FR2758889A1 - LASER PULSE RADAR OPERATING METHOD FOR DETERMINATION OF TARGET DISTANCE AND RADIAL SPEED - Google Patents

Abstract

According to this method carried out with a Doppler evaluation in distance intervals using a device obtaining, for each interval and from Doppler signals of the laser radar present in the form of pairs of sampled amplitude / time values, an average Doppler frequency f | - formed by the direction of the vector sum of several conjugate complex multiplications of two sampling values in the complex plane (real part, imaginary part) the formation of the sum of the autocorrelation function limited to a sector of the complex plane, this sector determined by a destination frequency (f0) and a variation width (DELTAf) being shifted in the direction 0 deg. for decision making (the limitation) being the differentiation determining whether or not the individual vector is within the predetermined area. Application in particular to the detection of small particles of dust in the atmosphere.

Description

METHOD FOR OPERATING LASER PULSED RADAR

  FOR DETERMINING DISTANCE AND SPEED

RADIALE OF TARGETS

  The invention relates to a method for operating a laser pulse radar for determining the

  distance and radial velocity of targets, with a

  Doppler evaluation, performed on the reception side,

  distances by means of a frequency analysis device, which for each

  distances determined by means of the autocova-

  couples of pulses, from some Doppler laser radar signals present in the form of

  pairs of amplitude / time values sampled, a frequency

  mean Doppler quence f, which is formed by the direction

  of the vector sum of several multiplications com-

  conjugate plexes of two successive sampling values

  respective directions in the complex number plane, the determination of the direction being performed by means of the formation of the argument using the function arctg2, which by definition has a point of discontinuity when

from the passage of n to -n.

  If we analyze, from the point of view of frequency and distance, Doppler signals from a laser pulse radar, which are present in the form of

  pairs of amplitude / time values sampled,

  the average Doppler frequency of a few sample values, typically 10 to 20 values

  during each distance interval.

  This frequency appears on the basis of the Doppler shift in the direction of the laser beam, which represents the

  radial displacement of targets, particularly targets

  persion, such as small particles of

  in each distance interval.

  Problems arising in the determination of

  Doppler frequency appears due to a high noise level and due to the large bandwidth during the measurement. If, for example, Doppler shifts in the +26 m / s area are to be detected, then the total bandwidth (referred to in the following

  BW band) is equal to about 10 MHz. In a general way

  We have: BW = 4Vmax / À, where A is the wavelength of the laser radiation and Vmax is the maximum speed to be measured. The actual useful signal P, however, has only one

  spectral width (referred to in the supplementary narrow band

BN) of about 1 MHz.

  Another problem when determining the

  Doppler frequency occurs because of weak

  the reception signal due to the small intensities of retrodispersion of small targets (eg very small dust particles) and because of a great measurement distance, for example a few kilometers. If, as is customary, a high resolution of distance is sought, because of this requirement, in many very small intervals of distances, respectively, only a few sampling values are obtained, which can also lead to larger problems. when determining a

Doppler frequency.

  There is an additional problem when determining the Doppler frequency by the fact that

  formation of the average value of several individual signals

  duels is possible only after a preliminary estimate of

  frequency, that is, what is known as incoherent averaging, since in the case of laser pulse radar,

  no phase adaptation is possible between the different

  shots with the laser. The error, which appears under

  the effect of this necessary preliminary estimate of the frequency

  quence, therefore continues to be simply cumulated by addition. An improvement in frequency estimation can be expected only in the case of a

  accumulation of "good" frequencies.

  For example, it is known from German Patent Application 32 36 101 A1 to determine the radial velocity of a target in a pulse radar apparatus, in the range of distances using a device which is sensitive to the phase and which determines the vector phase shift, which appears under the effect of the radial displacement of the target, between two successive echo pulses (pulse pair). To obtain the frequency

  Mean Doppler, from which we obtain directly

  For information on the radial velocity of the detected target in the respective range of distances, filter blocks or better FFT (Fast Fourier Transform) evaluation may be used. This known process

  However, it has the disadvantage of a relative

  badly in the laser pulse radar.

  Another known method for determining the frequency

  average Doppler quence during the range of distances

  respective is what is called the autocovarian-

  this with pulse pairs, which has a net return

  higher than the traditional Fourier transformation

  nelle. The average Doppler frequency f is provided by the direction of the complex vector of the function

  of autocorrelation in the plane of complex numbers con-

  fs n-1 formally to f = - arg (Z xi * xi + 1), fs representing the 2n i = 0

  sampling frequency, xi + 1 a measurement value com-

  plex and x i the previous measurement value of a pair of pulses in conjugate complex form. The spectrum of the measurement signal is also complex and ranges from

  -fs / 2 to + fs / 2. The average Doppler frequency corresponds to

  therefore sets a number between 0 and n for positive frequencies and at an angle between 0 and -n

  for negative frequencies. (See Day Meteor Appl.

22, pp. 1776-1787).

  The desired average Doppler frequency is therefore formed by the direction of the vector sum of several complex conjugate multiplications of successive sampling values. The determination of direction is made by the formation of the argument using the function "arctg2", which has by definition a discontinuity for i to -Z. This point of discontinuity leads to too widely dispersed estimates of

  frequencies for a low signal-to-noise ratio, i.e.

  say bad signal / noise ratio and / or high frequencies. The aim of the invention is to establish the method, which is based on the autocovariance algorithm of a pair of pulses, explained above, to determine the average Doppler frequency in the case of pulsed radar.

  laser so that on the one hand we obtain a fil-

  a precise limitation of the bandwidth in the frequency domain, and that on the other hand we obtain an elimination of the points of discontinuity of the function "arctg2"

  In accordance with the invention, which relates to a

  laser pulse radar operation of the type described above, this problem is solved by virtue of the fact that

  that the formation of the sum of the function of autocorre-

  lation is limited to an angular zone, which can be determined, that is to say a sector of the complex number plane, that this angular zone determined by the estimated frequency f0 and a variation width Af, is shifted in the direction 0 before a differentiation of

  case, ie a limitation, that the limitation, that is,

  that is to say the differentiation used to indicate whether or not the individual vector is situated in the predetermined zone, is effected on the absolute value of the direction, and that the formation of the sum is executed exclusively on

  vectors that are located in the limited angular area.

  The limitation, carried out in accordance with the

  the range of assessments provides a limitation of

  the bandwidth, the signal and therefore an improvement

  improvement of the signal-to-noise ratio. Since, in accordance with

  According to the invention, only the vectors which are situated in a limited sector must be involved, and the filtering is much faster and more accurate than previously possible with filters.

  numerical and by means of the transformation of Fou-

  mulberry. The rotation, executed in accordance with the invention, of the sector in the plane at 0 and the study of absolute values

  of angle prevent abrupt discontinuity from 180 to -

  1800. The vectors, which, in a manner conditioned by noise and broadening of the spectrum, could "overshoot", are eliminated in accordance with the invention, so that this point of discontinuity is eliminated. An advantageous embodiment of the invention is that a medium value formation, which corresponds to a consistent average value formation, is performed during another sum formation of the vectors. The formation of the average value within the determination of frequencies over an extended vector sum provides an improvement, weighted by

  the intensity of the signal, the estimate of the frequency.

  Phase matching between individual signals is suppressed due to complex conjugate multiplication

  and the subsequent addition of the vectors in the

  plex of frequencies. This procedure according to the invention

  therefore corresponds for the first time to a

  consistency of the average, in the case of

Doppler laser radar.

  Hereinafter, the invention will be explained in detail with reference to the appended drawing, the single figure of which shows how, in accordance with the invention, the average Doppler frequency in a complex plane of frequencies is obtained.

  quences (abscissae = real part, ordered = part of

  ginaire). For the determination of the average Doppler frequency f during the respective interval of distances,

  uses the pulse torque autocovariance algorithm

sions.

  The average Doppler frequency f is provided by the direction of the complex vector of the function f5 n-1

  of autocorrelation f = - arg (x xi Xi + 1).

  2i i = 0 The scan frequency is fs and xi is the value of

  complex measure. The spectrum of the measurement signal is also

  complexity and ranges from -fs / 2 to + fs / 2. The frequency

  quence f therefore corresponds to an angle between 0 and n for positive powers and between 0 and -n for

negative frequencies.

  The desired average Doppler frequency f is

  consequently formed by the management of the vec-

  of multiple complex complex multiplications of successive sampling values. The determination of direction is made by the formation of the argument using the function arctg2 which has by definition a

discontinuity for +180 to -180.

  According to the invention, the formation of the sum of the autocorrelation function is limited to a

  angular area that can be determined, for example a sec-

  complex plan. This sector is represented in gray

  on the drawing. This angular zone (which corresponds to the

  shown in gray) is defined by a frequency

  Theoretical f0 and the variation width Af. The angular area

  is rotated in direction 0 before the decision

  to be taken for the case, that is to say before the

tation.

  According to a feature of the invention

  the limitation, ie the differentiation to determine whether the individual vector is located in the

  predetermined zone, is carried out using the absolute value

  read from the direction. The formation of the sum is then made only on criteria, which are located in the limited area. It is important that the rotation of the sector represented in gray to come in the direction to 0 and

  the study of the absolute values of the angles prevent a dis-

abrupt continuity from n to -Z.

  We eliminate vectors that are produced by the

  noise and spectrum enlargement and perform a sur-

  cillation in the area represented by hatching in the figures. The point of disruptive discontinuity is of this

  avoided and no longer plays a role in the determination of

  the average Doppler frequency according to

the invention.

  Hereinafter, a suitable embodiment of the determination according to the invention of the average Doppler frequency and therefore the radial displacement of the targets, which in the example are dust particles located in the atmosphere, will be explained in detail. .

  First, the frequency of estimation is determined.

  f0 in the ranges of distances, in which the signal-to-noise ratio is high, for example near the ground or by means of an echo returned by a cloud. The variation width Af can take here the entire bandwidth Bw because of the high signal-to-noise ratio. The estimation frequency f0 can be provided in the case of satellite measurements, also by means of a time forecast

or an atmospheric model.

  The estimated frequency f0 is then transmitted to zones having a lower signal-to-noise ratio, in which case the variation width Δf is firstly still chosen relatively high for the first calculation of the height profile (for example half of the total bandwidth Bw = BW / 2), and is reduced more and more strongly during repeated calculations. The estimated frequency can then be transmitted, after reduction of the bandwidth, to

  average of the variation width Af, and to other

  distance ranges with a signal-to-noise ratio

even weaker.

  The initially high value of the variation width Δf ensures that the algorithm can follow the changes in the atmosphere (wind shear). The lower limit of the variation width Af is fixed

  both by the respective laser radar system (sta-

  the frequency of the fLAS laser) only by the characteristics

  characteristics of the atmosphere (turbulence) and should therefore be at least twice the width

spectral laser radiation.

  In the case of the coherent formation of the average value by broadening the vector sum to neighboring measurements, that is to say in the case of several

  laser shots, the varia-

  (LAS (in radians) the frequency of the laser before the formation of the vector sum, by complex mixing This process can be realized by means of a complex multiplication of the time signal in a much simpler way than in the case of the Fourier analysis usually used, for example by means of a transformation

Hilbert.

  To obtain a new corrected temporal signal xn, we therefore perform advantageously a complex multiplication of the initial temporal signal x (= Xreel, Ximaginaire), and this in the form xnouveauréel = XrelC S (PLAS) + Ximaginairesin (YLAS) xnew .. .. xv imaginary = XimaginaireC S (PLAS) + XréelSin (PLAS) Hereinafter, we will explain again the different steps of the invention and their variants of implementation. The exclusive consideration of only

  amplitude / time sampling values

  when forming the vector sum of the autocorrelation function, which are located inside

  the variation width Af around the estimated frequency

  f0 according to ABS (arctg2 (xi * xi +) - fO-fLAS) <Af, represents precise filtering or precise limitation of

the bandwidth.

  The study of ABS absolute values also eliminates the +180 / -180 discontinuity point for training

  arguments with the arctg2 function.

  Variations in the frequency of the fLAS laser are taken into account during the differentiation of the case, by subtraction during the formation of the vector sum

  by a complex multiplication in the temporal range.

  The extension of the vector sum to neighboring measurements, that is to say in the case of several shots by the

  laser, corresponds to the formation of a coherent mean value

annuity.

Claims (11)

  1. Pulse radar operating method   laser probes for determining the distance and radial velocity of targets, with Doppler evaluation, performed on the receiving side, in intervals of distances   through the use of a frequency analysis device   quences, which for each range of distances determined by means of the autocovariance algorithm of pulse pairs, from a few Doppler signals from the radar   lasers present in the form of pairs of   sampled time, an average Doppler frequency f, which is formed by the direction of the vector sum of several conjugate complex multiplications of two respective successive sampling values in the   plan of complex numbers, the determination of the   tion by the formation of the argument using the function arctg2, which by definition   a point of discontinuity when passing from n to -a,   characterized in that the formation of the sum of the autocorrelation function is limited to an angular area that can be determined, that is to say a sector of the plane of the complex numbers, that this angular zone, determined by the estimated frequency f0 and a variation width Af, is shifted in the direction 0 before a differentiation of   case, ie a limitation, that the limitation, that is,   that is to say, the differentiation used to indicate whether or not the individual vector is situated in the predetermined zone, is effected on the absolute value of the direction, and that the formation of the sum is executed exclusively on   vectors that are located in the limited angular area.   2. Pulse radar operating method   Laser siezes according to claim 1, characterized in that a medium value formation, which corresponds to a consistent formation of the average value, is performed from a plurality of individual signals or from fairly long intervals within the range. an individual signal, by means of   additional training of sums of vectors.   3. Pulse radar operating method   lasers according to one of claims 1 or 2, which is   characterized in that the estimated frequency fo is determined in the range of distances, in which the signal-to-noise ratio is high, the variation width Af may extend over the entire bandwidth due to high signal-to-noise ratio.   4. Pulse radar operating method   Laser siezes according to claim 3, characterized in that the estimation frequency fo is transferred to areas in which the signal-to-noise ratio is lower, the variation width Af being firstly still relatively large when first calculation, that is to say on the order of half of the total bandwidth, whereas it is reduced more and more during the repetition calculations.   5. Pulse radar operating method   laser seals according to claims 3 and 4 taken in   together, characterized in that after the reduction of   the bandwidth by the variation width Af, the frequency   estimated frequency f0 is transferred to other ranges of distances with an even higher signal-to-noise ratio low.   6. Pulse radar operating method   laser lasers according to one of claims 1 to   characterized in that the lower limit of the variation width is at least twice the width spectral laser radiation.   7. Pulse radar operating method   laser lasers according to claim 2, characterized in that in the case of the coherent formation of the average value by means of the extension of the vector sum to neighboring measurements, that is to say in the case of several   shots with the laser, the variations, which may appear   the frequency of the laser (yLAS in radians) being eliminated before the formation of the vector sum, by complex mixture.   8. A method of operating a laser pulse radar according to claim 7, characterized in that for obtaining a new corrected time signal x n ueau, a complex multiplication of the original time signal x (= Xreel, Ximaginaire) is executed in the form xn uveauréel = XrelcoS ((PLAS) + XimaginarSin (gLAS) xnew. = x * o () + xsny xnv imaginary XimaginaryC S (WLAS) + Xrealin (LAS)   9. Operating method of pulse radar   laser lasers according to one of claims 1 to   8, characterized in that variations of the frequency of the laser are taken into account by means of a subtraction, when differentiating cases.   10. Pulse radar operating method   laser lasers according to one of claims 1 to   9, characterized in that it is used for measuring the   distance and displacements of dispersion targets.   11. Pulse radar operating method   laser sensors according to claim 10, characterized in that the dispersion targets to be measured are small   dust particles in the atmosphere.