GB1602094A

GB1602094A - Adaptive cancellation arrangements

Info

Publication number: GB1602094A
Application number: GB2866377A
Authority: GB
Original assignee: Marconi Co Ltd
Current assignee: BAE Systems Electronics Ltd
Priority date: 1977-07-08
Filing date: 1977-07-08
Publication date: 1981-11-04
Also published as: DE2750151C2; DE2750151B1

Description

(54) ADAPTIVE CANCELLATION ARRANGEMENTS (71) We, THE MARCONI COM PANY LIMITED, a British Company, of Marconi House, New Street, Chelmsford, Essex, CMl lPL, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement:-- This invention relates to adaptive cancellation arrangements.

Adaptive cancellation arrangements may be used in signal receiving systems where, in addition to the reception of a wanted signal, unwanted continuous inteference signals may be received from one or more sources.

In, for example, a radar system, the wanted signals may be return echoes from targets.

whilst the unwanted interference signals may be noise signals received from one or more noise jammers located at various bearings relative to a radar receiving aerial system.

Adaptive cancellation operates, in its broadest aspect, by arranging for signal reception in a plurality of signal channels each including, in the presence of interference, both wanted and unwanted signals, the relationship between the wanted and unwanted signals ideally being different in each channel.

The signals in the different channels are combined in a summing circuit, after imparting to each channel an appropriate weighting value, to provide a final single output signal in which interference signals are minimised or under ideal conditions, nullified.

The weighting values may be represented by complex numbers which multiply the complex numbers representing the amplitudes and phases of the received signal waves in each channel.

The plurality of signal channels required m an adaptive concellation arrangement can be derived from two known aerial systems.

(a) a linear or planar array of similar antenna elements regularly spaced (b) a main aerial receiving dominantly with in a narrow beam and one or more auxiliary aerials which are usually omnidirectional.

The system of case (b) is known as coherent sidelobe cancellation.

The calculation of weighting values for cases (a) and (b) above is carried out in a decorrelation stage or in the case of more than one weighting value, an arrangement of stages, so that the unwanted signals received at the different aerials are decorrelated and do not combine constructively.

Two types of decorrelation stage or arrangements are known, the closed loop type and the open loop type. The closed loop type is essentially a feedback servomechanism and known arrangements will now be described with referto the drawings accompanying the provisional specification in which, Figure 1 is a simplified illustration of an arrangement of closed loop decorrelation stages - for a linear array antenna, Figure 2 shows the corresponding arrangement for the sidelobe cancellation case.

Adaptive cancellation arrangements using closed loop decorrelation stages are described in "Adaptive Arrays - An Introduction" by W. F. Gabriel, Proc. IEEE. Vol. 64, No. 2, February 1976.

Referring to Figure 1 a linear array includes, in this example, four aerial elements 1 each providing an individual signal channel 2. The signal channels 2 may be at R.F. frequency, but for the present purposes it will be assumed that each channel 2 carries both in-phase and quadrature-phase signals at I.F. frequency provided by frequency changers (not shown) from R.F. to I.F. The frequency changers are fed from a local oscillator (again not shown) which is common to all of the channels 2, and which provides in-phase and quadrature-phase oscillations.

Each channel 2 is connected as one of two double inputs to a respective correlation mixer 3. The second input for each correlation mixer 3 is derived from an output channel 7, which, as will be described later, carries just one signal at I.F. frequency. The output from each correlation mixer 3 is a double output consisting of a pair of low frequency signals which rep resent in analogue form the complex number product of the complex number which represents the pair of signals in the respective channel 2 and the complex number which represents the amplitude and phase of the signal in output channel 7. In practice the signals in the channels 2 are limited at the inputs to the correlation mixers 3, but means for achieving this are not shown.

Each channel 2 is also connected as one of two double inputs to a respective control mixer 4, but now without limiting. The second double input to each control mixer 4 is a pair of low frequency signals from a respective comparator 9.

Each comparator 9 is connected to receive on one of its two double inputs the pair of low frequency signals from the output of its respective correlation mixer 3 via a respective filter amplifier 8. The latter are each double units which act so as to remove unwanted high frequency components from the outputs of their respective correlation mixers 3. The second double input for each comparator 9 is an input signal at D.C. or very low frequency connected to respective input terminals 10. The double output signal of each comparator 9 applied to the second double input of its respective control mixer 4 represents in analogue form the difference of the complex numbers representing the double inputs to the comparator. The purpose of the comparators 9 will be described later.

The output 5 from each control mixer 4 is a single signal at I.F. frequency, such that the complex number representing the amplitude and phase of this signal is the complex number product of the complex numbers representing the pair of signals at I.F. frequency in the respective channel 2 and the pair of signals at I.F. frequency which are the output of the respective comparator 9.

Each output 5 is connected to a respective input of a summer 6 which combines the s; nals present at the outputs 5 and is comma to each decorrelation stage in the respective channels.

1fhe summer 6 provides the output 7 already referred to which, as well as providing the feedback signal which is connected as a second input to each correlation mixer 3, also provides the final output of the arrangement.

The whole arrangement effectively consists of four decorrelation stages operating in parallel, one for each channel, each stage comprising a correlation mixer 3, a filter amplifier 8, comparator 9 a control mixer 4 and a summing circuit 6, the summer 6 being common to each stage.

In operation, the output 7, in which it is desired to minimise interference signals, is multiplied in the appropriate correlation mixer 3 with the signal in the respective channel 2.

As the interference is reduced in the output signals the product formed by the correlation mixer 3 will tend to zero. The output value of the mixer 3 filtered and amplified by the filter amplifier 8, forms the weighting value which is used to multiply the signals in the channel 2 in the control mixer 4.

The comparator 9 is provided to enable beam steering to be achieved for the array of aerials 1 and beam steering reference signals are fed to the terminals 10. These beam steering signals are modified by the weighting signals provided by the filter amplifiers 8 to provide resulting weighting values. The resulting aerial aiming point will, thus, not be quite as determined by the steering reference signals.

The corresponding arrangement for the sidelobe cancellation arrangement is shown in Figure 2. Referring to Figure 2, where like parts to those in Figure 1 have like reference numerals, a main directional receiving aerial 11 provides a main signal channel 12 which is fed directly to one input of 2 summer 6.

Three further channels 22 are provided as auxiliary signal channels and are fed from respective auxiliary omnidirectional aerials 21.

Each auxiliary channel forms part of a decorrelation stage formed as before from a respective correlation mixer 3, filter amplifier 8, control mixer 4, summer 6 and the direct channel 12. The beam steering comparators are, of course, omitted.

For purposes of stability, between each filter amplifier 8 and the corresponding control mixer 4 is an inverter 13. The inverters 13 provide the inverting effect which, in Figure 1, is provided by the comparators 9.

The operation is identical to that of Figure 1 with the exception that the weighted auxiliary signals are combined in the summer 7 with the signal in the main channel 12.

As can be seen, in the case of Figure 1, each aerial and in Figure 2 each auxiliary aerial, provides a signal channel which includes a feedback loop forming part of a decorrelation stage. The number of loops is usually chosen in dependence upon the number of interference sources for which it is required to compensate, one loop being required for each source. In the case of Figure 2, therefore, three interference sources can be dealt with.

A problem with closed loop arrangements is obtaining a satisfactorily short response time for achieving an adequate level of interference cancellation.

A simplified mathematical model of adaptive arrangements shows that the complex weightings are changed from their initial values at the commencement of operation, in the form of a superposition of decaying exponential responses. The time constants of these responses depend upon the eigenvalues of the covariance matrix which defines a set of equations expressing the correlations between the antenna elements, and upon the gain the loops.

Satisfactory reduction of interference demands, in the sidelobe cancellation case, at least as many auxiliary aerials and loops as there are sources of interference and in the array case, one extra element and loop.

It is found that loop conditions (gain and power level), which give an adequately fast response for a single loop often give a very slow response when there are two or more loops. This arises because of interaction between the loops, as determined by the eigenvalues of the covariance matrix. A typical situation which causes slow response is when two jammers are almost in the same direction, but not excessively so.

Typically, some time constants of the res ponses are shorter than, or of the same order of magnitude as, the time constant for a single loop, but often at least one time constant is very long. This can result in a rapid approach to partial cancellation of the interference, folowed by a slow approach to the final value of cancellation.

One way round this problem is to use an open loop system. In this case the complex weightings are found by solving the system of complex linear simultaneous algebraic equations with coefficients specified by the covariance matrix.

This process demands a powerful computer, but the calculation time is almost independent of interference conditions. There are several disadvantages in that high accuracy is required for converting the analogue signals or correlations to digital form and for converting the calculated weightings back to analogue form.

Further, the self-compensating effect of a servo-mechanism has been lost I11 conditioning of the equations must also be considered and this results in small changes in the correlation coefficients in the equations producing large changes in the weightings.

This invention seeks to provide a closed loop adaptive cancellation arrangement in which the above mentioned disadvantages are mitigated.

According to this invention, a closed loop adaptive cancellation arrangement for minimising unwanted signals comprises a plurality of first signal channels each forming part of the feedback loop of a respective one of a corresponding plurality of closed loop decor- relation stages which operate to decorrelate unwanted signals, and wherein there is provided at least one further signal channel forming part of the feedback loop of a respective further closed loop decorrelation stage, the input signals for each further channel comprising a combination of input signals of the said first signal channels.

Where the cancellation arrangement is of the sidelobe cancellation type, an unweighted main signal channel is provided as part of at least one decorrelation stage.

The said plurality of closed loop decor relation stages may be arranged in parallel with one another but, preferably, at least two cas- caded groups each of a plurality of parallel connected decorrelation stages is provided.

Advantageously, the said plurality of decorrelation stages are in parallel with one another as a first of said at least two cascaded groups.

The means for combining input signals may include means for forming sum and difference signals from pairs of input signals of said plurality, but preferably, includes means for forming both sum and difference signals.

A phase shift may be introduced into one or more selected chanels prior to forming sum or difference signals as the case may be.

This invention will now be described fur ther, by way of example, with reference to Figures 3 to 6 of the drawings accompanying the provisional specification in which, Figure 3 shows a two auxiliary aerial sidelobe adaptive cancellation arrangement in accordance with this invention, Figure 4 shows a modification of the arrangement of Figure 3, Figure 5 shows a cascaded version of the arrangement of Figure 3 and Figure 6 illustrates the application of the invention to a three auxiliary aerial arrangement.

Like parts to those of Figure 2 bear like reference numerals.

Referring to Figure 3, a main receiving aerial 11 and two auxiliary aerials 21 respectively feed a main signal channel 12 and auxiliary channels 22. The main channel 12 and the auxiliary channels 22 form two parallel connected closed loop decorrelation stages in exactly similar manner to the three channel arrangement of Figure 2, the summer 6 again being common to each stage.

In accordance with the invention, each of the two channels 22, which form part of the actual feedback loop of a decorrelation stage, is used to derive a further pair of channels 32. Each channel 22 is connected to respective inputs of an adder 14 and to the respective inputs of a subtractor 15. The outputs of the adder 14 and the subtractor 15- provide the two further channels 32.

Each channel 32 forms part of the feedback loop of a further closed loop decorrelation stage and hence includes a respective correlation mixer 30, filter amplifier 80 and control mixer 40. The control mixers 30 receive, as input signals, the signals in the derived channels 32 and the control mixers 40 provide respective output signals to the summer 6. The output signal 7 from the summer 6 provides a feedback signal input to each correlation mixer 30.

The additional channels 32, thus, form part of the feedback loops of respective decor relation stages connected in parallel with the stages of which the auxiliary channels 22 form part, the summer 6 being common to all stages.

By generating the additional channels 32, from arbitrary combinations of the input signals to each feedback loop, a situation arises where there are more feedback loop decorrelation stages than auxiliary aerials. This provides no improvement in the steady state condition, since the additional derived loops are redundant, but it has been found that the provision of the extra loops generally results in a fast approach to the steady state condition.

A modification to the Figure 3 arrangement is shown in Figure 4 which shows an arrangement identical to that of Figure 3 but with signals obtained for combination from one of the channels 22 being passed through a 90" phase shifter before being fed to one input of the adder 14 and the subtractor 15 In each of the arrangements of Figures 3 and 4, the additional derived loops are included in parallel with the original loops. With such an arrangement of parallel loops the algebraic equations for the steady state condition are singular, i.e. there are infinitely many solutions.

A further improvement can be provided by providing loops in cascade and this is shown in Figure 5.

Referring to Figure 5, the loops including the channels 22 are arranged with correlation mixers 3, filter amplifiers 8, control mixers 4, summer 6 and inverter 13 in exactly corresponding manner to their arrangement in Figures 3 and 4 to form a first group of decorrelation stages. The output 7 from the summer 6 provides the feedback signal for each feedback loop.

The channels 32 are derived as before from the channels 22 by combination in the adder 14 and the subtractor 15. As before, the channels 32 provide two additional loops each including correlation mixer 30, filter amplifier 80 and control mixer 40. In this case, however, the control mixers 40 do not feed the summer 6 but an additional summer 60 to provide a second group of decorrelation stages.

This provides an output signal 70 which provides via an inverter 23 a feedback signal for each of the additional loops.

The whole arrangement of additional loops is connected as a second group in cascade with the loops including the aerials 21 by the output 7 from the summer 6 providing a main signal input for decorrelation stages which include the additional loops, in similar manner to the way in which the channel 12 provides a main signal for the decorrelation stages which are fed from the aerials 11 and 21.

By providing groups of decorrelation stages in cascade as in Figure 5, it is possible to arrange that each group has a complete set of loops with no redundancy. In addition, by cascading, the gain of each closed loop may be made smaller.

Referring now to Figure 6, there is shown a possible derivation of further channels in the case of three original loops. Three adders 14 and three subtractors 15 are provided and each adder adds and each subtractor subtracts the signals in each of the three possible pairings of the channels 22 to provide six additional channels 32.

The channels 32 may all be arranged in parallel connected loops, but it is preferred to provide cascaded groups of decorrelation stages. This could be done by having three groups of stages, the first including the loops containing the original input channels 22, the second including the feedback loops of the summation channels and the third group the loops of the difference channels.

Although six additional loops have been generated, it is possible to use only some of these, for example, only those of channels 32 provided by the adders 14 or the subtractors 15.

Adaptive cancellation using cascaded stages can also be applied to an adaptive linear array of the kind shown in Figure 1. The arrangement shown in Figure 5, for example, would be modified by substituting for the aerials 11 and 21 and the circuit features between the aerials 11, 21 and the summer 6 (which as shown in Figure 5), are similar to the corresponding aerials and circuit features in Figure 2), aerials and circuit features similar to the corresponding aerials and circuit features shown in Figure 1.

WHAT WE CLAIM IS: 1. A closed loop adaptive cancellation arrangement for minimising unwanted signals comprising a plurality of first signal channels each forming part of the feedback loop of a respective one of a corresponding plurality of closed loop decorrelation stages which operate to decorrelate unwanted signals, and wherein there is provided at least one further signal channel forming part of the feedback loop of a respective further closed loop decorrelation stage, the input signals for each further channel comprising a combination of input signals of the said first signal channels.

2. An arrangement as claimed in claim 1 and wherein the cancellation arrangement is of the sidelobe cancellation type, an unweighted main signal channel being provided as part of at least one decorrelation stage.

3. An arrangement as claimed in claim 1 or 2 and wherein said plurality of closed loop decorrelation stages are arranged in parallel with one another.

4. An arrangement as claimed in claim 1 or 2 and wherein asid plurality of closed loop decorrelation stages are arranged such that at least two cascaded groups each of a plurality of parallel connected decorrelation stages are provided.

5. An arrangement as claimed in claim 4 and wherein the said plurality of decorrelation stages are irl parallel with one another as a first of said at least two cascaded groups.

6. An arrangement as claimed in any of

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **. signals to each feedback loop, a situation arises where there are more feedback loop decorrelation stages than auxiliary aerials. This provides no improvement in the steady state condition, since the additional derived loops are redundant, but it has been found that the provision of the extra loops generally results in a fast approach to the steady state condition. A modification to the Figure 3 arrangement is shown in Figure 4 which shows an arrangement identical to that of Figure 3 but with signals obtained for combination from one of the channels 22 being passed through a 90" phase shifter before being fed to one input of the adder 14 and the subtractor 15 In each of the arrangements of Figures 3 and 4, the additional derived loops are included in parallel with the original loops. With such an arrangement of parallel loops the algebraic equations for the steady state condition are singular, i.e. there are infinitely many solutions. A further improvement can be provided by providing loops in cascade and this is shown in Figure 5. Referring to Figure 5, the loops including the channels 22 are arranged with correlation mixers 3, filter amplifiers 8, control mixers 4, summer 6 and inverter 13 in exactly corresponding manner to their arrangement in Figures 3 and 4 to form a first group of decorrelation stages. The output 7 from the summer 6 provides the feedback signal for each feedback loop. The channels 32 are derived as before from the channels 22 by combination in the adder 14 and the subtractor 15. As before, the channels 32 provide two additional loops each including correlation mixer 30, filter amplifier 80 and control mixer 40. In this case, however, the control mixers 40 do not feed the summer 6 but an additional summer 60 to provide a second group of decorrelation stages. This provides an output signal 70 which provides via an inverter 23 a feedback signal for each of the additional loops. The whole arrangement of additional loops is connected as a second group in cascade with the loops including the aerials 21 by the output 7 from the summer 6 providing a main signal input for decorrelation stages which include the additional loops, in similar manner to the way in which the channel 12 provides a main signal for the decorrelation stages which are fed from the aerials 11 and 21. By providing groups of decorrelation stages in cascade as in Figure 5, it is possible to arrange that each group has a complete set of loops with no redundancy. In addition, by cascading, the gain of each closed loop may be made smaller. Referring now to Figure 6, there is shown a possible derivation of further channels in the case of three original loops. Three adders 14 and three subtractors 15 are provided and each adder adds and each subtractor subtracts the signals in each of the three possible pairings of the channels 22 to provide six additional channels 32. The channels 32 may all be arranged in parallel connected loops, but it is preferred to provide cascaded groups of decorrelation stages. This could be done by having three groups of stages, the first including the loops containing the original input channels 22, the second including the feedback loops of the summation channels and the third group the loops of the difference channels. Although six additional loops have been generated, it is possible to use only some of these, for example, only those of channels 32 provided by the adders 14 or the subtractors 15. Adaptive cancellation using cascaded stages can also be applied to an adaptive linear array of the kind shown in Figure 1. The arrangement shown in Figure 5, for example, would be modified by substituting for the aerials 11 and 21 and the circuit features between the aerials 11, 21 and the summer 6 (which as shown in Figure 5), are similar to the corresponding aerials and circuit features in Figure 2), aerials and circuit features similar to the corresponding aerials and circuit features shown in Figure 1. WHAT WE CLAIM IS:

1. A closed loop adaptive cancellation arrangement for minimising unwanted signals comprising a plurality of first signal channels each forming part of the feedback loop of a respective one of a corresponding plurality of closed loop decorrelation stages which operate to decorrelate unwanted signals, and wherein there is provided at least one further signal channel forming part of the feedback loop of a respective further closed loop decorrelation stage, the input signals for each further channel comprising a combination of input signals of the said first signal channels.

6. An arrangement as claimed in any of

claims 1 to 5 and wherein the input signal for each further channel is a sum or difference signal obtained from pairs of input signals of said first signal channels.

7. An arrangement as claimed in claim 6 and wherein means are provided whereby a phase shift is introduced into one or more selected channels prior to forming sum or difference signals as the case may be.

8. A two auxiliary aerial sidelobe adaptive cancellation arrangement substantially as herein described with reference to Figure 3 of the drawings accompanying the provisional specification.

9. A two auxiliary aerial sidelobe adaptive cancellation arrangement substantially as herein described with reference to Figure 4 of the drawings accompanying the provisional specification.

10. A two auxiliary aerial sidelobe adaptive cancellation arrangement substantially as herein described with reference to Figure 5 of the drawings accompanying the provisional specification.

11. A three auxiliary aerial sidelobe adaptive cancellation arrangement substantially as herein described with reference to Figure 6 of the drawings accompanying the provisional specification.