Title of Invention

"A DATA COMMUNICATION APPARATUS FOR COMMUNICATION BETWEEN A MOBILE STATION AND BASE STATION"

Abstract

Method and apparatus for transmitting data frames, and a method and apparatus for data rate matching Elements to be transmitted are distributed over a number of radio frames by means of an interleaver and are punctured or repeated, with the puncturing or repetition being carried out in such a manner that, when it is related to the original arrangement of the element before interleaving, the pattern avoids puncturing or repetition of adjacent elements, or of elements which are not far apart from one another.

Full Text

Description Method and apparatus for transmitting data frames, and a method and apparatus for data rate matching The present invention relates to a method and an apparatus for transmitting data frames, and to a method and an apparatus for data rate matching, in particular using puncturing and/or repetition. Digital communications systems are designed for transmitting data by representing the data in a form "which makes it easier to transmit the data via a communication medium. For example, in the case of radio transmissions, the data is transmitted between transmitters and receivers in the communications system in the form of radio signals. In the case of broadband telecommunications networks, the data can be in the form of light, and can be transmitted, for example, via a fiber-optical network between transmitters and receivers in the system. During data transmission, bits or symbols in the transmitted data may be corrupted, which means that these bits or symbols cannot be determined correctly in the receiver. For this reason, the data communications systems frequently contain means for ameliorating the corruption of the data which occurs during transmission. One of these means is to equip transmitters in the system with coders, which use an error control code to code the data before transmission. The error control code is designed such that it adds redundancy to the data, in a controlled manner. In the receiver, errors which occur during transmission can be corrected by decoding the error control code, as a result of which the original data is reproduced. The decoding is carried out using an error decoding algorithm, which corresponds to the error control code, which is known to the receiver. Once the data has been decoded, it is often necessary, for data rate matching, to puncture or to repeat data bits or symbols from a block of coded data, before such data is transmitted. In this context, the term puncturing means a process of removing or deleting bits from a coded data block, with the effect that the punctured bit is not transmitted with this data block. Puncturing could be required, for example, because a multiple access method which is used for transmitting the data via the data-carrying media requires formatting of the data to form blocks of predetermined size, which size does not correspond to the size of the coded data frame. In order to accommodate a coded data frame in a transport data block having a predetermined size, data bits are therefore either punctured from the coded data frame in order to reduce the size of the coded data block in a situation in which the coded data frame is larger than the size of the transport block, or bits in the coded data frame are repeated, in a situation in which the coded data frame is smaller than the predetermined size of the transport block. In a situation in which the data frame is smaller than the transport data block, the data bits or symbols are repeated to the extent necessary to fill the rest of the transport data block. Those skilled in the art are familiar with the fact that one effect of puncturing a coded data frame is that the probability of correct reproduction of the original data is reduced. Furthermore, the performance of known error control codes and the decoders of these error control codes is best when the errors which occur during the transmission of the data are caused by Gaussian noise, since this has the effect that the errors are distributed independently throughout the transport data block. When a coded data frame is intended to be punctured, the positions in the coded data frame at which bits are punctured should be separated as far as possible from one another. To this extent, the puncturing positions should be distributed uniformly throughout the data frames. Since errors during transmission frequently occur in bursts, particularly in the case of radio communications systems which do not use interleaving, and since the repetitions are not intended to particularly improve the quality just in a certain region of the data frame but should be as uniform as possible, positions in a coded or uncoded data frame in which data bits are intended to be repeated should be arranged similarly so that they are uniformly separated from one another throughout the entire data frame . Known methods for selecting positions of bits or symbols which are intended to be punctured or repeated in a coded data frame include the division of the number of bits or symbols in a frame by the number of bits or symbols which are intended to be punctured, and the selection of positions with integer values corresponding to the division. In a situation in which the number of bits to be punctured is not an integer division of the number of bits in the data frame, this does not, however, lead to uniform spacings between the punctured or repeated positions, thus resulting in the disadvantage that certain positions are closer than this integer number, or in some cases even alongside one another. In order to describe the complex invention, the technical field of the invention and the problems that occur in this case will be briefly explained in the following text with reference to Figures 1 to 6 but, at least partially, also result from the state of standardization for the 3rd mobile radio generation (UMTS (Universal Mobile Telecommunications System)) prior to the invention, which is specified in particular in the following document: SI.12 vO.0.1, 3GPP FDD, Multiplexing, channel coding and interleaving description. The interleaving in a transport multiplexing method is frequently carried out in two steps. The various solutions for carrying out the puncturing/repetition have specific consequences if the puncturing is carried out after the first interleaver, as is envisaged from the UMTS system. It can be assumed that the puncturing will be useful both in the uplink direction and in the downlink direction in order, for example, to avoid multicode. The current state of the specification for the UMTS system results in a potential problem, since, when using FS-MIL (FS-Multistage Interleaver) as the interleaver in the uplink direction multiplexing methods (Figure 1) in conjunction with the current rate matching algorithm proposed for UMTS, the performance could deteriorate. As an example, let us consider a situation in which layer 2 results in a transport block with 160 bits on a transport channel with a transmission interval of 80 ms. This bit sequence can also be described as a data frame, or as a sequence of data frames. This means that, after the first interleaver, (first interleaving), the data is interleaved over eight frames (also often referred to as a radio frame in the following text) (see Figure 2) . Let us now assume that four bits are intended to be punctured in each frame (radio frame) in order to produce a balance for the requirements for the quality of the service of this transport channel together with other channels. The result of the rate matching algorithm (which is intended for the UMTS system and is also, for simplicity, referred to as the rate matching algorithm in the following text) (where e=Nc) is that the bits 4, 9, 14 and 19 (index starts at 0, numbering based on the sequence of the bits after the first interleaving) should be punctured in each frame (radio frame) . In Figure 2, a punctured bit is illustrated in bold text. In consequence, eight adjacent bits are punctured, and this, as explained above, is undesirable. One obvious procedure to avoid .this problem would be to shift the puncturing pattern in each frame. Let us assume that NI is the number of bits in a frame before rate matching, Nc is the number of bits after rate matching, mi is the index of the punctured/repeated bit, k is the frame number and K is the number of interleaved frames. Let us then consider the situation where Ni>Nc, that is to say puncturing. In the above example, Ni=20, Nc=16, mi=4, m2=9, m3=14, m4=19, k=1...7 and K=8. The shift could then be achieved using the following formula : mj shift = (MJ + k*fNc/ (NC-NC) /Kl) mod Ni, where [ 1 means round up. The same example as above would then give the result in Figure 3. As can be seen from Figure 3, the puncturing of adjacent bits is admittedly avoided to a certain extent, but, however, there is a cyclic effect or edge effect, that is to say for example, the two bits 43 and 44 are punctured. If the puncturing ratio is low, the probability of puncturing adjacent bits decreases. Figure 4 shows an example with 10% puncturing. As can be seen from Figure 4, some adjacent bits are still punctured. It is thus possible for a decrease in performance to occur. If the first interleaver is optimized and the second interleaver is kept simple, then the puncturing no longer requires the described rate matching algorithm. An optimized first interleaver should reorder the bits such that adjacent bits are separated. The puncturing can accordingly be carried out simply by removing successive bits after the interleaving process. However, there are two options. Let us consider the scenario illustrated in Figure 5. The four blocks on TrCH A are interleaved together, and the rate matching is then carried out. When puncturing is used, successive bits are removed in each frame. It is therefore highly improbable that any punctured bits would be adjacent in a frame after the coding process. However, there is no guarantee that punctured bits would not be adjacent in different frames after the coding process. In consequence, a reduction in performance could occur when using this approach. One alternative is to puncture successive bits only occasionally in individual transmission time intervals. The disadvantage of this approach is that bits on TrCH A are repeated at a time of 30 ms, since there is no data on TrCH B. It would probably have been better to reduce the extent of puncturing instead of puncturing a number of further bits. This problem has already been mentioned and was one of the reasons for combining static and dynamic rate matching. However, combined rate matching would also result in further advantages if this approach were to be used. Non-realtime transport blocks (NRT transport blocks) can still be transmitted, provided modifications are carried out to the original NRT concept. In the original proposal, it was possible to increase the puncturing and in this way to create space for the NRT block - although this would not be feasible with this new approach. The restriction in the above example was that the NRT block or the NRT blocks had to be shorter than, or precisely the same length as, the transport blocks in TrCH B. In situations in which repetition is used, the number of repeated bits may, however, naturally be reduced, in order to create space for the NRT blocks. The problem for puncturing when FS-MIL is used in the uplink path multiplexing method has been mentioned. This problem occurs when rate matching is carried out after the first interleaving. When the current rate matching algorithm is used for an output from the first interleaver (intermediate frame FS-MIL) , the number of adjacent bits in the specific line are punctured as shown in Figure 2. In order to avoid this, the shifting of the puncturing patterns is then introduced, in Figure 3. However, some adjacent bits are still punctured as a result of a cyclic effect or edge effect, resulting in certain deteriorations in performance. The following modification for rate matching at that particular time could be effective to solve the above problem; that is to say puncturing using a simple shift rule prior to column randomization of the intermediate frame FS-MIL (the expression "line-by-line processing" has been changed to "line-by-line randomization" in order to make it easier to understand the major characteristics of the processing blocks). Figure 6 shows an example of puncturing patterns when this modification is carried out for the same bit sequence example as before. The rate matching with a shift is carried out immediately after the block interleaving in the first stage. No puncturing of adjacent bits can now be seen in this figure. This puncturing should thus not result in any reduction in performance. In fact, there is no need to carry out the above rate matching before the column randomization. The equivalent rate matching could be carried out after the column randomization by taking account of the column randomization rules, and this could easily be achieved just by replacing the initial offset value of the puncturing by a simple formula. The details of the modified rate matching algorithm are shown in List 1. This list introduces e0ffset in order to set the initial offset in each frame for uplink path rate matching. The offset is not calculated on the basis of the column number after column randomization, but before column randomization, and this can be calculated using the inverse column interchanging rule. Furthermore, e0ffset is not used just for puncturing, but also for repetition. Repetition bits could thus also be positioned more uniformly. The interleaving in the transport multiplexing method is carried out in two steps. As explained in the above sections, consequences of the various solutions have specific consequences on the uplink path. The following text shows that the previously proposed solutions, that is to say the proposed puncturing pattern, is still not always optimum in all situations. Against this background, the invention is based on the object of reducing these disadvantages of the prior art. This object is achieved by the features of the independent claims. Developments of the invention can be found in the dependent claims. Embodiments of the present invention will now be described just by way of example with reference to the attached drawings, in which: Figures 1 to 6 show the prior art; Figure 7 shows a block diagram of a mobile radio communications system; Figure 8 shows a block diagram of a data communications apparatus, which forms a path between the mobile station and a base station in the communications network shown in Figure 1; Figure 9 shows first interleaving of 80 ms and 1:8 puncturing with an improved algorithm: Figure 10 shows the principle of optimized puncturing Figure 11 shows a reference table Figure 12 shows first interleaving of 80 ms and 1:5 puncturing Figure 13 shows 1:8 puncturing using the proposed algorithm Figure 14 shows an odd number of bits per frame Figure 15 shows puncturing patterns An exemplary embodiment of the present invention will be described in the context of a mobile radio communications system. Mobile radio communications systems are equipped with multiple access systems which operate, for example, on the basis of time division multiple access (TDMA) as is used, for example, in the global mobile radio system (GSM), a mobile radio communications standard which is standardized by the European Telecommunications Standard Institution. As an alternative, the mobile radio communications system could be equipped with a multiple access system operating using code division multiple access (CDMA), such as the UMTS system proposed for the third-generation universal mobile telecommunications system. However, as can be seen, any desired data communications system could be used to represent an exemplary embodiment of the present invention, such as a local data network or a broadband telecommunications network operating using the asynchronous transmission mode. These examples of data communications systems are characterized in particular in that data is transmitted as frames, packets or blocks. In the case of a mobile radio communications system, the data is transmitted within radio signals which carry data and represent a predetermined amount of data. Figure 7 shows one example of such a mobile radio communications system. Figure 7 shows three base stations BS which exchange radio signals with mobile stations MS in a radio coverage area which is formed by cells 1, which are defined by dashed lines 2. The base 17- stations BS are coupled to a network relay system NET. The mobile stations MS and the base stations BS exchange data by using radio signals, which are annotated 4, to transmit between antennas 6, which are coupled to the mobile stations MS and to the base stations BS . The data is transmitted between the mobile stations MS and the base stations BS, using a data communications apparatus, in which the data is transformed into radio signals 4, which are transmitted to the receiving antenna 6, which identifies the radio signals. The data is reproduced from the radio signals by the receiver. Figure 8 shows an example of a data communications apparatus which forms a radio communication path between one of the mobile stations MS and one of the base stations BS, with parts which also appear in Figure 7 having identical numerical designations. In Figure 8, a data source 10 produces data frames 8 at a rate which is governed by the type of data produced by the source. The data frames 8 produced by the source 10 are supplied to a rate converter 12, which converts the data frames 8 to form transport data blocks 14 . The transport data blocks 14 are designed such that they are of essentially the same size, with a predetermined size and an amount of data which can be carried by frames in data-carrying radio signals, via which data is transmitted by a radio interface which is formed from a pair comprising a transmitter 18 and a receiver 22. The data transport block 14 is supplied to a radio access processor 16, which controls the sequence of transmission of the transport data block 14 via the radio access interface. The transport data block 14 is supplied at an appropriate time by the radio access processor 16 to a transmitter 18, which converts the transport data block to the frame of data-carrying radio signals, which are transmitted in a time interval which is allocated to that transmitter, in order to transmit the radio signals. In the receiver 22, a receiver antenna 6' identifies the radio signals and carries out downward conversion and reproduction of the data frame, and this is supplied to a radio access sequence control reversing apparatus 24. The radio access sequence control reversing apparatus 24 supplies the received data transport block to a frame conversion reversing apparatus 26 which is controlled by the multiple access sequence control reversing apparatus 24, and is supplied via a conductor 28. The rate conversion reversing apparatus 26 then supplies a representation of the reproduced data frame 8 to a destination or sink for the data frame 8, which is represented by the block 30. The rate converter 12 and the rate conversion reversing apparatus 26 are designed such that, as far as possible, they utilize the data-carrying capacity available in the transport data block 14 optimally. According to the exemplary embodiment of the present invention, this is done by means of the rate matching converter 12, which is used to code the data frame and then puncture or repeat data bits or symbols which are selected from the coded data frame, with the effect of producing a transport data block which fits into the data blocks 14. The rate converter 12 has a coder and a puncturer. The data frame 8 which is supplied to the coder is coded, in order to produce a coded data frame which is supplied to the puncturer. The coded data frame is then punctured by the puncturer, in order to produce the data transport block 14. It is assumed that the puncturing can be carried out both in the uplink direction and in the downlink direction. When the ETSI and ARIB specifications were joined together to form the UMTS specification, ARIB made the assumption that no puncturing is carried out in the uplink direction. It is assumed that the puncturing will also be useful in the uplink direction, in order, for example, to avoid multicode. There is then a potential problem since the performance could deteriorate when using FS-MIL in the uplink path multiplexing method in conjunction with the present rate matching algorithm. This has been shown with reference to Figure 2 by an example of an analysis of a situation in which layer 2 supplies a transport block with 160 bits on a transport channel with a transmission interval of 80 ms, subject to the precondition that four bits should be punctured in each frame. This means that eight adjacent bits are punctured, which is obviously undesirable. The proposal as shown in Figure 3 is to shift the puncturing pattern in each frame. This is then also equivalent to the use of puncturing before column mixing, if it is actually carried out before the intermediate frame interleaving. In fact, in contrast to the example in Figure 2, no adjacent punctured bits are produced in this example. However, in a method as shown in Figure 2, there are always still situations in which adjacent bits are punctured, depending on the puncturing rate. Figure 9 shows, for example, the situation Ni=16, Nc=14, mi=4, m2=14, k=1...7 and K=8. For the sake of simplicity, Figure 9 and Figure 10 show only the area before interleaving, in which, however, those bit positions which are punctured after interleaving are illustrated by marking them in bold print. As can be seen, the adjacent bits 31-32 and 95-96 are punctured, which is obviously undesirable. A first aim of a good puncturing algorithm is to distribute punctured bits as uniformly as possible over the bit positions in their original sequence. This was also the critical principle which was used for the definition of the puncturing algorithm for UMTS, as is described, for example, in the abovementioned Specification SI. 12. This is best done by puncturing every n-th bit or, in some cases, every (n+first) bit if the puncturing rates are not integral. A second aim is to puncture the various columns (in the following text, frames are also often referred to as columns) with equal frequency, and hence also to distribute the punctured bits uniformly over all the radio frames (frames), and also to achieve uniform puncturing in the various columns. The expressions puncturing or repetition of a column (for the frame) also mean the puncturing or repetition of an element in the column (in the frame). However, if the principle explained above is also applied to puncturing after interleaving, then the second aim can no longer be adequately achieved. Let us consider, for example, 80-ms interleaving and a puncturing rate of 1:6. Puncturing every sixth bit would result in only the columns 0, 2, 4, 6, but not 1, 3, 5, 7 being punctured, which is, of course, impossible. In order to achieve both aims, one embodiment variant of the invention provides for the puncturing interval to be changed at least once, and if necessary more than once, in order to avoid some columns being preferred for puncturing, while others, on the other hand, are not punctured at all. This is shown in Figure 10. Horizontal arrows (P6) with thin surrounding lines show a puncturing distance of 6, and the horizontal arrow (P5) with thick surrounding lines shows a puncturing distance of 5, in order to avoid puncturing the first column twice. Once each column has been punctured once, the pattern (as shown by the vertical arrows) can be shifted six lines downward, in order to define the next bits to be punctured. This obviously corresponds to the puncturing of every sixth bit in each column, that is to say it corresponds to the use of a standard rate matching algorithm, and to the shifting of puncturing patterns with respect to one another in different columns. STATEMENT OF THE INVENTION According to the present invention there is provided data rate matching apparatus for bits which are to be transmitted and are distributed by a first interleaver over a set of K frames, comprising a rate converter designed to carry out a puncturing or repetition method after the interleaving to puncture or repeat the same number of bits in each frame, the distance between punctured or repeated bits is varied with respect to the sequence of the bits before the first interleaver, bits to be punctured or to be repeated being obtainable by the following steps: a) determination of the integer component q of the mean puncturing distance where q:=(LNc/(lNrNcl) ), where 1 means rounding down and 11 means an absolute value, and where: Nj— the number of bits after rate matching, and Nc:= the number of bits before rate matching; b) selection of a bit to be punctured or to be repeated in a first frame; c) selection of the next bit to be punctured or to be repeated in a next frame on the basis of the most recent bit to be punctured or to be repeated in the previous frame, in that, starting with this most recent bit to be punctured or to be repeated, the next bit is in each case selected at a distance q, with respect to the original sequence before the interleaver, provided that this does not lead to a frame being punctured or repeated twice, otherwise a bit at a distance q-1 or q+1, which is not the same as q, is selected for puncturing or repetition; d) repetition of step c) until all of the frames have been punctured or repeated once. This method will now be described using formulae in the following text: Let us assume that NI is the number of bits in a frame before rate matching, Nc is the number of bits after rate matching, nij is the index of the punctured/repeated bits, k is the frame number and K is the number of interleaved frames. The aim is to consider mainly the situation Ni>Nc, that is to say puncturing, but the formulae are also applicable to repetition. In the above example, Ni=20, Nc=16, m1=4, m2=9, m3=14, m4=19, k=l...7 and K=8. The shifting could then be achieved using the following formula: -- Calculation of the mean puncturing distance q: = (l.Nc/(/Ni-Nc/) J) mod K — where L J means round down and I I means absolute value. Q:= (lNc/(/Ni-Nj)J) div K if q even -- deal with as a special case: then q = q - lcd(q, K)/K -- where lcd(q, K) means the highest common denominator of q and K -- It should be remembered that led can easily be calculated by bit manipulation, since K is a power of 2. -- For the same reason, calculations with q can easily be carried out using binary fixed-point arithmetic (or integer arithmetic and a small number of shift operations). endif -- Calculation of S and T; S represents the shift in the line mod K, and T represents the shift magnitude div K; S thus represents the shift in the line with respect to q (that is to say mod K) and T the magnitude of the shift with respect to Q (that is to say div K); for i = 0 to K-l S (RK ([i*q] mod K)) = ([i*q] div K) — where f 1 means round up. T((RK ([~i*ql mod K)) = i — RK(k) reverses the interleaver, end for In an actual implementation, these formulae can be implemented as shown in Figure 11, as a reference table. The table also includes the effect of remapping the column randomization achieved by RK(k). S can obviously also be calculated from T, as a further implementation option. eoffset can then be calculated as follows : eoffset (k) = ((2*S) + 2*T Q +1)* y + 1) mod 2Nc Using e0ffset (k) , e is then preloaded in the rate matching method for UMTS. This choice of eoffset obviously results in a shift in the puncturing patterns in the columns relative to one another by the amount S + T * Q. The following text describes a simplified representation which simply results from the calculation of q and Q not being carried out separately for the remainder in the division by K and the multiple of K, but being combined for both components. In the same way, S and T cannot be calculated separately for q and Q, but likewise combined. The substitutions q+K*Q --> q and S+Q*T --> S result in the following equivalent representation. Depending on the details of the implementation, it may be better to carry out one calculation method or the other calculation method (or further methods which are likewise equivalent to them). -- Calculation of the mean puncturing distance q:= (l.Nc/ (/Ni-Nc/) J) -- where L J means round down and / I means absolute value. if q even -- deal with as a special case: then q = q - lcd(q, K)/K -- where lcd(q, K) means the highest common denominator of q and K -- It should be noted that led can easily be calculated by bit manipulation, since K is a power of 2. -- For the same reason, calculations with q can easily be carried out using binary fixed-point arithmetic (or integer arithmetic and a small number of shift operations), endif -- Calculation of S(k) for the shift in the column k; for i = 0 to K-l S (RK ([i*q] mod K)) = ([i*q] div K) — where [ ] means round up. -- RK(k) reverses the interleaver end for ^offset can then be calculated as follows: eoffset (k) = ((2*S) * y + 1) mod 2Nc Using e0ffset (k) , e is then initialized in advance in the rate matching method. If the puncturing rate is an odd-numbered fraction, that is to say 1:5 or 1:9, this method produces the same perfect puncturing pattern as that which would be used directly before interleaving by puncturing using the rate matching method. In other situations, adjacent bits are never punctured, but the distance between punctured bits may be greater than the others by up to lcd(q,K)+l. This method can also be applied in a corresponding manner to bit repetitions. Although the repetition of adjacent bits does not have such a severe influence on the performance of the error correction codes as is the case when puncturing adjacent bits, it is nevertheless advantageous to distribute repeated bits as uniformly as possible. The fundamental objective of this method is to achieve a uniform distance between the punctured bits in the original sequence, but taking account of the constraint that the same number of bits must be punctured in the various frames. This is achieved be reducing the puncturing distance by 1 in certain cases. The described method is optimum to the extent that the distance is never reduced by more than 1, and it is reduced only as often as is necessary. This results in the best-possible puncturing pattern subject to the constraints mentioned above. The following example shows the use of the first set of parameters, that is to say puncturing with 1:5 (Figure 12) . The optimized algorithm obviously not only avoids the puncturing of adjacent bits, but punctured bits are also distributed with the same spacing in the original sequence. In fact, the same characteristics are achieved as if the puncturing were to be carried out directly after the coding and before the interleaving. We will now investigate the next case, that is to say puncturing with 1:8 (Figure 13) . Once again, the puncturing of adjacent bits is avoided. In this case, it is impossible to achieve uniformly spaced puncturing, since all the bits in an individual frame would then be punctured, which is completely unacceptable. In this case, most of the distances between adjacent bits are 7 (only one less than with an optimum distribution) . In this case, some distances are greater (every eighth) . In two situations, the rate matching may vary during the transmission time interval: a) The number Ni of input bits is not divisible by K. The last frames then have one bit less than the first, and therefore also have a somewhat lower puncturing rate. It should be remembered that it is not clear whether this situation will be permissible or whether it is expected that the coding will supply a suitable number. b) Owing to fluctuations in other services which are multiplexed onto the same link, the puncturing may be weakened in later frames . In these situations, the balanced puncturing method could still suffer from disadvantages. Owing to the unpredictable nature of case b) , it appears to be improbable that it will be possible to find any method whatsoever which could lead to a virtually perfect puncturing pattern, and in this situation it is therefore necessary to accept a certain unpredictable behavior in each case. However, in case a), it is proposed that the puncturing pattern should not be varied in the last lines. Instead of this, it is proposed that the same puncturing algorithm be used as for the first columns, but simply with the last puncturing being omitted. Let us consider, as an example, a situation in which 125 input bits are intended to be punctured, in order to obtain 104 output bits which are interleaved over eight frames. The puncturing pattern would then appear as shown in Figure 14. The last columns have one input bit less than the first while, due to the omission of the last puncturing, the columns all have 13 bits. Furthermore, it is proposed as an alternative that an optimized first interleaver be used, with a simple second interleaver and a simple puncturing method being used. This is based on the expectation that an optimized interleaver will distribute bits such that the puncturing of blocks of bits after the interleaving will distribute these punctured bits uniformly before interleaving. However, experience with puncturing after a simple first interleaver has shown that this is not an easy task. Since the individual interleaver cannot be optimized for all puncturing rates, it is virtually impossible to achieve good characteristics: the reason for this is as follows: the puncturing patterns (Figure 15) for n+1 bits must be identical to the puncturing pattern for n bits, although an additional bit can be chosen for puncturing. If the puncturing pattern is good for n bits (see the first line in the table in Figure 15), then it is impossible to achieve an optimum distribution of n+1 bits (last line) irrespective of which specific bit is additionally punctured in order to obtain n+1 bits (second line). Furthermore, such an interleaver would need to be a compromise between good puncturing characteristics for block puncturing and, at the same time, good general interleaving characteristics (for example in order to achieve good transmission characteristics for transmission via fading channels). Since no such method and no such interleaver are known, the method described in the present application is particularly advantageous, in which puncturing is carried out after a simple first interleaver with a subsequent second interleaver with optimized interleaving characteristics. Virtually optimum puncturing patterns are thus possible, if the rate matching is carried out after the first interleaving. The method is simple, requires little computation power and need be carried out only once per frame, and not once per bit. The method described above is not restricted to radio transmission systems. WE CLAIM:- 1. A data communication apparatus for communication between a mobile station and base station comprising: a rate converter (12) for converting data frames (8) to form transport data block (14), the transport data block being essentially of same size; a radio access processor (16) to control the sequence of transmission of the transport data blocks (14) via the radio access interface to a transmitter (18); a receiver antenna (6') to identify the radio signal from the transmitter (18) and to carry out downward conversion; a radio access sequence control reversing apparatus (24) to receive the data transport block (14) and supply it to a frame conversion reversing apparatus (26) which is controlled by said sequence control reversing apparatus (24); said rate conversion reversing apparatus (26) configured to supply the representation of data frame (8) to a destination (30) or sink (3); wherein the rate converter carries out a puncturing or repetition after the interleaving to puncture or repeat the same number of bits in each frame, the distance between punctured or repeated bits is varied with respect to the sequence of the bits before the first interleaver. 2. A data communication apparatus for communication between a mobile station and base station, substantially as hereinbefore described with the accompanying drawings.

Documents:

abstract.jpg

in-pct-2001-00912-del-abstract.pdf

in-pct-2001-00912-del-claims.pdf

in-pct-2001-00912-del-correspondence-others.pdf

in-pct-2001-00912-del-correspondence-po.pdf

in-pct-2001-00912-del-description (complete).pdf

in-pct-2001-00912-del-drawings.pdf

in-pct-2001-00912-del-form-1.pdf

in-pct-2001-00912-del-form-13.pdf

in-pct-2001-00912-del-form-19.pdf

in-pct-2001-00912-del-form-2.pdf

in-pct-2001-00912-del-form-3.pdf

in-pct-2001-00912-del-form-5.pdf

in-pct-2001-00912-del-gpa.pdf

in-pct-2001-00912-del-petition-137.pdf

in-pct-2001-00912-del-petition-138.pdf