DE3722716B3 - Pulsdopller radar method - Google Patents

Abstract

A time window covering the duration of the envelopes of a rotor blade echo signal (single flash) is sequentially supplied with sections of the received echo signal train, which are firstly cross-correlated with a post-subdued, correspondingly timed rotor blade echo signal section and examined on the other hand with regard to the respectively present frequency spectrum. The comparison of the cross-correlation result with a detection threshold indicates whether or not it is a rotor blade echo. After formation of the envelope of the respective frequency spectrum, a recognition of whether it is the spectrum of a helicopter rotor and a classification with respect to the even or odd number of blades of the rotor takes place due to the knowledge of characteristic rotor blade echo spectra. Subsequently, the periodicity of the rotor echoes can be determined. The invention is intended for rapid detection of helicopters, including classification, whereby a high false alarm strength is achieved with good cumulative detection and classification behavior.

Description

The invention relates to a pulse Doppler radar method with measures for helicopter detection and classification, in which the received echo pulse signals are examined after a hard character suppression using a detection threshold for the presence of such echo pulse signals derived from helicopters, so that these signals with additional distinction With regard to origin of helicopters with odd or even number of rotor blades, demarcate signals that do not originate from helicopters. Such a radar method is out US 4,275,369 which will be discussed later.

The radar echo signal signatures of helicopters differ from those of other moving or stationary objects by a characteristic Radarechomodulation.

While in general the frequency components of this modulation produced by moving parts (speed range up to more than 200 m / s) also lead to a discovery with conventional MTI radar methods, a helicopter classification and possibly the recognition of the respective helicopter type is based on the evaluation of specific characteristics of the radar echo signal bound.

Since the radar echo signal of a helicopter is characteristically modulated by the rotating parts, namely the rotors and their linkages, the modulation frequencies are directly related to the number of blades and the shaft speed of the respective rotor. Depending on the radar wavelength, the particularly effective rotor blades have a more or less sharply focused directional beam characteristic, which sweeps over the location of the radar device with the rotor shaft frequency and is scanned by the radar pulses, corresponding to long rod antennas. If the number of sheets is odd, each sheet delivers twice per turn. If the number of sheets is even, each pair of sheets delivers a bundled echo twice, or 2a or a echo if a is the number of sheets.

The time interval of the echoes from the main and / or tail rotor is a distinguishing feature, another is the shape and symmetry of the individual sheet echo spectra, i. E. whether an asymmetrical or symmetrical echo spectrum is present relative to the helicopter cell.

The time interval between the rotor echoes is only dependent on the number of rotor blades, the blade length and the rotor speed and is typically between 25 ms and 90 ms for main rotors, and between 4 ms and 20 ms for tail rotors. Since the sheet echoes recur with high echo amplitude, but only for a very short time ("rotor flash") and unfavorably large distances for detection, the tail rotor echoes may also be shadowed by the cell, a pulse Doppler radar to classify must pass through these conditions correspondingly long target illumination times (TOT) and a sufficiently high sampling rate are adapted.

If smaller target illumination times are chosen, the likelihood that the rotor will be scanned by the radar at precisely the time when it would coherently backscatter the energy absorbed in the direction of the radar, that is, is less likely. the echo signals coming from the rotor have a detectable signal power.

Cumulatively then several target lighting times must be considered until the return of an echo. After the first detection of a characteristic echo, temporal expectation spaces are defined, which result from the next target illumination times at which rotor echoes are possible at the same time.

There are several suggested solutions in connection with helicopter classification. A proposal (German Patent Application P 36 00 827.3) reaches a classification as a helicopter by time determination between main rotor echoes, which are defined as individual pulses without structure. The minimum target illumination time (TOT) is at least equal to the single pulse distance. A distinction of interference pulses is possible only by multiple identification of the same pulse echo distance.

Other proposals include spectral analysis with high resolution and correspondingly long continuous target illumination time (T0T). For a resolution of the characteristic frequencies (leaf and wave frequencies: a few Hz) target illumination times of the order of 1 second are required.

An other suggestion ( US 4,275,396 ) reaches the helicopter detection and a classification of the helicopter type with respect to even or odd number of rotor blades by distance measurement of the rotor flashes in the time domain, which are evaluated by a special Doppler filtering signals that exceed a certain threshold.

Out DE 30 02 148 For example, a method for classifying moving targets is known in which the echo signal is correlated with the pattern functions of flying objects stored in a sample bank in a correlator.

The proposed methods can not be reconciled with a reasonable data renewal rate with conventional circular search radars.

The object of the invention is to supplement Pulsdopplerradarverfahren known and suitable type so that a discovery and classification of helicopters can be achieved with minimal target illumination times. In this case, a high level of assurance is to be achieved.

According to the invention, this object is achieved by extracting from the receiving side arriving, freed of fixed signal Pulsecho signal train successive, each extending over a first and a second time window area, which cover the duration of the characteristic envelope of a rotor blade echo signal (rotor flash) of the helicopter, extending signal sections be that on the one hand for each occurring during the first time window range N1 echo pulse of a signal section with a temporally corresponding signal portion of a simulated rotor blade echo signal cross-correlation and after a comparison of the respective present cross-correlation result with the local noise mean value dependent detection threshold is decided whether it is each occurred signal portion is a rotor blade echo signal or not, on the other hand for each during the second time window area a A spectrum analysis is performed on the N2 echo pulses and that envelopes of the frequency spectrum resulting for the respective second time window range are averaged and then examined, the distance of these mean values, individually and integrally, based on the noise average of the same radar resolution cell and / or neighboring cells and that the change in these averages is used to distinguish the desired characteristic helicopter spectra from spectra of other target classes and to discriminate between spectra originating from helicopters with an even number or that with an odd number of rotor blades.

The timeslot area T2 , which is assigned to a spectral analysis, is much larger than required for the correlation, since the necessary spectral resolution is defined by 1 / T (T = time window duration). The second time slot area T2 is about 10 times longer than the first timeslot area T1 ,

The essential idea of the invention is thus to supplement suitable radars by a combination of signal processing methods such that with minimum target illumination time (TOT), each echo flash detection first results in a classification which, because of the single-flash processing in the spectral range, also provides the statement as to whether it is is a helicopter with even or odd number of rotor blades. The verification and classification of the helicopter type is done with a flash distance measurement in the time domain. The single-flash detector according to the invention distinguishes between glitches and rotor echoes on the basis of the characteristic shape and duration of the individual flashes. Main rotor echoes and tail rotor echoes - differentiated by the included modulation frequencies, their echo shape and length - are processed separately in the time range for the flash distance measurement. Single-flash detection and in principle also the determination of the flash distance in the time and frequency domain are redundant. The link gives the increased reliability.

After estimating the temporal rotor echo distance, the average flash distance after several flashes observed can then also be determined e.g. calculate according to a predictor-corrector procedure, if lightning could not be detected in individual expectation windows.

A reliability criterion (variance) can be specified.

In the signal processing, both the main rotor echo and the tail rotor echo can be evaluated, which is only amplitude modulated when the plane of rotation is illuminated from the rotational axis, is also phase-modulated under all other aspect angles, with maximum phase modulation when the helicopter illuminates in the plane of rotation of the tail rotor.

The phase-modulated component of the Heckrotorechos can be understood over a large aspect angle, since the associated antenna characteristic of the tail rotor blades is weakly bundled in the azimuth plane. Their spectrum is detected like that of the main rotor echo with the help of a filter bank.

The amplitude modulation caused by the tail rotor can be examined after splitting off by a low-pass filter of suitable cut-off frequency via a correlation to successive echoes, in analogy to the lightning detector for the main rotor. The phase modulated component is treated as that of the main rotor.

• 1 die Zeitfunktion der Einhüllenden eines beispielsweise empfangenen Signalzuges,
• 2 das Frequenzspektrum zu einem Zeitfensterbereich T2, in dem ein Hubschrauber-Echosignal vorliegt, hier ohne Festzeichenunterdrückung (MTI) gezeigt,
• 3 die Zeitfunktion eines nachgebildeten Rotorblattechosignals eines Hubschraubers, das zur Korrelation benutzt wird,
• 4 eine Anordnung zur Durchführung des Verfahrens nach der Erfindung in Blockschaltbildform.

The invention will be explained below with reference to four figures. Show it

• 1 the time function of the envelope of a signal train received, for example,
• 2 the frequency spectrum to a time window area T2 in which a helicopter Echo signal is present, here shown without hard-character suppression (MTI),
• 3 the time function of a simulated rotor blade echo signal of a helicopter used for correlation
• 4 an arrangement for carrying out the method according to the invention in block diagram form.

In 1 an example of a time function of the envelope of a received signal train is shown. It forms T1 the duration of a time window area, which is pushed over the incoming signal. .delta.t is the time range during which occur in this example echo pulses of a rotor flash with significant power. During the time window area T1 occur N1 Echo pulses on. T1 comes from 3 out.

In 2 is that to a timeslot area T2 belonging frequency spectrum shown. It is PRF the pulse repetition frequency of a suitable radar, R the noise average outside the sheet echo spectrum and B the frequency range of the leaf spectrum, with the central peak S essentially from the helicopter cell and the underlying narrow plateau F be formed by the rotor head. The representation is without mark suppression ( MTI ). T2 comes from 1 out.

In 3 an example of a time function of a simulated rotor blade echo signal of a helicopter is shown. This time function applies to the time window area T1 and serves to cross-correlate with each during the time-window area T1 present section of the received signal train. The characteristic time function is not constant over time. Therefore, a set of different replica functions is required. The length of the correlation can be restricted to the time interval _Δt , ie the rotor echo duration between the first zeros of the rotor choke characteristic. The function in the time interval _Δt in the example corresponds approximately to the function sin x / x.

With the given by the invention method of spectral analysis of short periods of time T2 each N2 Echo pulses in preferably K overlapping pulse groups, only one rotor flash is processed. With a the N2 Echo pulses corresponding number of N2 Filters can be used to distinguish low bandwidth targets of the Doppler echo (eg, aircraft, clutter) from high bandwidth targets (helicopters).

Nullen ersetzt werden.For this purpose, a relatively good spectral resolution is required at the expense of the signal-to-noise ratio, which would not be given optimal processing of the few actually received by the rotor coherent, a detectable signal power having n echo pulses. However, on both sides of the expected n rotor echoes of appreciable power, the true samples, which essentially only contribute to the noise, can pass through $\frac{\text{N2}-\text{<figure-callout id="2" label="n" filenames="DE000003722716B3\_0002.png" state="{{state}}">n</figure-callout>}}{2}$

Zeros are replaced.

In detail, the following procedure results in the spectral analysis:

By the overlapping processing already mentioned, the n samples of the rotor echo are roughly centered on the center of a group of N2 samples that span the time window area T2 expand. This is necessary because otherwise the weighting function would suppress the few echoes that contain the rotor modulation.

The subsequent spectral decomposition, with complex transformation for odd numbered rotors, results in an asymmetric spectrum on one side of the helicopter cell's Doppler echo, for even number rotors, more or less unbalanced spectrum depending on the reflective properties of the leading and trailing edges of the rotor blade both sides of the "cell echo" S. The spectra are band-limited. The envelope of each spectrum provides, if properly processed, whether the discovered helicopter has an even or odd number of rotor blades. A signal / noise improvement or a smoothing of the spectrum can be achieved by re-integration of several individual spectra. In this case, the spectra for the alternating echoes of the front and rear edge of a rotor blade are to be integrated separately from one another if the information concerning the even or odd rotor blade number is not to be lost.

The statement "helicopter echo" or "other echo" is made possible by a spectral envelope detector having the following properties:

After sufficient smoothing of the spectra either by a previously suitable calculation method (eg autoregressive spectral estimation with suitable parameters) or by Fourier transformation with subsequent smoothing over averaging, the distance between the envelope becomes below the band limit f _g of the rotor echo compared to the noise average above the band limit f _g or the average noise level - above the cut-off frequency of the rotor head portions - of spectra which lie temporally between the sheet echoes. It is assumed here that the sampling frequency is higher than the frequency 2 f _g lies. In addition, the gradient of power density versus frequency is determined in short sections .DELTA.f on either side of the helicopter cell line, each containing the power averages of several filters. The sections Δf are sequential or overlapping. An approximately constant gradient for f <f _g and f> f _g in conjunction with the distance of the mean values yields the desired statement. The method is equally suitable for hover and cruise flight. In cruise flight, the Doppler frequency of the cell is considered additively.

The above analysis of the frequency spectrum is supplemented according to the invention by an independent detection of the rotor blade echoes and their periodicity in the time domain.

In order to improve the signal-to-noise ratio and to suppress echo formations which do not correspond to rotor echoes, successive overlapping sections of the received signal train (FIG. 1 ) and the temporal excerpt of the simulated rotor echo ( 3 ) first carried out the mentioned cross-correlation. It can also be differentiated between main and Heckrotorechosignalen if necessary.

This correlation is performed for both quadrature channels in complex processing.

After suitable further processing, e.g. Summing the quadraecho cross-correlation function with a local noise average dependent threshold, the characteristic period of the rotor echoes can be estimated and compared to known periodicity data from helicopters of different types.

After a first detection, expectation spaces for possible echo distances can be searched for recurring echoes at a reduced detection threshold. The change in the detection threshold is advantageously made dependent on the signal-to-noise ratio.

In 4 is shown in block diagram form an arrangement for carrying out the signal processing method according to the invention. With this arrangement, by combining both sub-processes a high false alarm resistance is achieved with good cumulative detection and classification behavior. The failure of individual information is replaced by a correspondingly longer evaluation of the remaining one.

The received echo signals are via two quadrature channels I and Q a facility 1 for sign suppression ( MTI ). The device 1 For spectral analysis follows an FM detector formed, for example, by a filter bank 2 , The device 1 are also two correlators 3a and 3b downstream, with which the cross-correlation for analysis in the time domain takes place. As a comparison signals serve the simulated echo signals (replica) of rotor blades (main and tail rotors). Since the time function for the tail rotor echo by the hard character suppression device ( MTI ) 1 at least in a pure side view would be suppressed (no phase modulation), a time function comparison with a replica function before MTI or, as in 4 shown, the MTI Fixed part (f <f _gMTI ) with a replica function. This is done by the AM detector correlator 3b , The moving-sign portion at the output of the MTI device 1 with f ≥ f _gMTI becomes the correlator 3a fed. In the correlator 3a Primarily a cross correlation with main rotor replica takes place. The tail rotor also almost always delivers FM-modulated echoes. In the AM detector correlator 3b can be cross-correlated in addition to the tail rotor AM shares and AM main rotor shares, namely in so far as the flight of a helicopter are conceivable that would lead to a Doppler frequency-free modulation. The FM detector 2 and the correlators 3a and 3b is a threshold detector 4 followed by classifier. Herein, on the one hand, the envelope of the spectrum is generated over a respective time window range, so that a classification can be made as to whether or not the examined signal traction excerpt is that of a helicopter. Information regarding even or odd number of sheets can also be detected. On the other hand takes place in the detector 4 also the further processing with regard to the analysis in the time domain and that for example the summation of the quadrature channels, magnitude formation and peak detection of the rotor echo cross-correlation function with a local noise average value dependent threshold. The detector 4 followed by a license plate generator 5 with which the target identified as a helicopter can be identified with information about the even or odd number of sheets.

The determination of each available helicopter type can in the detector 4 also be made due to a Rotorblitzabstandsmessung, since the typical periodicities of different types of helicopters are generally known and therefore can be used for comparison. The helicopter type can also be via the license plate generator 5 mark.

Claims (10)

A pulse Doppler radar method comprising measures for helicopter detection and classification, wherein the received echo pulse signals after a hard-character suppression using a detection threshold for existence such echo pulse signals are derived from helicopters, so that these signals can be distinguished with additional distinction with respect to origin of helicopters with odd or even number of rotor blades against signals that do not originate from helicopters, characterized in that from the receiving side, arriving from hard tags liberated Pulsecho signal train successive, each over a first and a second time window area (T1 or T2), which cover the duration of the characteristic envelope of a rotor blade (rotor flash) of the helicopter, extending signal sections are removed, firstly for each during the first time window region (T1) occurring N1 echo pulses of a signal section with a temporally corresponding signal portion of a simulated rotor blade echo signal cross-correlation and after a comparison of the respective present Kr euzkorrelationsergebnisses is decided with the dependent on the local noise average detection threshold, whether or not the respective signal section is a rotor blade echo signal that on the other for the N2 echo pulses occurring during the second time window area (T2) a spectral analysis is performed and that of the envelope averages are formed and then examined for the frequency spectrum resulting from the respective second time-window region (T2), the distance of these mean values, individually and integrally with respect to the noise mean value of the same radar resolution cell and / or neighboring cells, and the change of these mean values for distinguishing the characteristic helicopter spectra of spectra of other target classes and to differentiate between spectra used by helicopters with an even numbered or even those with an odd number of roto come from. Pulse Doppler radar method according to Claim 1 , characterized in that following the detection, the characteristic period of the rotor blade echo signals is detected and compared with known helicopter data. Pulse Doppler radar method according to Claim 1 or 2 , characterized in that following a first detection expectation spaces for possible echo intervals are searched for recurrent echo signals at a reduced detection threshold. Pulse Doppler radar method according to Claim 3 , characterized in that the change of the detection threshold is made dependent on the signal-to-noise ratio. Pulse Doppler radar method according to Claim 1 , characterized in that overlap each extending over a time window region signal portions sequentially, so as to ensure that in the time window shift belonging to a Rotorblattechosignal n echo pulses are also once approximately in the central region of the time window. Pulse Doppler radar method according to Claim 1 or 5 , characterized in that each (2k - 1) th spectrum and every 2k th spectrum is alternately re-integrated, where k = 1, 2, 3, .... Pulse Doppler radar method according to one of Claims 1 . 5 or 6 , characterized in that the spectra are pre-smoothed either by a suitable calculation method, eg the appropriate autoregressive spectral estimation with suitable parameters or by Fourier transformation with downstream averaging. Arrangement for carrying out the method according to one of the preceding claims, characterized in that the received echo pulse signals are fed via two quadrature channels (I, Q) to a device (1) for solid-state suppression, the one for spectral analysis, for example, an FM detector formed by a filter bank (2) and an envelope of the spectrum generating threshold detector (4) are followed by the classifier and the other for cross-correlation generation in the time domain two the motion sign (f ≥ f _gMTI ) or the fixed-component (f <f _gMTI ) receiving correlators (3a , b) are connected downstream of which at least one simulated rotor blade echo signal required for cross-correlation formation is present, and in that the correlators follow with a second or third input the threshold value detector (4) which is dependent on its local noise average threshold value, to which the crossbar input orrelationsergebnis and the exceeding of the detection threshold value indicating the presence of a helicopter signal is removed. Arrangement according to Claim 8 , Characterized in that in the threshold detector (4) is a device for Rotorblitzzeitabstandsmesasung provided and that a comparison device is provided in this detector for verification and classification of each of the present type of helicopter, in which the measured time intervals are compared with the special characteristic periodicities of different types of helicopters. Arrangement according to Claim 8 , characterized in that the threshold value detector (4) is followed by a number plate generator (5).

Images