And distinguish the multipath signal and the same frequency heterogeneous signal by the difference of the spectrum structure and the difference of the signal amplitude, and distinguish the direct signal and the multipath signal by using the difference of the signal amplitude, and initially determine the composition and structure of the signal. If the same-frequency heterogeneous signal is not found for each possible direction of the incoming wave, it can be judged that there is no interference signal of the same-frequency heterogeneous signal or the same-frequency alien in the space. The data processing unit 3 mainly analyzes and analyzes the characteristics of the directional antenna based on the gain of the different directional signals arriving at the directional antenna. The analysis and processing method of the data processing unit 3 will be described in detail below.

 The difference of the same-frequency heterogeneous signal is mainly reflected in the spectrum structure and the spectrum variation law. The spectrum structure includes the center frequency, the frequency bandwidth and the spectrum shape. The spectrum variation law mainly refers to the variation of the amplitude and phase of each frequency component of the signal with time. .

In the analysis process, the set signal uses the baseband or baseband modulation signal, and the high frequency modulation and demodulation process is not analyzed. For each manner of screening out the direction of the incoming wave of two possible co-frequency heterogeneous signals, there are two co-channel heterogeneous signals /i) and / ₂ (0) in the space, and the two signals arrive at the present invention. There are multipath signals in the system. - - No. / (0 direct signal, each multipath signal reaches the main rotating antenna (first antenna 11) attenuation factor is

K _u , ₂ ... _LM , the delay is 7^, ? ...:^, the incident angles are ^, θ _η ... θ _1πι , and the attenuation factors reaching the auxiliary rotating antenna (second antenna 12) are respectively „, ₂ ... _m , and the delays are Γ' _Ι Γ' ₁₂ ... Γ ' _1μ , the incident angle is ₂ ... _m . Signal / ₂ (0 direct signal, the attenuation factor of each multipath signal reaching the main rotating antenna is ₂₁ , ₂₂ delays are respectively Γ ₂₁ , Γ ₂₂ ... Γ ₂ „, the incident angles are respectively θ ₂₂ ... θ _2η , and the attenuation factors reaching the auxiliary rotating antenna are ₂₁ , ^ ₂₂ .. ' ₂ „, and the delay is 7” ₂₁ , Γ ' ₂₂ ...Γ' ₂ „, the angle of incidence is ₂₁ , ₂ .. ' ₂ „. The number of multipath signals of the signal / ₂ (t) is m_l, n-\, m, n are greater than or equal to The natural number of 1.

 When the short-term statistical analysis method employs the short-time Fourier transform, the specific steps of the analytical processing method of the present invention are as follows:

 Step 1: Keep the main rotating antenna and the auxiliary rotating antenna fixed at a certain angle and roughly point to the possible incoming wave direction; wherein the two antennas have different directivity, the so-called different directivity may be the directivity of the two antennas. The functions are different (so that their directivity is different at any time), or the directivity functions of the two antennas are the same, but the antennas are pointing differently;

 Step 2: Perform short-term statistical analysis on the signals simultaneously received by the two fixed antennas, and obtain the amplitude spectrum ratios of the short-term statistical analysis values at a frequency point. Here, the two fixed antenna units can be synchronously received. The magnitude of the short-term statistical analysis values of the signals collectively determine the validity of the amplitude spectral ratio. The short-term statistical analysis method used in the analysis below is the short-time Fourier spectrum calculation value in the time window. Of course, other statistical values may be used for some applications, such as integral measurement of voltage over time in time window, wavelet transform, and the like.

The time window for short-time statistical analysis of the signal needs to have a certain length, which is greater than the longest delay of all the major multipath signals. For the implementation of the single channel time switching as shown in FIG. 2, the time is The length of the window is also greater than the time interval of channel switching. If the length of the time window does not satisfy the above requirements, the ratio of the short-term statistical analysis values of the signal may also be sharply fluctuated without interference of the same-frequency heterogeneous signal. Taking the influence of the multipath signal as an example, when the antenna signal is windowed, there are differences in the signal at the edge except for the different delays of the signal data with different delays, for example, for the time window (^, ₂ ), the data of the direct signal is ^(0 ( <ί<ί ₂ ), and the multipath signal data with a delay of Γ ₀ is 3⁄4 (2 {t _x +T ₀ <t<t ₂ +T ₀ ) _a In fact, when there is only a multipath signal, since the difference data exists in the time window, the amplitude spectrum ratio may fluctuate with time, which also illustrates the prior art dual channel ratio direction finding method. Two antennas measured - - The ratio of the voltage value (corresponding to the length of the time window is 0) is not desirable because it may fluctuate drastically due to the influence of the multipath signal. According to the general distribution of multipath delay, we believe that the length of the time window is generally greater than or equal to 3 microseconds, otherwise the stability of the signal short-term statistical analysis value ratio will be seriously affected. The larger the time window length, the more advantageous it is to reduce the effect of the multipath signal on the stability of the ratio. However, if the time window is too long, the data length will be too long, which will increase the difficulty of subsequent data processing. Therefore, on the one hand, the time window cannot be too long. On the other hand, the signal should be fully down-converted before statistical analysis is performed. It is best to perform at the baseband with the lowest sampling rate. For example, for a modulated signal with a bandwidth of 300 kHz, the signal band can be moved to a position of 0 Hz to 300 kHz by analog or digital filtering and mixing methods, and then the signal is statistically analyzed. In addition, when the statistical analysis uses the short-time Fourier transform, the multipath delay T corresponds to the phase shift of e- ⁷ ^, which brings many difficulties to the model analysis, and also makes the ratio of the short-term statistical analysis value of the signal with time. The fluctuation increases, but when the frequency is reduced to baseband processing, if the multipath delay is not too large, the term is usually close to 1, which simplifies the signal model and reduces the influence of the multipath delay contrast value fluctuation. , The baseband width referred to here does not exceed the signal bandwidth.

The directivity functions of the main rotating antenna and the auxiliary rotating antenna are respectively, and the short-time Fourier spectrum qO^, and <3⁄4 ( ί, which can be measured by the main rotating antenna and the auxiliary rotating antenna, respectively, can be approximated as:

G ₂ ( jw, t, τ) = F _x { jw, t, τ) Σ[Κ

)β- ^ν ' ^ϊΐί ] + F ₂ { jw, t, τ) Σ[Κ _2ί ξ(θ' _2ί )β- ^]κΤ ' ² ·

Where t is the starting point of the data recording time and r is the length of the data recording time window (generally lms)

Fourier transform values of F ₂ ( , signal and / ₂ respectively). Thus, the spectral values of the short-time Fourier spectrum and (3⁄4(t,r) at the frequency ν are:

_ _ In general, the propagation path of multipath signals is relatively stable in a short time, then] = Μ' ₀ ) and

G ₁ (jw ₀ , t, T) = F ₁ (jw ₀ , t, T) M(w ₀ ) + F ₂ (jw ₀ , t, T)N(w ₀ )

 That is, the short-time spectrum of the signal measured by the two antennas in the time window is at w. The single-frequency point amplitude spectrum ratio is: 八 / . , 1 G, ( jw , t, τ) I

Here the ratio) (jV., t, τ) is valid by G. , t, τ) and G ₂ G. , , r) jointly decides that if both are small, that is, a short-term statistical analysis value is less than a constant, and another short-term statistical analysis value is less than a constant " ₂ , then the ratio has no validity, where; The mode is not unique and can be selected according to the actual measurement environment, such as a value greater than 2~5db of the background noise, or a quarter of the statistical average of the spectrum of the frequency signal. If one or both of them are large, the ratio is valid. Due to the value and signal with ^^ 1

\M (w ₀ )\ \N(w ₀ )\

Path / 1 (0, Λ) and the system reaches the point of the two directional antennas and other relevant factors, so in most cases determined) _(., Variation range. Directional antenna

The stronger the directionality of the line, the easier it is to obtain, and the larger difference between \[with: \[

\M (w ₀ )\ \N(w ₀ )\ )( ., ,r) has a larger range of variation. The limit case is that when the directivity of the two directional antennas is good enough, the directions of the two different frequency signals are sufficiently different, so that each pair of directional antennas receives only one of the direct and multipath signals of one of the signals. When the signal is signaled, (^ν., , Γ) varies from zero to infinity. In addition, in this process, the influence of background noise should be considered. The influence of background noise is that a constant is superimposed on the numerator and denominator of (ν., ,Γ). Therefore, the background noise should be measured first in the analysis process, and then The corresponding background noise estimates are subtracted from the numerator and denominator of ^ν., ,Γ).

If there is no inter-frequency heterogeneous signal, there is only one signal source, ie / ₂ () is /; For the _ _ path signal, then D(jw ₀ , t, T) will approximate a constant between II and II. When / (0, Λ ) is

|M(w ₀ )| |N(w ₀ )|

The same frequency heterogeneous signal is at the frequency point w. When the variation of the spectral value is inconsistent, the frequency point w. ^^, ,Γ) will change with time _t , two sources /; (0, / ₂ (0 spectrum change is greater, the difference between (,, η) is more obvious. The limit case is for the same frequency single The fixed source of the two different locations of the communication, ,, r) will be in two constants

|M(w.)| Jumps back and forth between ^^, and the same, for _n fixed sources and obey the simplex communication rules of \M(w ₀ )\ |N )|

In the case, it should be a transition between n constants, where n is a natural number greater than or equal to 2, from which the number of sources can be estimated. If one of the antennas receives a relatively small signal in the direction of its alignment, it may cause the (..., τ) change to be inconspicuous and subject to systematic deviation. In this case, a proper fixed gain can be applied to the corresponding receiving path. Make) G. , , r) change significantly; can also adjust the phase of the local oscillator signal to be as close as possible to the phase of the signal to be received, to enhance the receiving effect of the signal to be received and weaken the receiving effect of other signals.

 The fluctuation caused by the multipath signal in the time window can be reduced by increasing the length of the time window, while the fluctuation of the index £>(w,,,r) caused by the same-frequency heterogeneous signal is independent of the length of the time window. When the fluctuation of the index (w, ,,r) is very obvious, it can be determined that there is a co-channel heterogeneous signal. When the fluctuation of the index £>(,,, is small, the fluctuation may be caused by a small inter-frequency heterogeneous signal, or it may be caused only by a multipath signal, which can be based on the fluctuation of £>(,,,r). Whether there is sensitivity to the length of the time window to determine whether the subject causing the fluctuation is a co-channel heterogeneous signal. Therefore, increasing the length of the time window can improve the ability of the system to resolve multipath signals and co-channel heterogeneous signals. When one of the powers of the heterogeneous signal is very small, it is necessary to increase the length of the time window to increase the separation rate. When the time window is not large, the index at each frequency point in the frequency band is £>(w.,, r) When there is almost no fluctuation, it can be considered that there is only one incoming wave direction. Since the delay caused by the path difference lkm is about 3.333 |iS, the time window for analysis can take the reference value 1ms.

 When it is necessary to carefully investigate whether there is multi-path interference caused by co-channel interference or multi-path signal caused by the same-frequency heterogeneous signal at each frequency point in a frequency band, the frequency point ^> index can be separately examined. The variation over time, including the maximum, minimum, and variance (and several time windows) within the indicator £>( ., ,Γ) is related to the variance

统计 statistical value σ ² , σ ² = ](1 _η ( )(^, .,Γ))- )) ² , where D=^=i—— ; D(jw _i ,t _j ,T) _ _ is the ratio in the first time window calculated according to step 2; for natural numbers, l≤ ≤M, M is a natural number greater than 1; when the statistical value is greater than constant ,, then the signal must have the same frequency heterogeneous signal . Thereby, the interfered portion and the undisturbed portion can be found, and for the disturbed portion, the degree of interference can be evaluated. In addition, signal separation and reduction can also be performed according to this. For example, for a signal whose spectrum is symmetric about the center frequency, it is found by the foregoing analysis that when the upper side frequency is disturbed and the lower side frequency is undisturbed, the lower side frequency can be used to recover the original signal.

When the statistical value _σ ^{2 of the} DOW, when the time t changes, is greater than a certain value ,, it can be considered that there is necessarily a co-channel interference. When the two signal sources reach the system, the closer the signal power is, the more the relative change of the spectrum is, and the larger the statistical value σ ² is. When the statistical value σ ^{2 is} less than or equal to a certain value C ₂ , it is generally considered that there is only one incoming wave direction; when the statistical value σ ^{2 is} greater than the constant ς less than or equal to ς, it is generally considered that there is no co-channel interference and multipath interference exists, but When the statistical value σ ^{2 is} less than or equal to ς, the possibility of co-channel interference is not excluded. Among them, ς and ς need to be selected according to the actual measurement environment. Generally, multiple pre-analysis can be performed in the actual analysis, and according to the pre-analysis statistics. The values select constants ς and ς, and the constants ς and ς increase with increasing , and decrease with increasing length of time window. When the time window is lengthened, the values of ς and ς are correspondingly smaller, that is, the resolution is improved. Of course, for some very short bursts of co-channel interference signals, the time window should not be too long to be much longer than the interference signal length, resulting in smoothing of the variance. If only the overall interference of the frequency band is judged, the average value of the index Ζ ν, ί, Γ at the main frequency point in the frequency band; ζ ^, Γ) can be determined according to the change of time, so that a better result can be obtained. Stability, where Ν is the number of selected frequencies. The linear difference in the reception frequency characteristics of the two channels does not affect the stability of the index Ζ ν, ί, Γ), and therefore the equalization of the reception channel is low.

In the same time window, the normalized spectrum is obtained by dividing the spectrum in the frequency band of the two antenna signals by the respective maximum values (^, and (3⁄4(>^, in the presence of co-channel interference). In the case of normalized spectrum (>^, and (3⁄4(>^, there will be significant differences). In the case where co-channel interference does not exist and multipath interference exists, the normalized spectrum will also appear certain. The difference is only that the difference is not significant. Of course, the difference in the receiving frequency characteristics of the two channels (mainly caused by factors such as frequency nonlinearity) also causes the difference in the normalized spectrum, which can improve the equalization of the receiving channel. Therefore, in most cases, the difference in normalized spectrum can also be used as an important indicator for determining co-channel interference and multi-path interference. - - When the same-frequency heterogeneous signal is not found for each possible incoming wave direction, it can be judged that there is no interference signal of the same-frequency heterogeneous signal or the same-frequency heterogeneous in the space.

 Ζ ^ό, , Γ) caused by the same-frequency heterogeneous signal fluctuates rapidly with time. If there is no same-frequency heterogeneous signal, if Ζ^ 'Μ^,Γ) fluctuates greatly with time, It may be caused by two factors: First, the change of the position of the source, which leads to the change of the propagation path and the reflection and refraction paths; and second, the change of the propagation channel. However, in a general electromagnetic environment, the propagation channel does not change rapidly, even if the change only affects the propagation of those minor multipath signals. Therefore, the main factor under normal conditions is the change in the position of the source, from which it can be judged whether the source is a mobile source. If the source is a mobile source, and all or part of the path of the propagation changes abruptly due to the blockage of the obstacle during the movement, /3(^, ) will jump. This sharp change process mainly occurs when the propagation path meets the edge of the obstacle. After the intersection, the originally accessible path is blocked or the originally blocked path becomes reachable. At other times during the movement, the propagation path will also undergo a relatively slow change, resulting in a slow change. According to this principle, it is possible to make a judgment as to whether or not the source is a source of movement. In the case where it is confirmed that there is no co-channel heterogeneous signal and the transmission source is not a moving source, if Ζ^ 'Μ^,Γ) fluctuates greatly with time, it can only be caused by a change in the propagation environment. This does not happen in a normal communication environment. But if we construct a relatively closed electromagnetic environment and place a wireless signal source and one or more of the systems of the present invention in it, we can detect whether new obstacles appear in the environment and whether obstacles move. . Of course, the transceiving between the wireless signal source and the receiving system can be frequency hopping to prevent external narrowband electromagnetic interference.

 After a thorough analysis of the signals in the frequency band under investigation, the next direction finding process can begin. The technical solution of the present invention focuses on three key technical features: one is to synchronously receive signals by using the first antenna and the second antenna, wherein at least the first antenna is a rotatable directional antenna; and the second is to use a certain length of time window. Short-term statistical analysis; The third is to use the ratio of the short-term statistical analysis values to complete the analysis and direction finding. Since the technical solution of the present invention utilizes the ratio analysis method, the specific implementation method of the short-term statistical analysis is not a necessary technical feature for implementing the technical solution of the present invention, and a wavelet can be used in addition to the short-time Fourier spectrum analysis. Analysis and other analytical methods, that is, analysis at various scales in the scale domain.

The traditional method of amplitude measurement of rotating antenna is affected by many factors, and the accuracy and reliability of direction finding are very limited. These factors include changes in signal levels during rotation, multipath propagation, and interference from coherent heterogeneous signals. Traditional methods only consider the statistical maximum and minimum values of the passing signals (ie, large and small points). To determine the direction of the incoming wave, there are major limitations. The multi-antenna unit signal analysis processing and direction finding system of the present invention performs the accurate measurement of the signal by combining the multi-antenna unit and the directional function of the directional antenna on the basis of the above-mentioned signal analysis processing. to. The influence of the multipath signal is excluded, and the orientation of the direct signal is obtained. One of the main methods is the amplitude direction finding method based on the amplitude directional function superposition model regression analysis. Among them, the regression analysis is a kind of optimization method for numerical analysis. Here, numerical analysis methods such as least squares method can also be used.

According to the aforementioned signal / ₂ (0 propagation model, keep the auxiliary rotating antenna fixed at a certain angle and roughly point to the possible incoming wave direction, and the main rotating antenna starts from a certain angle = . at a certain angular rate "rotation. The amplitude directionality function of the rotating antenna is ί ^). Considering the case where the main rotating antenna rotates once a second, if the time window of the short-time Fourier transform is selected as lms, the angle at which the antenna rotates in the time window is 0.36°. The directional function can be seen as a constant in this process. Of course, this condition can be better satisfied by slowing down the rotation speed of the main rotating antenna or shortening the time window. If the time window - ) is regarded as a constant, the short-time Fourier spectrum q( t,r) of the signal measured by the main rotating antenna in a certain time window can be approximated as:

F _l ( jw, t, τ) Ύ _Κ _Χί φ{ω ΐ― θ _Χί )e~ ^]wTl ' ] + F ₂ ( jw, t, τ)∑[Κ _2ί φ(ω ί θ _2ί )e~ ^]wT2 ' ]

It is at some point w. The spectral value can be expressed as

F _l ( jw ₀ , t, τ) ΐ κ _λί φ{ω ΐ θ _Χί )e ^lw ^ ] + F ₂ ( jw ₀ , t, τ)∑[Κ _2ί φ(ω ί θ _2ί )e~ ^]wT2 ' ]

The short-time spectrum of the signal measured by the auxiliary rotating antenna is at w. The spectral value can still be expressed as:

Still consider the short-time spectrum of the signal measured by the two antennas in the time window at w. Single frequency point spectral ratio:

This ratio can also be expressed as a logarithmic difference. The effective degree of (jV., t, τ) here is determined by G ₂ (>v., t, τ), which can be set in practical use. If (3⁄4(> _3⁄4 ) is less than the constant " ₃ , then it is considered The ratio has no validity. - The number is determined not to be unique. It can be selected according to the actual measurement environment, such as a value greater than 2~5db of background noise, or 1/4 of the statistical average of the spectrum of the frequency signal.

If there is no co-channel heterogeneous signal in space, ie Λ (0 is also a multipath signal of / O or Λ (0 = ο, here both are reduced to / ₂ (0 = 0.) The weighted superposition model is:

Here, each incoming signal has three parameters: amplitude coefficient ^: _u , incoming wave azimuth and delay time r _h ., where the amplitude coefficient ^: _u may no longer reflect the original relative amplitude of the signal in space, but Multiplied by a number of different coefficients during the downconversion process. By establishing a model, an optimal solution for the measured data can be obtained by using an optimal analysis method of numerical analysis. For example, a nonlinear optimization numerical analysis method can be used to perform regression analysis on the measured actual data according to the above model, thereby obtaining an accurate solution vector, and some other types of numerical methods can also be used to obtain an optimal solution. Rotating the main rotating antenna for one week, you can measure a set of data in / ^ '1⁄4 ^ in each direction, for example, lms is a time window, and Is rotating will measure about 1000 data. According to the results of the previous signal analysis, when there is no co-channel interference and the signal propagation path is stable, the /3 measured by the rotation of the main rotating antenna (^ data will be in the angle of 2 τ for the period, in fact, it can be rotated a few times. Week, and then select the data with higher validity index to form the ratio data curve corresponding to each azimuth angle. By analyzing the data curve, the number of directions of the wave direction and the corresponding amplitude, orientation and direction of each direction wave can be estimated. For example, the data measured by one rotation of the directional main rotating antenna can estimate the envelope position and amplitude of each incoming direction signal, and the delay is generally arbitrarily within a certain range, and the amplitude is obtained. The delay of the maximum path can be taken as 0. The estimated number of directions of the incoming wave is taken as the dimension of the model (ie, the number of directions of the incoming wave), and the estimated values of the parameters are substituted into the model as the initial value for optimization operation, such as using matlab software. The nonlinear fitting algorithm described by the nl inf it () function performs the optimization operation to obtain the optimal solution and the error value under the optimal solution. If the error value is large, the dimension of the model can be further optimized. The general method is to subtract the estimated value of the low-dimensional model from the original data, and then use the difference to estimate the new possible incoming signal. In this case, according to The result of the optimal solution of the incoming wave signal changes with time, and it can also be judged whether the signal source is a moving source. It is necessary to emphasize that the delay time for solving the above model has a certain uncertainty, and its value is ² , which is a period. When "small, the range of values of Γ _1; can be used to exclude the blur caused by periodicity, which is also a benefit of signal baseband processing.

When we obtain the arrival direction of each incoming wave direction, and the delay time of 7., we can achieve the work of single station positioning. can. The principle is as follows: Assume that the position of the source is A, the position of the system is B, and the position of the multipath signal with only one reflection is C, and the angle between AB and BC is measured.

\AC\ =

And + = cAT, where c is the speed of light, ΔΓ is

The delay difference between BC and AB direction, ΔΓ can also be calculated from the previous solution result, so if the value can be measured, the value can be obtained, and the azimuth of ΑΒ is known, thus determining the position of the source Α. . The measurement can measure the position of the reflector intersecting the BC direction and can be realized by physical ranging such as line-of-sight measurement.

 In the absence of co-channel interference, due to the influence of multipath signals, the main rotating antenna is measured at a constant speed.

Z v. , , r) There may be some high frequency rapid fluctuations that affect Z v. , , r) the smoothness of the curve, and may also affect the regression analysis. Since the main direction range of the Fourier transform of the standard pattern curve and its multipath superposition curve is limited, it can be 对. , , Γ) The data is filtered, and the extra components (generally high frequency components) are filtered out to obtain a smoother curve, and then regression analysis is performed.

 When co-channel interference exists, the data required for direction finding of one of the signal sources can be obtained by a method of comprehensive selection in the frequency domain, the air domain, and the time domain. The characteristics of this type of data are relatively smooth, conforming to the characteristics of the standard pattern curve and its multipath superposition curve, and the error obtained by regression analysis is small and the model is highly consistent. If the resulting set of data (not necessarily a complete cycle) meets the above characteristics, it can be called a coincidence test. When co-channel interference exists, the spectrums of the two sources overlap, but there may be only partial overlap. Through the previous signal analysis, we can find the unoverlapping parts and select the frequency points from them. Direction finding. If the frequency domain selection is difficult to achieve better results, we can consider using two directional antennas for airspace selection, that is, the auxiliary rotating antenna is directed to one of the signal sources, and try to avoid the incoming wave of the other signal source. If the methods selected in the frequency domain and the airspace do not achieve the expected results, we can comprehensively adopt the methods of frequency domain, airspace and time domain selection. That is, the auxiliary rotating antenna is pointed to a suitable position or an omnidirectional antenna is used as described above, and the main rotating antenna is repeatedly rotated at a constant speed for each frequency point.

^Measured, the curve is selected. If a certain frequency component of the same frequency signal disappears or is extremely small within a certain period of time, it is possible to obtain the curve data that meets the direction finding requirements. The frequency domain analysis method can also be replaced by a wavelet transform or the like. When the same-frequency heterogeneous signal exists but the signal is small, it is also conceivable to filter the β(., data, filter out those high-frequency components to obtain a smoother curve, and then perform regression analysis. For the same-frequency heterogeneous The interference signal can be the direction in which the maximum fluctuation of the curve (the azimuth corresponding to the point at which the amplitude spectrum is instantaneously changed) is the maximum possible direction of the incoming wave. - - one of the preferred embodiments of the invention:

 In this embodiment, signal analysis and signal source direction finding are mainly performed by a combination of a directional antenna and an omnidirectional antenna. As shown in FIG. 3, the first antenna 1 can adopt directional antennas of different shapes, but should be light and easy to carry. The double-strand cable 14 is used to transmit two different antenna signals. The anti-coupling bracket 16 is used to keep the two antennas on different horizontal planes and keep a certain distance between them, so as to prevent the antennas from interacting with each other when receiving signals. Of course, the antenna unit 1 can also be designed in other ways. For example, when the directional antenna adopts a composite loop antenna, the output of the calibrated antenna in the composite loop antenna can be simultaneously used as the output of one omnidirectional antenna. The handle 13 in Fig. 3 facilitates hand-held direction finding and the bracket 15 is used to mount the directional antenna 11. The anti-coupling bracket 16 can also be bent back in a certain arc and has a large height so that when the user holds the device, the omnidirectional antenna 12 is vertically above the top of the user's head.

 The following is a brief description of the use of the system. The hand-held direction finding device rotates the measurement results to screen out the possible signal directions, and the directional antennas are sequentially pointed to these directions. Then, based on the measured values of the short-time spectral ratios of the signals measured by the directional antenna and the omnidirectional antenna at each frequency point, the presence or absence of the same-frequency heterogeneous signal is analyzed. When the directional antenna receives a relatively small signal in a certain direction of the wave, so that the ratio change is not significant and the system deviation is easy, a fixed gain can be added to the directional antenna receiving path, or multiplied by a fixed multiple before the ratio is taken. Make (iv., t, r) change relatively significant. If there is a large continuous fast fluctuation over time, it is considered that there is a large co-channel heterogeneous signal; if the fluctuation occurs with time, and the fluctuation becomes smaller when the time window width is increased, it means that the direction does not exist. The same-frequency heterogeneous signal or the same-frequency heterogeneous signal is small; if the index fluctuation is moderate in most of the time, and only a large jump occurs at some time, the signal source may be the moving source.

Figures 4 and 5 are graphs showing the results of one experiment. In the experiment, the three-way frequency modulated actual voice signal is used. The three voice signals are different. One of them is modulated in a lower frequency band, and only the multipath signal interference exists in its passband. The other two The road signals are modulated in the same higher frequency band and constitute the same frequency interference. Wherein the direction of the directional antenna diagram in FIG standard heart-shaped directivity function / (^ = + ∞ _8, wherein, k _0, k _x are constants, the multipath reflection coefficient is typically 0.5, multipath The angle between the arrival direction of the signal and the direction of arrival of the direct signal is typically 45 degrees. Figure 4 shows the maximum and minimum values of the short-time spectral ratio Ζ , , τ) of each frequency point over time in time. The maximum value is shown by the solid line, and the minimum value is shown by the dotted line. Among them, some frequency points with smaller amplitude values have a short-time spectral ratio >^, which is directly assigned to a zero value. In Figure 5, the * line shows the short-time spectrum ratio at a certain frequency in the lower frequency band, and Γ) changes with time. The frequency is affected by multipath interference, and the index fluctuation is relatively flat; The solid line in 5 shows the short-time spectral ratio Ζ , , , Γ at a certain frequency point in the higher frequency band as a function of time. The frequency point is more severe due to the presence of co-channel interference. - - When the same-frequency heterogeneous signal is not found for each possible incoming wave direction, it can be judged that there is no interference signal of the same-frequency heterogeneous signal or the same-frequency heterogeneous source in the space.

If there is no co-channel hetero-source signal in the space, or the co-channel hetero-source signal is small, as analyzed in the previous section, there are:

 When the signal measured by the omnidirectional antenna has a zero or very small amplitude at the wo frequency, the ratio loses its validity and a weak warning sign should be given. Since the antenna signal is generally down-converted to a baseband signal and then processed, the value of 1⁄43⁄4 is generally small, and in the case of handheld direction-finding equipment, the multi-path signal delay with a relatively large amplitude is generally small, so e can Approximate is taken as 1. IJ:

 At this time, as shown in Figs. 6 and 7, in the case of uniform rotation, D0. , i, r) curve can be regarded as a standard directional function with different amplitude coefficients, different angular offsets and the same angular expansion coefficient. In this case, it is easier to obtain DO'i3⁄4,i,r) Correspondence between data and antenna angle. In the case of non-uniform rotation, if the direction annotation function is taken as = cos( , the superposition result can be expressed as ^ 0 + ^ , 008(6^-^), which can be reduced to the cosine form, where , . , are all constants. The corresponding relationship between the amplitude and the angle of the cosine function can be used to obtain the corresponding antenna angle of each >3⁄4, that is, the DO'iv., i, r) the amplitude of the curve and the amplitude of the cosine function are the same. The angular value of the cosine function at this point, so that the exact solution of each incoming wave direction signal can still be obtained by regression analysis of the model. Preferred Embodiment 2 of the present invention

 The amplitude of the general rotating antenna is affected by various factors such as the change of the signal level during the rotation, the interference of the multipath signal and the same-frequency heterogeneous signal, and only considers the statistical maximum or minimum value of the signal (ie, the large sound point). And small sound points) to determine the direction of the incoming wave, there are major limitations.

 In this embodiment, the main rotating antenna rotates with time, the auxiliary rotating antenna (substantially aligned with the incoming wave direction) or the omnidirectional antenna is fixed, and the signal level measured by each antenna and the angular position of the main rotating antenna are simultaneously recorded.

Short-time spectrum of signals measured by two antennas in each time window under the condition of single signal (without multipath signal) - - At frequency w. The single-frequency point spectral ratio and the angular position of the main rotating antenna theoretically constitute a pattern of the main rotating antenna. Under the condition that the multipath signal exists and the coherent heterogeneous signal does not exist, the above calculation result can be regarded as a weighted superposition of a plurality of patterns. By establishing a model and using the optimization method to perform regression analysis on the model, an accurate solution of each incoming wave direction signal can be obtained.

 When co-channel interference exists, the data required for direction finding of one of the signal sources can be obtained by a method of comprehensive selection in the frequency domain, the air domain, and the time domain. Third preferred embodiment of the present invention

 Under the condition of single signal (or multipath signal is small), the direction of the incoming wave can also be measured by the fixed application of the omnidirectional antenna and the two directional antennas, and the two directional antennas are roughly aligned with the direction of the incoming wave, and the two directional antennas Open a certain angle between them. By pre-calibration, when the directional antenna is aligned with the incoming wave direction, the directional antenna and the omnidirectional antenna synchronize the measured short-time spectrum at the frequency point w. The ratio of the spectrum is /). When the signal is off the forward angle of the directional antenna, the ratio is

D. Where is the directional function, and ^) )=1. Considering the symmetry, the short-time spectrum of the signal is measured at w according to the synchronization of the two directional antennas and the omnidirectional antenna. The spectral ratio can uniquely determine the direction of the incoming wave. This application is suitable for single small and instant signals.

Claims

WO 2011/085629 _ ^ g_ PCT/CN2010/080138 Claims

 A multi-antenna unit signal analysis processing method, comprising the steps of:

 Step 1, using the first antenna and the second antenna to synchronously receive signals, at least the first antenna is a directional antenna, and the two antennas have different directivity;

 Step 2: Perform short-term statistical analysis on the signals received by the first antenna and the second antenna simultaneously in a certain length of time window, and obtain a ratio of the short-term statistical analysis values of the two in an analysis point of the corresponding analysis domain. Wherein the length of the time window is adjustable and greater than the longest multipath delay of all signals;

 Step 3: Determine a composition of the signal and/or an interference strength according to the variation characteristic of the ratio.

2. The multi-antenna unit signal analysis processing method according to claim 1, wherein in step 3, the validity of the ratio is first determined, if one of the short-term statistical analysis values is less than a constant, and the other is short. When the statistical analysis value is less than the constant " ₂ , then the ratio has no validity; then, the composition of the signal is determined according to the variation characteristic of the effective ratio.

The multi-antenna unit signal analysis processing method according to claim 1 or 2, wherein the method for determining the signal composition in step 3 is: if the ratio does not substantially change with time, at the analysis point There is only one signal source, the signal being a direct signal and a multipath signal of the signal source, or only a direct signal of the signal source; if the ratio changes rapidly with time and/or the ratio is substantially insensitive to the length of the time window Characteristic, the same frequency heterogeneous signal exists at the analysis point; and/or, if the ratio changes stepwise over time, the signal source is a mobile source or a different signal source of multiple simplex communication .

The multi-antenna unit signal analysis processing method according to claim 1 or 2, wherein in step 3, the intensity of the interference to the signal is determined according to the statistical value σ ² of the magnitude of the fluctuation of the signal, and the calculation of σ ² The way is:

σ ² = ; Z)( , t ., r) is according to step 2

The ratio of the first time window calculated at an analysis point; for a natural number, 1≤ ≤M, M is a natural number greater than 1; when the statistical value is greater than a constant ς, then the signal must have a co-frequency heterogeneous signal.

A multi-antenna unit signal direction finding method, comprising the steps of:

Step 1: Synchronously receiving signals by using first and second antennas, wherein the first antenna is a rotatable directional antenna, The direction of the first antenna can be relatively independently adjusted, and the second antenna is an omnidirectional antenna or a fixed directional antenna, and when the signal is received, the first antenna is rotated;

 Step 2: Perform short-term statistical analysis on the signals synchronously received by the first antenna and the second antenna in a certain length of time window, and obtain a ratio of the short-term statistical analysis values of the two in an analysis point of the analysis domain, where The length of the time window is greater than the longest delay of all multipath signals;

 Step 3: Determine the correspondence between the ratio and the azimuth, and obtain a data curve of the ratio with respect to the azimuth.

The multi-antenna unit signal direction finding method according to claim 5, further comprising the following step 4: Step 4, estimating a possible number of incoming directions according to the data curve, and each of the incoming wave direction signals Parameters, and the determined values are used as initial values. The weighted superposition model of the directivity function of the first antenna is used to solve the measured data curve by using an optimization method of numerical analysis to obtain the incoming wave direction number and each incoming wave direction signal. The optimal optimization value for the parameter.

The multi-antenna unit signal direction finding method according to claim 5 or 6, wherein in step 3, the validity of the ratio is first determined according to a short-term statistical analysis value of the signal received by the second antenna. If the short-term statistical analysis value of the signal received by the second antenna is less than a constant " ₃ , the ratio is not valid; then, according to the magnitude of the validity, each of the same azimuth angles obtained from the multiple-period scanning In the ratio data, the data with the strongest selection is selected to form a data curve of the ratio with respect to the azimuth.

The multi-antenna unit signal direction finding method according to claim 5 or 6, wherein the first antenna adopts a directivity function of ^ + COS (where form, wherein k is a constant, Azimuth; In step 4, the data curve is first reduced to a cosine function form, and then the antenna angle corresponding to each amplitude spectrum ratio is obtained according to the correspondence between the magnitude of the cosine function and the angle.

9 . The multi-antenna unit signal direction finding method according to claim 6 , wherein in step 4, the ratio data curve is first filtered to estimate a possible number of incoming directions, and each incoming direction signal parameter. 9 . .

10. A multi-antenna unit signal analysis processing and direction finding system, comprising: an antenna unit, a receiving unit, a data processing unit, and a display unit, wherein: the antenna unit includes at least two antennas, and at least one of all antennas is The directional antennas, the orientations of the directional antennas can be relatively independently adjusted; the signals measured by the antennas are respectively input into the data processing unit through the receiving unit for synchronous analysis processing, and the data processing unit pairs the directional antenna and another Synchronizing the received signals of the antennas to perform short-term statistical analysis within a certain length of time window, and obtaining a ratio of the short-term statistical analysis values of the two at an analysis point of the corresponding analysis domain, wherein the length of the time window is greater than Place WO 2011/085629 _2Q. PCT/CN2010/080138 has the longest multipath delay of the signal; the result of the data processing unit analysis processing is input to the display unit for output display.

11. The multi-antenna unit signal analysis processing and direction finding system according to claim 10, wherein: the antenna unit comprises a dual antenna combination comprising a pair of directional antennas and a pair of omnidirectional antennas.

12. The multi-antenna unit signal analysis processing and direction finding system according to claim 10, wherein the length of the time window is greater than 4 microseconds.

The multi-antenna unit signal analysis processing and direction finding system according to claim 10, wherein the short-term statistical analysis value is a short-time Fourier spectral spectrum value, and the analysis point is a frequency point.

The multi-antenna unit signal analysis processing and direction finding system according to any one of claims 10 to 13, wherein the receiving unit further comprises a mixing circuit and a lock controlled by the data processing unit Phase loop, each signal input to the receiving unit is first input to the mixing circuit for down-conversion to form a baseband signal corresponding to each, and each baseband signal is input to the data processing unit for short-term statistical analysis; When the data processing unit starts the short-term statistical analysis, the local oscillator signal phase lock signal is sent to the receiving unit, and the phase-locked loop stops the tracking and capturing process of the signal frequency phase, so as to keep the down-conversion of each signal. The frequency and phase of the vibration signal are unchanged.