Multiplexed Noise Coded Switching System

This application is related to another application filed in the name of the present inventor, and comprises U.S. Ser. No. 407,409 (CERCOM D-2058), entitled, "Orthogonal Code Division Multiple Access Communication System", filed on Aug. 12, 1982. This application is referenced inasmuch as the same type of noise codes are employed in both inventions.

This invention relates generally to signal switching systems and more particularly to a switching system which performs line to line or group to group switching for signals modulated by multiplexed noise codes.

Communications systems utilizing noise coded signals are well known and are particularly desirable because they exhibit immunity against self interference and jamming. In particular, one class of noise codes are known wherein pairs of coded signals termed "code mates" have autocorrelation functions so that when they are detected and the resultant detected outputs are linearly added, there is provided an impulse autocorrelation function meaning that there is an impulse output at a given time and a zero output at all other times. Such codes and systems utilizing such codes are typically shown and described in U.S. Pat. No. 3,461,451, entitled, "Code Generator To Produce Permutations Of Code Mates", which issued to F. S. Gutleber, the present inventor, on Aug. 12, 1969; U.S. Pat. No. 3,519,746, entitled, "Means And Method To Obtain An Impulse Autocorrelation Function", which issued to F. S. Gutleber on July 7, 1970; and U.S. Pat. No. 3,634,765, entitled, "System To Provide An Impulse Autocorrelation Function . . . ", which issued to F. S. Gutleber on Jan. 11, 1972. The teachings of these patents are meant to be incorporated herein by reference, since they provide a basis for the subject invention which will be described subsequently.

Prior art approaches to switching systems for performing line to line or group to group switching between multiple users have resorted to the use of frequency filters, time gates or space division switching. Such apparatus as a result has been found to be relatively complex, bulky and vulnerable to noise interference, cross talk, unauthorized interception and jamming.

Accordingly, it is an object of the present invention to provide an improvement in secure communications switching systems.

Another object of the present invention is to provide an improved switching system capable of automatically accommodating digital and/or analog signals.

Still another object of the invention is to provide an improved switching system for secure communications systems that automatically accommodate both digital and analog signals and which operates with no interference between multiple users while completely eliminating the unauthorized interception of messages therebetween.

Accordingly, these and other objects are achieved by means of a switching system utilizing multiplexed noise codes which compress to a lobeless impulse upon detection and which additionally have a characteristic whereby different codes of a selected subset crosscorrelate to zero at a time when they exhibit an autocorrelation impulse function. The system includes a plurality of input lines coupled to respective noise code mate generators and a plurality of output lines respectively coupled to noise code pulse compressors. The code generators and the pulse compressors operate in accordance with a switching algorithm that sets an input line code generator and an output line pulse compressor to the same code mate pair which automatically directs or translates the specific signal applied to an input line to a selected output line that is matched to the input code.

FIG. 1 is a functional block diagram illustrative of the preferred embodiment of the invention; and

FIG. 2 is a functional block diagram further illustrative of the embodiment shown in FIG. 1.

Referring now to the drawings and more particularly to FIG. 1, shown thereat in block diagrammatic form is a multiplexed noise coded switching system in accordance with the subject invention which is comprised of a plurality of input circuits 10.sub.1, 10.sub.2 . . . 10.sub.n and a corresponding number of output circuits 12.sub.1, 12.sub.2 . . . 12.sub.n. The input units commonly couple to a pair of linear adders 14 and 16 used for signal multiplxing and which respectively connect to a pair of multiplexed noise code busses 18 and 20 commonly coupled to the output circuits 12.sub.1, 12.sub.2 . . . 12.sub.n. Interconnection between the adders 14 and 16 and the noise code busses 18 and 20 is further made by a pair of signal lines 22 and 24. Each of the n input circuits 10.sub.1 . . . 10.sub.n include a respective input line 26.sub.1, 26.sub.2 . . . 26.sub.n which is adapted to receive an input signal S.sub.1, S.sub.2 . . . S.sub.n. In a like manner, each of the output circuits 12.sub.1, 12.sub.2 . . . 12.sub.n include a respective output line 28.sub.1, 28.sub.2 . . . 28.sub.n which are adapted to separately output the signals S'.sub.1, S'.sub.2 . . . S'.sub.n.

Selective signal coupling between one or more input circuits 10.sub.1 . . . 10.sub.n and one or more output circuits 12.sub.1 . . . 12.sub.n is achieved by utilizing multiplexed noise codes of a particular type, to be described, to modulate a particular input signal (S) of the signals S.sub.1 . . . S.sub.n which is then coupled to at least one output circuit having a detector or decoder which is matched to the noise code of the particular input circuit. The noise codes are of the well known general type of noise codes shown and described in the aforementioned U.S. Pat. Nos. 3,461,451; 3,519,746; and 3,634,765 and which consist of code mate pairs that possess an impulse autocorrelation function upon detection. The present invention, however, utilizes a specific subclass or set of these noise codes which have code mate pairs which exhibit a crosscorrelation function equal to zero at a time when the code mate pairs compress to an impulse. While spurious lobes may exist whose rms value has an amplitude that is down by at least the square root of the code sequence length, i.e. the time--bandwidth product, the absence of any correlation value in the same compressed time interval for all codes of this latter type enables the system as shown in FIG. 1 to operate orthogonally, i.e. without any interference between any of the input/output lines, while using different noise codes for the input/output circuits 10.sub.1 . . . 10.sub.n and 12.sub.1 . . . 12.sub.n.

Accordingly, each of the input circuits 10.sub.1, 10.sub.2 . . . 10.sub.n respectively include a noise code selector 30.sub.1, 30.sub.2 . . . 30.sub.n which are adapted to selectively control each of the desired noise code outputs of code mate pair generators 32.sub.1, 32.sub.2 and 32.sub.n and 34.sub.1, 34.sub.2 . . . 34.sub.n so that each input circuit is adapted to generate mutually different code mate pairs. As shown, code generators 32.sub.1 and 34.sub.1 are adapted to generate code mate pairs a.sub.1 and b.sub.1 whereas code generators 32.sub.2 and 34.sub.2 are adapted to generate code mate pairs a.sub.2 and b.sub.2, respectively. Each of the input circuits 10.sub.1, 10.sub.2 . . . 10.sub.n additionally include a pair of modulators 36.sub.1 and 38.sub.1, 36.sub.2 and 38.sub.2 . . . 36.sub.n and 38.sub.n. It can be seen, for example, that code generators 32.sub.1 and 34.sub.1 are coupled to modulators 36.sub.1 and 38.sub.1, respectively, and operate such that code mate a.sub.1 modulates the input signal S.sub.1 in modulator 36.sub.1 which is coupled to the linear adder 14 while code mate b.sub.1 modulates the input signal S.sub.1 via modulator 38.sub.1 whose output is coupled to the linear adder 16.

The input signal S.sub.1 modulated by the generated code mate pair a.sub.1 and b.sub.1 will transfer only to a selected output line 28.sub.1, 28.sub.2 . . . 28.sub.n coupled to output circuits 12.sub.1, 12.sub.2 . . . 12.sub.n whose noise code detectors are matched so that they are responsive to codes a.sub.1 and b.sub.1. Accordingly, the output circuits 12.sub.1, 12.sub.2 . . . 12.sub.n include respective code selectors 40.sub.1, 40.sub.2 . . . 40.sub.n which are coupled to decoders in the form of matched filters 42.sub.1, 44.sub.1, 42.sub.2 and 44.sub.2 . . . 42.sub.n and 44.sub.n. Additionally as shown, the various matched filters for code mates a.sub.1, b.sub.1, a.sub.2, b.sub.2 . . . a.sub.n and b.sub.n couple to respective linear adders 46.sub.1, 46.sub.2 and 46.sub.n wherein resultant compressed impulse signal outputs S'.sub.1, S'.sub.2 . . . S'.sub.n appear in a well known manner on lines 28.sub.1, 28.sub.2 . . . 28.sub.n.

Referring now to FIG. 2, the ith input and output circuits 10.sub.i and 12.sub.i are shown respectively coupled to the linear adders 14 and 16 and the busses 18 and 20 which additionally translate all codes a and b of different code mate pairs generated in the system shown in FIG. 1. What is additionally shown in FIG. 2 is a switch controller 48 coupled to the code selectors 30.sub.i and 40.sub.i for matching the codes a.sub.i and b.sub.i generated and detected by code generators 32.sub.i and 34.sub.i and the matched filters 42.sub.i and 44.sub.i. Also a clock 50 is shown coupled to the a.sub.i and b.sub.i code mate pair generators 32.sub.i and 34.sub.i. Its purpose is to start the generated codes a.sub.i and b.sub.i as well as all other codes at the same time. The clock 50 additionally is coupled to a gate generator 52 which is adapted to trigger a gate 54.sub.i at a time the detected code mate pairs a.sub.i and b.sub.i compress to an impulse output signal S.sub.i upon being added linearly by the linear adder 46.sub.i.

The code mate pairs a.sub.i and b.sub.i are of the general type disclosed in the aforementioned U.S. Pat. Nos. 3,461,451; 3,519,746; 3,634,765 but comprise a subclass thereof having an additional characteristic which permits different code mate pairs to be selectively switched concurrently between input and output lines 26.sub.1 . . . 26.sub.n and 28.sub.1 . . . 28.sub.n simultaneously without mutual interference therebetween. In order to fully understand the operation of the subject invention, it becomes necessary to describe the specific type or subclass of code mate pairs utilized.

As noted above, multiplexed noise codes are codes formed with mate pairs that compress to an impulse autocorrelation function upon detection. The autocorrelation function of two codes forming a mate pair must be of equal magnitude and opposite sense for all values of time outside of the main lobe. Expressed mathematically, for two code mate pairs a and b,

where

.phi..sub.a.sbsb.i (.GAMMA.) is the autocorrelation function of code a.sub.i,

.phi..sub.b.sbsb.i (.GAMMA.) is the autocorrelation function of code b.sub.i, and

.GAMMA..sub.0 is the location of the main lobe.

When the two codes a.sub.i and b.sub.i meet this requirement, then the simple linear sum .phi..sub.Ti (.GAMMA.) of their orthogonally multiplexed outputs from matched filters 42.sub.i and 44.sub.i coupled to busses 18 and 20 result in compressing the composite code structure into a single impulse at time .GAMMA.=0.

As an illustration and for clarification, consider the use of the following mate code pair in the switching system as a result of the operation of code selector 30.sub.i of FIG. 2. For code a.sub.i =1000 and code b.sub.i =1011, the code selector 40.sub.i would compress these codes by autocorrelation as follows.

The output .phi..sub.a.sbsb.i (.GAMMA.) of the filter 42.sub.i, matched to code a.sub.i by providing an inversion sequence of 0001 where 0=0.degree. and 1=180.degree. phase reversal, is generated graphically below as,

and the exponent is the amplitude of the 0 or 1 bits.

The output .phi..sub.b.sbsb.i (.GAMMA.) of the filter 44.sub.i matched to code b.sub.i by providing an inversion sequence of 1101, is graphically illustrated as,

The linear sum .phi..sub.t (f) of the two matched filter outputs results in

As noted above, orthogonal subsets of multiplexed noise codes exists whereby code pairs of the same subset have a crosscorrelation value of zero at .GAMMA.=0 which results in no interference (zero crosstalk) occurring between any of the codes when the code mate pairs compress to an impulse. Two specific sets of codes possessing the required crosscorrelation characteristic are shown in tables 1 and 2 where the codes identified in Table 1 form a subclass of 4 bit multiplexed codes and the codes identified in Table 2 form a subclass of 8 bit multiplexed codes.

Demonstrating that there would be no interference occurring between any of the input/output lines 26.sub.1 . . . 26.sub.n and 28.sub.1 . . . 28.sub.n utilizing any of the code mate pairs of Tables 1 and 2 requires establishing that the crosscorrelation between any two or more codes identified in either tables is zero at .GAMMA.=0.

Consider first the codes listed in Table 1. The crosscorrelation function of the codes a.sub.1, a.sub.2 and b.sub.1, b.sub.2 for code mate pairs 1 and 2 at .GAMMA.=0 is readily determined as follows:

For code a,

where:

1.times.0=-

0.times.1=-

1.times.1=+

0.times.0=+

For code b,

And the composite crosscorrelation function at .GAMMA.=0 becomes:

In a like manner, the crosscorrelation function value at .GAMMA.=0 was calculated for all 28 possible combinations. The results are shown in Appendix A appearing at the end of the specification and verifies that the crosscorrelation values between any 2 or more codes of the same subset is zero and accordingly the codes are orthogonal or totally non-interfering at .GAMMA.=0.

As a further clarification, consider the situation where eight input and output lines 26.sub.1 . . . 26.sub.8 and 28.sub.1 . . . 28.sub.8 are being utilized at the same time with each input/output line pair employing and matched to a different code mate pair of the set shown in Table 1. For the input/output circuits 10.sub.1 and 12.sub.1 responsive to code mate pair No. 1, the potential interference of codes 2 through 8 would be as follows:

The interference signal .SIGMA..sub.a present at linear adder 14 of FIG. 1 and coupled to code a bus 18 is graphically illustrated below as,

The interference signal .SIGMA..sub.b present at linear adder 16 and coupled to code b bus 20 would be,

where the exponent indicates the amplitude.

The output .phi..sub.a1 and .phi..sub.b.sbsb.1 of the matched filters 42.sub.1 and 44.sub.1 of output unit 12.sub.1 matched to code pair No. 1 would be as follows:

Code a matched filter 42.sub.1 responds to .SIGMA..sub.a and provides an inversion sequence of 0001 to output .phi..sub.a.sbsb.1 (.GAMMA.) in the following manner.

Code b matched filter 44.sub.1 responds to .SIGMA..sub.b and provides an inversion sequence of 1101 to output .phi..sub.b.sbsb.1 (.GAMMA.) as,

And the composite summed output .phi..sub.T (.GAMMA.) would result in,

The center or fourth bit time slot corresponds to .GAMMA.=0 and is where all the other selected codes utilized in the system would be totally non-interfering. The interfering energy that is present in the lobes of the compressed signal for .GAMMA..noteq.0 is simply gated out or it causes no interference at all. It is important to note that although this illustrated example used the same amplitude for each signal being switched, the actual amplitudes could all be different and theoretically have any value with the same result. Furthermore, there would be no interference problem present in the proposed concept whether digital, analog or a mix of digital and analog signals are employed.

Therefore, the desired signal S.sub.1 (FIG. 1) applied to input line 26.sub.1 and modulated with code mate pair No. 1 would simply linearly sum with the total input and compress to its peak value of S'.sub.1 =8 at .GAMMA.=0 which would then appear on output line 28.sub.1.

When a positive signal S.sub.1 is present and modulated by code mate pair No. 1 of Table 1 the composite signal on the code a bus 18 would be.

while the composite signal at the code b bus (20) would be

The output of the matched filters 42.sub.1 and 44.sub.1 for code No. 1 would then be as shown below. For the code a matched filter 41.sub.1 providing an inversion sequence of 0001, there is provided .phi..sub.a.sbsb.1 (.GAMMA.) and illustrated as,

The code b matched filter 44.sub.1 output would be zero, therefore the composite summed output .phi..sub.T.sbsb.1 (.GAMMA.) would result in

What is provided on output line 28.sub.1 is an output S'.sub.1 equal to the input that is utilizing code No. 1 with an amplitude 8 times larger. If the input signal is digital using a binary structure of +1 and -1, then its negative value would simply become -8 at the output line.

A similar demonstration for the code mate pairs of Table 2 would reveal identical results. Rather than present the crosscorrelation function values for all of the 66 possible combinations, however, a typical sample situation will be illustrated. This will suffice to demonstrate and clarify that the 8 bit code subset set forth can indeed be utilized for the noise coded switch of this invention and result in zero interference at the output lines.

Referring to Table 2, consider, for example, code mate pair Nos. 5 and 7 being used simultaneously with code mate pair No. 4. For the input/output circuits 10.sub.4 and 12.sub.4 that is utilizing code pair No. 4, then the potential interference of code pairs Nos. 5 and 7 would be as follows.

The code a matched filter 42.sub.4 of output unit 12.sub.4 would see a signal .SIGMA..sub.a.sbsb.4 which can be graphically illustrated as,

The code b matched filter 44.sub.4 would see a signal .SIGMA..sub.b.sbsb.4 as shown below as,

The output of the matched filters 42.sub.4 and 44.sub.4 responsive to code pair No. 4 would be as follows.

Code a matched filter 42.sub.4 would output .phi..sub.a.sbsb.4 (.GAMMA.) as,

Code b matched filter 44.sub.4 would output a signal .phi..sub.b.sbsb.4 (.GAMMA.) shown below as,

The composite summed output .phi..sub.T.sbsb.4 (.GAMMA.) of the filter matched to code pair No. 4 and provided by the adder 46.sub.4 is

which verifies that there would be no interference from the simultaneous presence of code pairs 5 and 7 at the output of the filter matched to code pair number 4 at .GAMMA.=0 .

Although 4 bit and 8 bit code mate pairs have been utilized above, for purposes of illustration, the quantity of available unique noise codes whose crosscorrelation value equals zero at .GAMMA.=0 will be greater than the time-bandwidth product, possibly by a very large factor, for moderate to large time-bandwidth values (i.e. 50 to 1000). The two sets of codes listed in Tables 1 and 2 are both equal to the time-bandwidth product (T.times.B=n=no. of code bits) even though the quantity of perfect noise codes that could be generated are somewhat restrictive for a limited 4 bit and 8 bit code pair. As the code bit quantity increases to moderate values, the number of different multiplexed noise codes that can be generated rapidly approaches infinity as a result of the powerful and general expansion rules that have been discovered for the general class of multiplexed noise codes. Accordingly for any application, the required number of code bits will never be greater than the required number of bits switched in a prior art digital switch using gates and, moreover, will be significantly less so that a coded switched system results which is less complex. In addition, very large switch line capacities that would ordinarily be impractical for pure time gate switching, due to bandwidth and bit rate limitations, now become feasible when utilizing coded switching as set forth in this specification. While orthogonal operation has been shown and described, it should also be pointed out that even larger quantities of noise code sets become available for a given application is some degree of non-orthogonal operation is allowed. The performance would be slightly degraded, but this would occur gracefully.

Thus what has been shown and described is a multiplexed noise coded line selection system wherein each selected input/output line utilizes a different unique noise code. This is made possible by the use of multiplexed noise codes having an impulse autocorrelation function and which have the unique property of comprising code subsets whose crosscorrelation value is zero at .GAMMA.=0 which is the time when the autocorrelation function of each specific code pair compresses to an impulse.

An important advantage associated with noise coded switching in accordance with the subject invention is that the noise codes can serve a multi-purpose function that is capable of satisfying the most difficult requirements associated with certain types of communications systems, for example, tactical communications systems. The codes can be utilized to perform the signalling function of a mobile access system with little or no self-interference, provide a large degree of anti-jam protection, enjoy a low probability of intercept and meet communications security requirements while interoperating directly with a multi-channel switching system that automatically switches the code to any required destination point over a trunk line network.

Having shown and described what is at present considered to be the preferred embodiment of the invention, it should be understood that the same has been made by way of illustration and not of limitation and accordingly all modifications, alternations and changes coming within the spirit and scope of the invention are herein meant to be included.