WO1986003014A1

WO1986003014A1 - A mach-zehnder acousto-optic signal processor for electronic support measures

Info

Publication number: WO1986003014A1
Application number: PCT/AU1985/000276
Authority: WO
Inventors: Ian Gordon FUSS
Original assignee: The Commonwealth Of Australia, Assistant Secretary
Priority date: 1984-11-12
Filing date: 1985-11-12
Publication date: 1986-05-22
Also published as: GB2183822A; GB8616375D0

Abstract

A method of and a means for acousto-optic signal processing for electronic support measures in which a signal is received on a two element spaced array (E1-E2) antenna and applied to Bragg cells (BC1-BC2) in a Mach-Zehnder interferometer to produce a first optical beam (P1) from one element of the array (E1), and a second beam (P2) which is a combination of the signal from the first element (E1) and the second element (E2) of the array, and passing the beams through opto-electrical transducers (L1-L2) and linear photo detector (D1-D2) to a computer.

Description

A MACH-ZEHNDER ACOUSTO-OPTIC SIGNAL PROCESSOR FOR

ELECTRONIC SUPPORT MEASURES INTRODUCTION:

This Invention relates to a signal processor for simultaneous direction finding and spectral analysis of radio signals.

The new generation of battlefield electronic support measures (ESM) receivers require extensive and rapid signal processing facilities. Optical processors which provide parallel processing, powerful operations, and a high throughput rate, are strong contenders for this signal processing role. An excellent example of such a processor is the acoustooptic (A0) spectrum analyser, which exploits the Fourier transform properties of a lens to provide a wideband power spectrum in real time. Such a processor is described by D.L. Hecht 1976 "Spectrum analysis using acousto-optic devices" SPIE 90 148-157. The A0 spectrum analyser performs tasks which are significant for ESM, it sorts radio signals on a frequency basis, plus provides signal parameters such as power and centre frequency. Further, this processor offers a number of desirable features such as a 100% probability of intercept, small size and weight, plus low cost and high reliability. However, it fails to provide direction of arrival (DOA) signal parameter measurement. This parameter is significant for ESM.

A number of A0 signal processors for simultaneous spectrum analysis and direction finding have been proposed and investigated. Perhaps the simplest of these uses the variation in gain of a directional antenna with signal incidence angle from boresight, see for instance A.E. Spezio 1978 "Acousto-optics for Electronic Warfare Applications" 12th Asilomar Conference on Circuits Systems and Computers, IEEE 596-608. In this type of ESM receiver, two or more antennas are placed with boresights pointing in different directions. The outputs of these antennas are independently spectrum analysed and a comparison of a particular signal's power from each antenna is used to determine its DOA.

For a fixed number of antennas and field of view, the DOA accuracy of such a system is limited by radio back scattter, antenna pattern distortion and noise in the radio section of the receiver. These problems are reduced by using phase interferometer methods for DOA determination.

An A0 signal processor for a phased array antenna is described by L.B. Lambert, M. Arm and A. Aimette

1965 "Electro-optical Signal Processors for Phased Array Antennas" Optical and Electro-optical Information processing ed Tippet, MIT 715-748. This system uses a linear periodic antenna with elements spaced a distance d apart. Each element is connected to an

A0 cell. The cells are also placed in a linear periodic array but with a spacing

, where λ is the optical wavelength, f is the rad equency and c the velocity of the radio signal incident on the antenna. These

A0 cells are arranged so that they deflect light in a common direction which is perpendicular to the array axis. The phase of a radio signal applied to an A0 cell is imparted to the light deflected by it. Thus, a phase difference in the excitation of two antenna elements results in a corresponding phase difference in the light deflected by the two cells connected to these elements. Hence, for each frequency f the light in the upper first diffraction orders of the A0 cells forms a beam which has the same radiation (far field) pattern as the antenna. By placing a lens after the A0 array, the deflected light forms an image at the focal plane of the lens which is the frequency, angle of arrival spectrum of the radio signals incident on the antenna. While in principle this is an admirable signal processor, the close spacing of A0 cells can result in cross talk due to acoustic diffraction effects, see A. Vanderlugt, G.S. Moore and S.S. Mathe 1983 "Multichannel Bragg cells: compensation for acoustic spreading" Applied Optics 22 3906-3912. Also the large number ( -100) of elements needed to obtain good angular resolution over a wide angle of view makes an ESM receiver based on this type of signal processor both large and expensive.

In order to reduce the number of elements needed to maintain accurate DOA information, aperiodic devices based on the same concept have been proposed and investigated. Typically, these aperiodic A0 signal processors have four or five elements.

However, the A0 array is a wavefront splitting interferometer and by using aperiodic spacing there are large regions of, the wavefront υndeflected, hence this device has a poor diffraction efficiency. Further, this type of signal processor does not produce an angular spectrum at the focal plane of the lens but a pattern which has to be interpreted by a postprocessor in order to obtain DOA information. The throughput rate of this signal processor is limited by the readout rate of the two dimensional photodetector array required to convert the optical information into electrical form, and the time required by the postprocessor. With the advent of frequency hopping communications systems on the battlefield, ESM receivers with high throughput rates are required.

GENERAL BASIS OF THE INVENTION:

The Mach-Zehnder A0 signal processor, the subject of this invention, has a higher throughput rate than the linear phased array processors. This increase in throughput rate is achieved by using an optical configuration which only requires two or three linear photodetector arrays to convert from optical to electrical signals, and which needs a simpler postprocessing algorithm to extract DOA information. The Mach-Zehnder A0 signal processor consists of two A0 cells mounted in opposite arms of a Mach-Zehnder (MZ) interferometer. In this type of interferometer, the spacing between the

A0 cells does not enter the interferometric equations and hence cross talk due to acoustic interaction between A0 cells can be eliminated and acoustic beams of any size can be used. Thus it is not necessary to construct special multitransducer cells as required by the linear array pr'ocessors. The MZ interferometer is an amplitude splitting interferometer and most of the light interacts with the A0 cells, hence a high diffraction efficiency is maintained for the MZAO signal processor.

The MZAO signal processor cannot sort two signals of exactly the same frequency on a DOA basis as the phased array signal procesors can, but it is unusual for two communications radios to operate at the same frequency simultaneously. The invention comprises a method of acousto-optic signal processing for electronic support measures comprising receiving a signal on a two element array having its antenna spaced apart, converting the two electromagnetic signals to optical signals and applying the signals to Bragg cells in a Mach-Zehnder interferometer to produce a first beam defining a signal flow one said element of the array and a second beam defining a signal which is a combination of the signal from said first and second elements of the array, passing each beam through opto-electrical transducer means and applying the two formatted signals to a computer.

To enable the invention to be fully appreciated it will now be described with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS:

Figure 1 shows the geometry for direction finding analysis of a two element antenna array. Elements E1 and E2 are spaced a distance d apart and give rise to voltage signals V1 and V2 respectively. V1 and V2 respectively. The antenna array, its normal n_a and the radio wave propaqation vector k_i lie in the azinuthal plane. The angle of arrival is φ and this determines the difference in the time of arrival τ_fof a wavefront at E1 and E2.

Figure 2 is a ray diagram of the Mach-Zehnder optical signal processing system in plan view. S, laser source; BE, beam expander; MO, primary beam splitter;

P1, path 1; P2, path 2; BC1 and BC2,

Bragg cells; M1 and M2, beam splitters;

M3, beam combiner; L1 and L2, Fourier transforming lenses; D1 and D2, linear photodetector arrays; BD2 and BD3, beam dumps.

Figure 3 is a schematic diagram of the electronics associated with the Mach-Zehnder acoustooptic signal processor.

Figure 4 is a series of plots of the output of the photodetector array D2. The pixcells have been assigned a frequency between 30 and 60 MHz. The 35 MHz signals applied to the MZAO signal processor have Oº phase difference for each output but the 45 MHz signals have phase differences varying between -90° and 90°.

Figure 5 is a plot of the phase difference estimate φ produced by the MZAO signal processor versus the phase difference φ the Hewlett Packard HP3335A's were programmed to produce. Figure 6 is a component diagram of a possible MZAO signal processor for field use. A comparison with figure 2 will indicate ray paths. S, laser diode; BE, beam expander; M0, primary beam splitter;

T1 and T2, piezoelectric transducers; BCl and BC2, Bragg cells; M1 and M2 beam splitters; M3, beam combiner; L1 and L2, Fourier transforming lenses; D1 and D2, linear photodetector arrays.

THE MZAO SIGNAL PROCESSOR:

An ESM receiver based on the MZAO signal processor has an antenna array consisting of two elements E1 and E2 spaced a distance d apart, see figure 1. The array, its normal n and the propagation vector k_i of the radio wave lie in the azimuthal plane. The direction of arrival angle θ is defined to be the angle between n and -k_i taken in a clockwise direction. The time difference between the arrival of a wavefront at elements E1 and E2 is

1

It is this time difference that carries the DOA informa tion to the MZAO signal processor.

The signals received by the signal process have a voltage which is of the general form

V(t-τ)=A(t-τ)cos(ω(t-τ))

where τ = τ_f for V1 _3a

= 0 for V2 3b and ω=2π

f

Usually τ-1/f hence if the modulation term A(t) varies only slightly over many cycles of the carrier cos(ωt) then

equation 2 can be approximated by

V(t-τ)=A(t)cos(ω(t-τ)) 5

A signal of this form is applied to each of the Bragg cells BC1 and BC2 in the MZAO signal processor see figure 2. This signal causes a piezoelectric transducer to launch an acoustic wave across the Bragg cell which is given within a constant as⁷

S(x-v(t-τ))=A(x-vt)exp(-

(f)[x ])cos(ω( -t+τ)) 6

The acoustic wave is attenuated as It propagates with a velocity v across the optical aperture which is of length D in the x direction, γ is the acoustic loss coefficient in Np/m and is a function of frequency.

The optical effects of the Bragg cell can be described via a transmission function. In the Bragg region of AO one of the first diffraction orders can contain most of the deflected energy . If the upper first diffraction order is chosen to dominate and

[K_BA(x-vt)]_max«1 7

where K_B is a constant for a given Bragg cell. Then the optical transmission function that projects out only the light in the upper first diffraction order is

T(x,t-τ)=rect )rect() K_BA(x-vt)exρ(-γ(f)[x+ ])

exp(i[ω(x-t+τ)] 8

where L is the width of the acoustic beam in the y direction and x=(x,y) . Equation 7 shows that the term containing the time delay τ can be factored out of the optical transmission function le

T(x,t-τ )=T(x, t )exp(iωτ ) 9 Figure 2 is a ray diagram of the optics of the MZAO signal processor. In relation to this diagram, the output of the beam expander is a monochromatic light wave polarised in the x direction. The x direction is defined to be normal to the page and directed out of it. This light is split by the mirror MO so that half its amplitude continues on to the path P1 and the remainder is reflected into path P2. The light deflected into the first diffraction order of the Bragg cell BC1 is then

E^T(x,t)=T(x,t-τ)2E(x)expivt 10

where v is the optical frequency of the incident optical field and 2E(x) is its amplitude distribution. Mirror M2 directs equal portions of the light into the paths P1 & P1' . The lens L1 lies an optical path length 0 from the Bragg cell BC1 along the path P1' and it has a focal length F. The linear photodetector array D1 is situated in the back focal plane x' of this lens. Hence, the light distribution at this detector is¹⁰

E(x',t)=exp(1[δ(x')+γt+ωτ]) {v(x-vt)] 11

where

12

with

W(x)=E(x)rect rect exp(-γ(f) 13

is the weighted Fourier transform of the undelayed input signal

V(x-vt)=A(x-vt)exp 14

and

15

The photodetector measures light intensity and also integrates across the y dimension of the field. Hence the output of the photodetector array is the weighted power spectrum

16

where

now acts only on the x coordinate and the y integration is incorporated in the constant a along with the time average required to obtain intensity.

Using the small deflection angle approximation the x' coordinate is proportional to the radio frequency⁹

17

Hence the weighted power spectrum of the radio signals can be written in terms of the frequency ie

18

The output of the interferometer passes through lens L2 and is then detected by D2. Light from BC1 travels along path F1 and is superimposed at M3 with the light from BC2 which travels along path P2. The transmission function for BC2 is an undelayed version of that for BC1. Thus, the interferometric power spectrum output by the photodetector array D2 is

P^, _I(f)=P(f)_(1+cos(δ₁₂(f)+ωτ_f)) 19

where

δ₁₂ (f)=[(O₂-O₁)

O₁ is the optical path from BC1 to L2 and O₂ is the optical path from BC2 to L2

The instrumental phase δ₁₂(f) can be set to zero by choosing

O₁=O₂.

However, consider a generalised irrtruncntal phase Δ(f) which includes phase differences introduced in all sections of the ESM receiver up to the output of the photodetector array D2 eg. the radio section and the piezoelectric transducers. This phase function can be expanded in a power series in f

Δ(f)=Δ_O +Δ₁f+Δ₂f ²....... 20

The zero order term of this series can be set to a chosen value (eg

) by an appropriate adjustment of the difference between the optical path length from mirror M0 to BC1 and that from M0 to BC2. The first order term can be nulled by adjusting the difference in the electrical path length of the cable carrying signal V1 from the antenna to the MZAO signal processor to that of the cable carrying V2. The second order term can be adjusted by choosing O₁ -O₂ appropriately as considered before. Hence it is possible to correct any instrumental phase error to second order in f. The higher terms in this phase error series were found to contribute only small phase errors (~1°) in the instrument tested in the lab.

These can be compensated digitally by a postprocessor.

It is desirable for numerical reasons to replace the cosine function in equation 19 by a sine function, by choosing

21

Using equations 1 and 22 the direction of arrival of a signal of frequency f is given by

22

in the linear region of the sine curve this equation becomes

23

Hence the output of the MZAO signal processor P(f ) and P_I (f ) can be converted by a digi tal postprocessor into the weighted power spectrum P(f ) of the radio signals and the direction of arrival θ( f ) of its components . AN EXPERIMENTAL MZAO SIGNAL PROCESSOR:

A MZAO signal processor was constructed and arranged so that when the signal processor was operating, the optics were enclosed in a large plastic foam box in order to stabilise the air around the interferometer and hence the interference fringes.

A model 105-1 Spectra Physics laser was used as a light source. This laser provided a vertically polarised light beam with a wavelength λ = 632.8 nm. The beam was spatially filtered and expanded 43 times.

All mirrors had a one to one, reflection to transmission ratio. These mirrors were fabricated using dielectric films on one face of 1 cm thick optical flats. The opposite face of the optical flats was antireflection coated.

The alignment of the mirrors M0-M3, see figure

2, consisted of two main stages. A Fabry-Perot interferometer was formed from a pair of mirrors Ml and M2 by placing a laser with its beam incident normal to one of these mirrors. The resulting Fabry-Perot fringes were used to accurately align M1 and M2 parallel to each other. In the second stage, M0 was adjusted to superimpose the beams, along paths P1 and P2, at mirror M3. The mirror M3 was then adjusted to superimpose the beams at the focal plane of the lens L2. Sometimes the second stage needed to be repeated in order to observe interference fringes at the output of the interferometer. When these interference fringes were observed, they were used to complete the alignment of mirror M3. Iεomet model OPT-1 Bragg cells were used. A pair of these was selected out of a batch of four. This choice was made on the grounds of matching optical path lengths. The Mach-Zehnder interferometer showed that the maximum path difference between these two cells was <λ/8.

The Fourier transforming lenses were constructed by following the design of W.H. Steel 1973 "A simple lens for optical Fourier transforms" Jnl of Opt 2 36f. Tests with a Twyman-Green interferometer verified that these lenses met the specifications given by Steel.

The optical to electrical transducers were Reticon HL1024H linear photodetector arrays. These devices have a dynamic range of 26 dB.

Analogue signals from these detectors were converted into 12 bit digital form in 2.8 ws by Analogue Devices ADC-EH12B3 analogue to digital converters, see figure 3. The throughput rate to the FIFO controlled RAM was 2024 words every 3 ms. This RAM was read at a much slower rate by a Hewlett Packard HP-87 microcomputer.

A high speed digital postprocessor is used to detect elements of the photodetector arrays that have changed in sequential scans and only data from those elements is processed.

The test radio signals were produced by three frequency synthesizers as shown in figure 3. The two Hewlett Packard HP3335A synthesizers were phase locked to a common crystal and the phase difference between their outputs could be adjusted by computer control. The output of the Rockland 5310A synthesizer was power split and each of the resulting components was summed with the output of a Hewlett Packard synthesizer. The electrical lengths of the cables connecting the power splitter to the summers were matched so that the signals from the Rockland provided a 0° phase difference reference. The phase differences were checked at the Bragg cells by using a Hewlett Packard HP8405A vector voltmeter.

In one series of measurements performed with the MZAO signal processor, the Rockland provided a 35 MHz signal and the Hewlett Packard synthesizers provided 45 MHz signals. Figure 5 is a series of plots of the output of the photodector array D2. In this diagram, each pixcell has been assigned a radio frequency value. These values were obtained by observing the position of the centroid of the diffracted beam on the photodetector array for radio frequencies at 5 MHz intervals between 30 and 60 MHz and linearly interpolating for the points in between. The frequency resolution of the signal processor as determined from these interferometric spectra is 60 kHz, at full width half maximum of the peaks, over a bandwidth of 30 MHz.

The change in intensity of the 45 MHz peak with changes in the phase difference clearly illustrates behaviour consistent with equation 22 where

φ=ωτ_f ²⁴

Though there is some residual intensity when φ = 90º, this can be calibrated out using the postprocessor. Figure 6 is a plot of the radio phase difference φ estimated by the MZAO signal processor versus the phase difference φ produced by the HP3335A synthesizers. The data in this diagram show that the error φ in the estimate is less than 1º over the phase difference range -45° to 45°. Similar graphs were obtained at 10 MHz intervals over the range 30 to 60 MHz. In order to produce these estimates, gain and offset corrections were made to the raw data from the photodetector arrays.

In order to provide a compact rugged MZAO signal processor for field use, the interferometer could be constructed out of an acousto-optic material as demonstrated in figure 6. In this type of construction, the face at the join of the two trapezoid blocks of A0 material .is mirrored and provides both the primary beam splitter M0 and the beam combiner M3. The acoustic beams are launched through the A0 material by the two piezoelectric transducers T1 and T2.

It is desirable to further integrate components such as the laser diode S and beam expander BE as well as the interferometer and lenses, however this development is beyond the scope of the present paper.

CONCLUSIONS:

The experimental data show that the MachZehnder acousto-optic signal processor described, can provide a wide band (30 MHz) power spectrum with 60 kHz resolution and simultaneously provide the direction of arrival of the power spectrum components with 1 accuracy over a 90 viewing angle. Phased array signal processors can sort signals of the same frequency on a direction of arrival basis. The MZAO signal processor sacrifices this capability for a higher throughput rate. This high throughput rate, coupled with the 100% probability of intercept and direction of arrival capabilities of the MZAO signal processor should enable it to be very effective in ESM roles against frequency hopping communications systems.

Claims

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS:

1. A method of acousto-optic signal processing for electronic support measures comprising

(a) receiving a signal on an antenna consisting of a two element spaced array (E1-E2)

(b) applying the signals to Bragg cells (BC1-BC2) in a Mach-Zehnder interferometer to produce a first optical beam (P1) defining a signal flow from one said element of the array (E1), and a second beam (P2) which is a combination of the signal from said first element of the array (E1) and the signal from the second element of the array (E2), and

(c) passing each beam (Pl-P2) to an opto-electrical transducer (L1-L2) and applying the two electrical outputs to a computer through linear photo detectors

(D1-D2).

2. A method of acousto-optic signal processing for electronic support measures comprising

(a) receiving an electro-magnetic signal on a two element spaced array (E1-E2),

(b) amplifying the said signal from each element

(E1-E2) and applying the converted signals to a pair of Bragg cells (BC1-BC2) mounted in the two paths of a Mach-Zehnder interferometer one signal to each, (c) difracting a pair of light beams from a splutter

(M0) through the said Bragg cells (BC1-BC2) to convert the format to a pair of light beams (P1-P2)

(d) passing a first said difracted beam (PI) from the light splitter (M1) through a Fourier transforming lense (L1) to a linear photodetector array (D1)

(e) passing the said first difracted beam (P1) and the said second said difracted beam (P2) to a light combiner (M3) and then through a Fourier transforming lense (L2) to a second linear photodetector array (D2), and

(f) passing the signals from the said detectors (D1-D2) to a computer.

3. An acousto-optic signal processor for electronic support measures comprising

(a) a two element array having its antenna (E1-E2) spaced apart to receive a signal,

(b) means for applying the two signals to Bragg cells (BC1-BC2) in a Mach-Zehnder interferometer to produce a first beam (P1) defining a signal from the first element of the array (E1) and a second beam (P2) defining a signal from said first element (E1) of the array and the said second element (E2) of the array, and

(c) opto-electrical transducer means (L1-L2) and linear photodetectors (D1-D2) arranged to apply two resulting signals (P1-P2) to a computer.

4. An acousto-optic signal processor according to Claim 3 comprising a laser (5) and beam expander (BE) arranged to direct a light beam to a beam splitter (M0), a Bragg cell (BC1) in the path of a first beam (P1) from the primary beam splitter (M0) and a Bragg cell (BC2) in the path of a second beam (P2) from the said primary beam splitter (M0), a pair of secondary beam splitters (M1) and (M2) to receive the output from the said Bragg cells (BC1-BC2), a Fourier transformer (L1) and linear photodetector (D1) to receive the beam (P1) from the secondary beam splitter (Ml), a Fourier transformer (L2) and linear photodetector (D2) to receive the beam (P2) from the secondary beam splitter (M2) combined with the beam (P1) directed from the secondary beam splitter (Ml) to a beam conbiner (M3) in the path of the beam (P2) from the secondary beam splitter (M2).

5. An acousto-optic signal processor according to Claim 3 or 4 wherein the interferometer is constructed of acousto-optic material comprising two trapezoidal blocks having a joining face, mirrored to provide both the primary beam splitter (M0) and the beam combiner (M3), two piezoelectric transducers (T1-T2) being incorporated to launch the acoustic beams through the acousto-optic block.

6. A method of acousto-optic signal processing substantially as described and illustrated with reference to the accompanying drawings.

7. An acousto-optic processor constructed and operating substantially as described and illustrated in the accompanying drawings.