CA2129224A1

CA2129224A1 - Procedure for recovering test data transmitted on a transmission line for data line for data flows with cellular structure

Info

Publication number: CA2129224A1
Application number: CA002129224A
Authority: CA
Inventors: Andreas Wolf
Original assignee: Individual
Current assignee: Siemens AG
Priority date: 1992-01-31
Filing date: 1993-01-22
Publication date: 1993-08-01
Also published as: NO942802D0; NO942802L; DE4203298C1; EP0624295A1; AU3343993A; WO1993015575A1

Abstract

(57) Abstract The invention relates to a process for rec-overing transmitted test data in cellular digital streams at the receiving end in which the test data are coded by means of code words which, in self-correlation, generate a Dirac pulse and in which a cross-correlation is undertaken on the reception side for decoding. In order to be able to transmit a larger number of test data in coded form, a binary zero word (stored at St=1) is added to the code words on the transmission side and the various test data (e.g. 11101) are allocated to the various code words and the zero word. The sum of the digits of the code words received and the zero word is made in addition to the cross-correlation on the reception side.

Description

FiLE. ~I THI~ A~ Fit-D 21Z~2~ -~ TRANSLATION
A Procedure for Recovering Test Data Transmitted on a Transmission Line for Data Flows with Cellular Structure The present invention relates to a procedure for recovering test data that are transmitted from the sending end to the receiving end of a transmission line for digital data flows with cellular structure at the receiving end - in which the various test data are encoded at the sending end of the transmission line by means of a binary codeword that, autocorrelated, results in a Dirac pulse, and by additional binary codewords that are generated from the codeword by cyclical shifting of the codeword and - in which, at the receiving end of the transmission line, for purposes of decoding, a cross-correlation of each codeword that is received is made, while evaluating the main maxima of each cross-correlation function in order to recover the test data. Such a procedure is used mainly for measurement techniques for wide-band ISDN transmission technology.

A known procedure of this kind is described in DE 40 12 850 Al.
In this known procedure to determine the quality parameters of a transmission line for digital data flows, numbers of the test cells that are introduced into the data flow are encoded at the ; -~
sending end of the tr~nsmission line, using a codeword that, autocorrelated, generates a Dirac pulse; a codeword of this kind is, for example, an m-sequence or a codeword from the codeword inventory of the Barker Code. In addition, at the sending end, within the framework of the encoding process, the codeword i5 shifted from test cell to test cell, in each instance by an identical amount. This means that additional codewords can be formed by cyclical shifting of the one codeword. At the receiving end of the transmission line, a correlator is associated with a receiver, and this performs a cross correlation

2~ 21f of the particular codeword that is received in each instance. A
conclusion with respect to the bit error rate can be formed from the level of the main maxima of the cross correlation functions that are formed in this manner, whereas the relative position of the main maxima with respect to time permits a statement with respect to test cell number, whereby the sending end test data can be recovered. This takes place very reliably in the known procedure, since it still works very accurately even at a high bit error rate; as example, with a 25-m-sequence as a codeword, -~
up to seven bit errors can occur without any degradation of the accuracy of the procedure.

The known procedure is described in the journal telecom rePort, No. 14 (1991), Vol. 2, pp 104-107. In the embodiment that is described all of the other binary codewords are formed from the one codeword by cyclical shifting.

It is the task of the present invention to so develop the known procedure as to permit the accurate recovery of a large number of -test data, present at the sending end, at the receiving end.

According to the present invention, in order to do this, in a procedure of the type described in the introduction hereto . . , at the sending end, a binary null-word is added to the codewords and the various test data are associated with the ~-different codewords and the null-word, and at the receiving end of the transmission line, for decoding, in addition to the cross correlation, the codewords and the null- word that have been received are cross summed, when main maxima that are above a preset value of the cross correlation function indicated codewords that have been transmitted and are used to recover the associated test data, and 2~Z~Z~

- a cross sum that is below a preset additional value indicates the null word that has been transmitted and is used to recover a test datum that is associated with the null word.
The important advantage of the procedure according to the present invention is that because of the use of a null-word (in such a word, all the binary positions are set at "0"), the inventory of the encoding words that are available is extended by a word. For example, if a 25-m-sequence is used, the present invention leads - -to an inventory of a total of 32 different encoding words. Thus, it is possible to encode a total of 32 different (each varying over a length of 5 bits) test data, and recover these reliably at the receiving end.

The inventory of encoding words at the sending end can be advantageously doubled in that - at the sending end, inverse binary code words are formed by -~
inversion, and a binary one-word is added to these inverse codewords, and additional test data are associated with the inverse codewords and the one word;
- at the receiving end, main maxima of the cross-correlation that lie above a preset lower limit characterize codewords that have been transmitted, and are used to recover the associated test data, while taking into consideration the sign of the main maxima, or - a cross sum that is below a preset, additional value indicates the null-word that has been transmitted and is used to recover the test datum that is associated with the null-word, and - a cross sum that is above an additional, preset value indicates a one-word and is used in recovering a test datum that is associated with the one-word. Of course, the registration of main maxima of various kinds can also be effected in that the value range of the cross correlation functions is shifted by such a value that the values of the ;~ 2Zd~

main maxima of various kinds are always in the positive range.

In the procedure according to the present invention, the test data that are transmitted from the sending end and recovered at -the receiving end can be further processed in different ways.

Like the procedure described in the introduction hereto, the procedure according to the present invention permits not only the recovery of the test data at the receiving end; in the evaluation of the level of the main maxima of the cross correlations functions of the codewords, it also permits a statement with respect to the number of bit errors; such a statement is also ~;
possible with respect to the null-word or the one-word, because the number of bit errors can be determined from the cross sum.

The present invention will be described below on the basis of the drawings appended hereto. These drawings show the following:
igure 1: a block schematic diagram of one embodiment of an arrangement for carrying out the procedure according to the present invention. ~ ;
Figure 2: An array of encoding words.
Figure 3: An embodiment of an encoder of an arrangement for an additional embodiment of the procedure according to the present invention.
Figure 4: An embodiment of the receiving end of the same arrangement for carrying out this additional embodiment of the procedure according to the present invention. ~ ;

The embodiment that is shown in Figure 1 incorporates, on the input side, an encoding generator 1 that consists, for example, of an addressable memory. In the example that is shown, this ~ `
memory has at position St = 2 (see Figure 2) a 31-bit long code 23 ;2~3~2~

word that corresponds to a 25-m-sequence. At the other positions

3 to 32 there are additional 31-bit long code words that are the result of cyclical shifting of the 25-m-sequence; the second column of Figure 2 shows the phases s of the code words that result from this. A 31-bit long null-word is stored in the position St = 1 of the memory and this consists entirely of zeros, as can be seen from Figure 2.

Thus, a total of 32 encoding words are available. As a consequence of this, 32 different test data can be encoded, and each of these differs from the other by one bit over a total of five bits. Consequently, only an uncoded 5-bit long data word is recorded as a test datum in a block 2 of Figure 1.

The encoded test data generated by means of the encoding word generator 1 (Block 3 in Figure 1) are transmitted on a transmission line 4 and, in the form of encoded test data that has been more or less corrupted (Block 5 in Figure 1), reach the receiving end 6, where decoding takes place. This is done, amongst other things, in that a cross correlation of the transmitted test data with reference data, for example, an m-sequence mO (Block 8 in Figure 1) is made in a cross-correlator 7. The resulting cross correlation functions, which appear largely as shown in DE 40 12 850 Al referred to in the introduction, are examined in a block 9 for the position of the main maxima. The relative time position of the main maximum of the particular cross correlation function contains information about the test datum transmitted in each instance, so that the test datum that has been transmitted can be recovered by way of an additional memory on the receiving side 6, not shown herein, from the relative time position of the particular main maximum that has been determined.

Identification of the null-word is made in Block 10, and is done such that a cross sum is formed here. If the cross sum is below ~Z~3~24 \

PCT~DE 93/00066 6 GR 92 P 4008 P -a preselected value, for example, below the value 8 if a 25-m-sequence is used, this is an indication that the coded datum that has been transmitted is the null-word. Accordingly, the associated test datum can be recovered by means of the memory (not shown herein), within the framework of the decoding process (Block 11 in Figure 1).

In order to determine the bit error, the main maxima of the cross correlation function are tested for their level. When this is done, it is ensured that this occurs only in the case of such main maxima whose level is above a preset value; in the case of a 25-m-sequence, there is an evaluation of only the main maxima -whose value is above the number 16. This ensures that a main ~ ;
maxima that refers back to the null word is not detected.
Determination of the bit error is effected from main maxima whose level is above the value 16 (Block 12 in figure 1).
:
The cross sum is examined in order to determine the number of bit errors by means of the null word; the level of this null word is a direct indication of the number of bit errors.
. ~, In addition, is should be pointed out that the procedure according to the present invention is not restricted to the use of a 25-m-sequence (with its cyclic shift), but can be carried out with other m-sequences. It is also possible to use the Barker Code, the Gordon, Mills and Welch sequence, or the Gold Code.

Figure 3 shows another example for a encoding-word generator on the sending end of a transmission line (not shown herein), for the case that test data that vary over a length of 6 bits, i.e., have a 6-bit long data word D(x), are to be encoded. Here, the generator incorporates a back-coupled shift register 20 with a plurality of stages 21 to 24, and exclusive OR elements 25 and 26. An additional exclusive OR element is associated with the " ~' ~3 r` 5--`~24 shift register 20, in that it is connected by an input 28 directly to the output of the exclusive OR element 26, and can be connected through a changeover switch 29 with the output of the stage 24 in its position a. When the changeover switch is in position b, one input 28 is acted upon with the signals of the lower five positions d1 to d5 of the data word D(x). The additional input 30 of the additional exclusive OR element 27 is acted upon by the uppermost (sixth) position of the data word D(x). The sixth position controls the inversion so that, unlike the process shown in Figures 1 and 2, here there are 64 different encoding words available, if one proceeds from a 25-m sequence as an example.

This means that data C(x) encoded with the data words D(x), corresponding to the test data that are to be transmitted, appear at the output 31 of the additional exclusive OR element 27 and, for example, in addition to a 25-m sequence as the "basic codeword," these can also contain thirty binary codewords obtained by cyclical shifting, thirty-one additional code words derived therefrom by inversion, the null-word, and the one-word.

These encoded data C(x) are transmitted on a transmission line (not shown herein) and arrive at the receiving end, the configuration of which is shown in Figure 4. Here, as in the example shown in figures 1 and 2, a cross-sum of the encoded data C(x) that has been transmitted is formed in a block 40. This cross sum is once again tested in an associated block 42, to ascertain whether or not it is less than eight. If this is so, then the test datum that has been transmitted is detected as a null-word. If the cross sum is greater than twenty-three, it is a binary one-word (all the binary positions are set at "1").

During cross correlation of the remaining codewords, in this example, inverse codewords result in negative main maxima of the ~
'~ '.

:; ,- . . . . , , . ; . . . `, .. ~ , -,- . .. .

2~3~2~ ::
\

cross-relation functions, and the "normal" codewords can be distinguished from the inverse codewords during decoding. The total of the main maxima is used here to differentiate between the remaining codewords and the null-words or one-words; totals of the main maxima that are above a preset lower value distinguish the remaining code words. In the case of a 25-m sequence as the basic codeword, the lower value is 16, as in the example described above.
:
According to the examples shown in Figures 1 and 2, the test data D(x) that are to be transmitted can be recovered in this way at the receiving end, in block 42.
:.
For the remainder, the receiving end shown in Figure 4 can be structured as is shown in Figure 1. Determination of the bit errors is effected in a similar way. Here, too, the codes listed above can be used.

~,~ .. : :-:

Claims

Patent Claims

1. A procedure for recovering test data that are transmitted from the sending end to the receiving end (4) of a transmission line for digital data flows with cellular structure at the receiving end - in which the various test data are encoded at the sending end of the transmission line by means of a binary codeword that, autocorrelated, results in a Dirac pulse, and by additional binary codewords that are generated from the codeword by cyclical shifting of the codeword and - in which, at the receiving end (6) of the transmission line (4), for purposes of decoding, a cross-correlation of each codeword that is received is made, while evaluating the main maxima of each cross-correlation function in order to recover the test data, characterized in that - at the sending end, a binary null-word (stored at St = 1) is added to the codewords and the various test data (e.g., 11101) are associated with the different codewords and the null-word, and - at the receiving end (6) of the transmission line, for decoding, in addition to the cross correlation, the codewords and the null- word that have been received are cross summed, when - main maxima that are above a preset value of the cross correlation function indicate codewords that have been transmitted and are used to recover the associated test data (e.g., 11101), and - a cross sum that is below a preset additional value indicates the null word that has been transmitted and is used to recover a test datum that is associated with the null word.

2. A procedure as defined in Claim 1, characterized in that - at the sending end, inverse, binary codewords are formed from the codewords by inversion, and a binary one-word is added to these inverse codewords and additional test data are associated with the inverse code words and the one-word;
- at the receiving end (6), main maxima of the cross correlation functions that are above a preset lower value indicated codewords that have been transmitted and are used in recovering the associated test data, while considering the sign of the main maxima, or - a cross sum that is below a preset, additional value indicates the null-word that has been transmitted and is used to recover the test datum that is associated with the null-word, and - a cross sum that is above an additional preset value indicates a one-word and is used to recover the test datum that is associated with the one-word.

3. A procedure as defined in Claim 1, characterized in that - the level of the main maxima and the size of the cross sum are evaluated as a measure of the bit-error rate.