EP1518345A1

EP1518345A1 - Rate matching method

Info

Publication number: EP1518345A1
Application number: EP03762396A
Authority: EP
Inventors: Martin DÖTTLING; Jürgen MICHEL; Bernhard Raaf; Ralf Wiedmann
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2002-07-02
Filing date: 2003-06-11
Publication date: 2005-03-30
Also published as: DE10229640A1; AU2003250748A1; WO2004006489A1

Abstract

The invention relates to method for matching the rate with a predetermined data rate. According to the method, the data whose data rate is matched is composed of at least two coded data streams and the rate of the data rate of the at least two data streams is matched with the predetermined data rate using a rate matching pattern by means of which the schema for repeating or puncturing individual data sequences from the data streams is determined and which is identical for the at least two data streams.

Description

description

Rate adjustment procedure

The present invention relates to a method for rate adaptation, in which the bit rate of data streams is adapted to a fixed bit rate, for example a physical channel.

Transmission channels in mobile radio systems, for example, only offer fixed data or raw data transmission rates due to their embedding in transmission formats, while the transmission or reception data rates of different signals or applications differ. It is therefore generally necessary to adapt the data rates to one another at an interface.

Such an adjustment or rate adjustment is described below using an example from the UMTS standardization:

In UMTS (Universal Mobile Telecommunication System) data packets are sent to a mobile station (UE = User Equipment) via the "High-Speed Downlink Shared Channel" (HS-DSCH). The associated control information is transmitted via the "High-Speed Shared Control Channel" (HS-SCCH), such as, for example, the channelization codes used for the HS-DSCH, which are codes with which transmissions are spread on a receiver-specific basis and the modulation scheme, for example QPSK (Quadrature Phase Shift Keying) or 16QAM (16 Quadrature Amplitude Modulation). So that the receiving mobile station can recognize that the information on the HS SCCH is intended for them, this control information or this user data is linked to identification information. In this context one speaks of a masking of the data. Before the link is made, both useful data and identification data are encoded and a subsequent rate adjustment is carried out.

However, this process is quite complex, which is disadvantageous in particular in the case of the mobile radio terminal in that these coding and rate adjustment processes are resolved again in order to obtain the original (user) data.

Starting from this prior art, it is the object of the present invention to carry out the rate adaptation on a channel which is shared by a plurality of communication participants and has low complexity.

This object is achieved by a method according to the features of independent claim 1. Advantageous pus formations result from the dependent claims.

The essence of the invention is to design the rate adaptation for user data and identification data in the overall coding of a channel used by a plurality of communication subscribers, with the aid of which it is made clear, for whom the data is intended, according to a common scheme. This has the advantage that the complexity of the decoding is reduced, in particular on the receiver side. Another aspect of the invention aims at the design of a rate adjustment pattern which allows the rate adjustment for user data and identification data according to a common scheme while maintaining the original information as well as possible. A rate adjustment pattern indicates which Bits from a data sequence are repeated or shortened (punctured) in order to obtain the desired data rate.

Advantages and refinements of the invention are explained below by way of example, partly with reference to figures.

Show it

1 shows an overview of the overall coding for a channel in which the data to be transmitted are masked using the identification data; FIG. 2 shows a diagram which represents individual processes in the overall coding;

3 shows the previous implementation of the overall coding in the HS-SCCH according to the prior art; FIG. 4 shows an exemplary embodiment of an implementation of the total coding in the HS-SCCH according to the invention;

FIG. 5 shows an exemplary implementation on the receiver side for receiving the HS-SCCH in the currently used specification (release 99); FIG. 6 shows an exemplary embodiment of the implementation on the receiver side in the case of overall coding in accordance with the proposal shown in FIG. 4.

Overall coding of useful and identification data FIG. 1

FIG. 1 schematically shows an overall coding for user data (LD: Load Data) and identification data (ID: Identification Data), which are sent via a shared channel in a communication system. Transmitted data (TD: Transferred Data) are composed of user data LD, which are linked to the identification data ID. are linked in order to indicate for which recipient the transmitted data TD are intended. Linking

User data LD and identification data ID take place in the context of an overall coding (CC: Channel Coding), mostly a channel coding. Channel coding is understood to mean the adaptation of digital values to the physical transmission medium, that is to say, for example, coding with subsequent

Rate adjustment.

In this case, total coding is understood to mean the coding, rate adaptation and linking of the useful and identification data. However, it is not absolutely necessary for all of the steps listed to take place; the overall coding can also consist, for example, of coding alone without rate adjustment.

The scheme shown in FIG. 1 is known per se, but the prior art and the invention differ in the procedure for the overall coding.

Figure 2

In FIG. 2, individual process blocks of the overall coding CC are broken down. The user data LD are first subjected to a coding C_LD. In the context of this coding, for which convolutional codes are used in particular, LD redundancy is added to the useful data, which means that the transmitted data TD can be recovered more reliably in the event of transmission errors on the receiver side. The code used in the encoding is characterized by its code rate R = K / N, where K denotes the number of data or message bits to be transmitted and N the number of bits present after encoding. The lower the code rate, the more efficient the code is. A problem associated with coding, however, is that the data rate is reduced by the factor R and the information is therefore packed and thus a partial loss can be more problematic than with less densely packed information.

In order to adapt the data rate of the coded data stream to the respective possible transmission rate, a rate adjustment (rate matching) RM_LD is carried out in the transmitter, bits being either removed from the data stream or repeated in the data stream according to a certain pattern. Removing bits is called "puncturing" and repeating bits is called "repeating".

In an analogous manner, the identification data ID are first subjected to a coding C_ID and then to a rate adjustment RM_ID. Subsequently, identification data and user data are linked to one another in a linking process L, as a result of which the data TD to be transmitted are formed.

The procedure shown in FIG. 2 is known in principle, but the state of the art and the invention differ in the implementation of the rate adjustment for user data LD and identification data ID.

FIG. 3 shows the implementation of the overall coding of the HS-SCCH part 1 in accordance with the current specification UMTS standard (FDD, Release 5). The user data LD are formed by the channel information bits x _CC s, ι, x _{Ccε, 2 /} • • • _< ■ x _ccs7 . The channel information bits are referred to in specialist circles as "channelization code set bits". Furthermore, the modulation scheme bit x _mSιl , which is also referred to as "modulation scheme bit", flows into the user data. This user data is encoded with a rate 1/3 convolutional encoder in accordance with the 1999 standard (Release 99). Eight tail bits attached before this coding at the end of the bit block enable easier and more reliable decoding on the receiver side. The Multiplexer MUX enables an alternate polling of channel information bits X _ccs and the modulation scheme bit X _^ . The entirety of the data present after the multiplexer is referred to as X. Thus there are 16 bits on the input side of the encoder or encoder or before the coding process C_LD, while on the output side of the encoder or after the coding process C_LD there are 48 bits due to the rate 1/3. This coded bit block is referred to as Z _x . The index 1 means here that it is a size that concerns part 1 of the HS-SCCH. Part 1 of this control channel contains data which the receiver must decode immediately in order to process incoming data on the HS-DSCH (HS Downlink Shared Channel) accordingly. Accordingly, the availability of the data of Part 2 (Part 2) is less time-critical.

On the physical channel, ie the actual transmission channel, only 40 bits are available for the part 1 of the control channel HS-SCCH for the transmission.

In order to get from 48 bits to the 40 bits that can be physically transmitted in part 1, a rate matching takes place according to the following rate matching pattern (pattern 1): from the bit block or the sequence Z _x , which results from coding process C_LD the bits are punctured at positions 1, 2, 4, 8, 4 ^' 2, 45, 47, 48. If a notation with a second index j is used, which identifies the bit position and runs from 1 to 48 in the case shown, then the bits to be punctured can be specified as Z _{1 (1} , Z _{12 /} Z _{1 (4f} Z _lfBι Z _lι42 _ Z _145> Z _lj47> Z _148. As before, the first index indicates that it is part 1 of the HS-SCCH. In this notation, the sequence R _{1 (1} R _lι2 , ••• Rι _{/ 40} before.

The control channel HS-SCCH is intercepted by several mobile stations or mobile radio terminals (UE: User Equipment). To identify the respective mobile station UE, or so that this mobile station can decode part 1 and also so that a mobile station that is not addressed recognizes this, the user data, consisting of channel information data and the modulation scheme, are identified by the identification data or a specific mask dependent on the identification number of the mobile station.

In the case shown here, based on the 16-bit identification number of the mobile station (UE ID) using rate ^ coding according to the 1999 standard (Release 99), a so-called scrambling code for the identification number of the mobile station is generated, with which the user data be masked. The identification number of the mobile station UE ID is assigned to the mobile station in the respective cell by the respective base station. Scrambling is therefore a "personalization" of the information. This is done using the so-called "scrambling codes", with which the signal is modified in order to separate or separate signals for individual terminals or base stations.

To generate the scrambling code, the 16 bits of the identification number of the MoBilstation UE ID X _{ue; 1} , ... X _{ue> 16} and the attached eight tail bits according to the 1999 standard (Release 99) with the Rate% convolutional encoder (C_ID) coded. This also results in (16 + 8) x2 = 48 bits of a sequence B at the output of the convolutional encoder. In order to achieve a length of 40 bits here, the rate adjustment algorithm from the 1999 standard (Release 99) for puncturing is used for the rate adjustment RM_ID (RM_ID), in which of the sequence B, consisting of bits b _x _ b _2p ... b ₄₈ , the index indicating the bit position, bits b _{1 #} b ₇ , b ₁₃ , b _19> ^ ₂ 5 'k ₃₁ , b ₃₇ , b _{43 are} punctured. The sequence C thus formed, consisting of bits c _x , c ₂ , ... c ₄₀ , results in the necessary reduction from 48 bits to 40 bits. Different rate adjustment patterns are therefore used for the branch of the user data LD and the branch of the identification data ID for their respective rate adjustment RM_LD or RM_ID. The reasons are as follows:

- In general, there is not the same number of bits in the branch with the identification data ID or in the branch with the user data LD after the coding level. This can be due both to the number of output bits, that is to say the number of bits of the identification number of the mobile station or channel information or modulation information bits, and to the rate of the coding. A different rate adjustment is then absolutely necessary. - The coding in the coding stage C _LD or Cι _D inter alia serves to interleave the bits, so that the original bit sequence Xi or X _0E can be restored on the receiver side even in the case of poor transmission conditions. A good entanglement in this sense naturally looks different for different or differently structured input data X _ue or Xi (= X _C s or X _ms ), in particular even if different coding rates are used. As a result, individual bits have different importance after the coding stage. This different importance depends on how many input bits of the coding stage an output bit of the coding stage is associated with. The more input bits flow into the output bit, the more important the output bit is to restore the original data. In the case of a rate adjustment pattern, in the case of puncturing data, bits are preferably punctured which are of less importance in this sense. In other words, with different coding, for example with different convolutional encoders and subsequent rate adaptation, different rate adaptation patterns lead to different distance properties with regard to the Hamming distances of the resulting code sequences or code words and thus determine the performance of the coding.

The use of different rate adjustment patterns and the associated computing and storage effort pose only a minor problem in the base station, since the appropriate hardware is available there to also handle computing processes with high complexity. However, this does not apply to the receiving mobile station.

As already mentioned above, the aim of the invention is to make the overall coding, in particular the rate adjustment, less complex than is currently the case, that is to say the specification according to Release 5.

A main aspect of the invention is to perform the rate adjustment for identification data ID and user data LD according to a common rate adjustment pattern SR. In principle, two solutions are conceivable for this: i) The use of a common rate adjustment pattern, but separate implementation of the rate adjustment for user data LD and identification data ID. ii) Using a common rate adjustment pattern and performing the rate adjustment together.

Figure 4

FIG. 4 now shows a process sequence designed according to solution ii), likewise for the example of the control channel HS-SCCH. In this case, the identification data ID, here referred to as identification bit sequence X _ue , and the channel information data, here X _ccs and X _,. _s already after the respective coding C_LD or C_ID, linked together and then subjected to a common rate adjustment. The linking is done, for example, by an XOR function if the two values that a bit can take are defined with 0 and 1. If the values - 1 and 1 are assumed, the link can be made by multiplication. However, other bitwise links can also be used.

In FIG. 4, analogously to FIG. 3, the data resulting from the coding process are designated Z ₁ . In deviation from FIG. 3, the bit block or the bit sequence or the sequence R _x denotes the data before the common rate adjustment, but after the link.

The following advantages are achieved by proceeding according to solutions i) or ii): Since the rate adjustment is carried out with only one rate adjustment pattern, the decoding in the receiver, for example the mobile station UE, is correspondingly easier. Less complexity is already achieved if the rate adjustment for identification data ID and user data LD is carried out separately according to the same pattern (solution i).

If the rate adjustment is summarized according to solution ii), this means a further simplification.

Different rate adjustment patterns

Another aspect of the invention is the design of a rate adjustment pattern, which is approximately equally suitable as a common scheme for user data LD and identification data ID. An important aspect here is, among other things, that the Hamming distance or the Hamming distance after the link is as large as possible, for Example of the linked data in the event of a fault

To be able to reconstruct the transmission as well as possible. “Hamming distance” is understood to mean the number of bits by which two code words of the same length differ from one another. This is used for error detection by comparing received data units with valid characters. Any correction is made according to the probability principle. To keep the information content of the user data LD as good as possible, a large Hamming

Distance desirable. However, these and other criteria, such as the signal-to-noise ratio, are not necessarily independent of one another, which can lead, among other things, to the fact that the attempt to find an "optimized" rate adjustment pattern leads to several different rate adjustment patterns, which in mathematical way of speaking could possibly also be referred to as secondary minima of the optimization problem. Among others, some variants have special advantages for the common rate adjustment pattern: a) Use of the current puncturing algorithm (Release99):

In the sequence r _{lι lr} r _{lr 2} , ..., rι, ₄₈ the bits n, ι, r ₁₇ , rι, ι ₃ , rι, i9, rι, ₂₅ , r _x , ₃ ι, rι, ₃₇ , n, ₄₃ dotted, and you get the sequence sι, ι, Sι, ₂ ... sι, ₄₀ . This has the advantage that only minor adjustments to the currently used system have to be made.

This puncturing pattern, like other rate adjustment patterns, can be shifted, for example by means of an offset 0 <= k <6. This means that for the case of the standard from 1999 (Release99), the bits r _{1 + k} , r _{7 + k} , rι _{3 + k} , rig + k, r _{25 + k} , r _{31 + k} , r _{37 + k} , r _{43 + k} . The size of the shelf k is determined by the distance of the last to bit position to the last bit position. If the last bit to be punctured is at position 42, the puncturing pattern can be shifted back by a maximum of 6 positions with a total length of 48 bits.

b) The puncturing pattern used is the puncturing pattern "Pattern 1" [1] optimized for the user data of Part 1 of the HS-SCCH:

From the sequence rl, l, rl, 2, ..., rl, 48, the bits rl, l, rl, 2, rl, 4, rl, 8, rl, 42, rl, 45, rl, 47, rl , 48 punctured, and thus the sequence sl, l, sl, 2 ... sl, 40 is obtained. This variant is particularly advantageous since it optimally encodes the HS-SCCH data and also unites the sequences for masking the data in the code space have a greater distance between them, i.e. a greater Hamming distance than with puncturing according to the Release99 puncturing algorithm.

c) A new puncturing pattern, which simultaneously optimizes the coding properties of the data of Part 1 of the HS-SCCH and the recognition possibilities of the masking with the UE ID, can be solved by an optimization, whereby the secondary conditions through the data structure in the identification data branch and are specified in the user data branch. With such an optimization, a target function is first formed, in which the variable to be optimized is mapped. Taking into account the secondary conditions caused by the system, an extremum is then sought for this target function, for example the target function is minimized. The optimization function serves as the target function to be minimized

Total error probability P (E) of detection and the correct decoding of the user data of the user data branch:

P {E) = P (not detected) + R (not decoded). (1) Detection is understood here to mean that the decision based on the scrambling with identification data that the packet is intended for the mobile station is made correctly. that, that is. This objective function can be calculated or approximated under system-related boundary conditions, such as the number of channels to be observed, or boundary conditions of the control channel allocation. Based on the usual detection criteria and decoding algorithms specifically proposed for this, a rate adaptation pattern is then sought which minimizes the target function.

This is explained below using the example of an application of Part 1 of the HS-SCCH. In the special case of the control channel HS-SCCH, much less energy is required for decoding than for detection, so that P (E) can be approximated by P (not detected) in the above formula (1). P (not detected) settles due to the above Boundary conditions composed of three linear terms, terms of higher order are neglected in the following due to their negligible influence:

P {E) «P (undetected) = P {FÄ) _ _X + P (MD) + P (FA), (2) where P (FA) is the overall false alarm probability and

P (MD) denotes the non-detection probability. The Tiefindex -1 indicates the influence of the previous time step. For the special case of four control channels, the procedure according to the "consecutive scheduling rule" and a random decision for a control channel, if more than one can meet the detection criteria, Eq. (2) can be further resolved (the above-mentioned requirements will be explained later in connection) ):

P {E) «P (not detected) = 3 • P {fa) _ _λ + P (MD) + - P {fa). (3) The terms are made up of:

1) The triple false alarm probability from the previous time period: False alarm probability is the case when the mobile station incorrectly assumes that the data packet is intended for it, although it is actually intended for another or no mobile station. This contribution is made as follows: If the mobile station in the previous period (also called TTI, Transmission Time Interval) incorrectly selected a different HS-SCCH than the HS-SCCH on which the information is currently being transmitted from the base station due to a false alarm the mobile station can use all resources for the current time slot that are required for "listening" to the other, unselected HS-SCCH's for other tasks. This is because the standard provides for the case that that the same HS-SCCH is always used in successive time steps from a base station to a mobile station. This is also referred to as the so-called "consecutive scheduling rule ^λ . This leads to a saving of resources in the mo- bilstation, but has the side effect that an error in the previous time step also has consequences for the n has the current time step.

2) The non-detection probability in the current time period. This simply means that the mobile station does not recognize that the message is intended for it.

3) 1.5 times the false alarm probability from the current time interval. This contribution comes about if the mobile station is due in the current time step of a false alarm incorrectly selects two HS-SCCHs, which means that due to e.g. B any transmission errors seem to mask both HS-SCCH's with the same UE-ID. While the information is currently being transmitted on one of the HS-SCCHs, information for another mobile station is being transmitted over the other HS-SCCH. The mobile station must then select one from these two candidates. For the sake of simplicity, it is assumed here that one of the two candidates is selected at random. Since there are up to three competing HS-SCCH channels in addition to the correct channel, the factor is 1.5 = 3/2. If other criteria can be used to distinguish the candidates, this factor may possibly be reduced.

The following are examples of criteria that can be used to decide which of several potential HS-SCCH channels is the correct one:

- The channel whose decoding probability is the highest or the decoding metric is the smallest.

- If necessary, not all possible useful data are useful or applicable for a given network configuration or mobile station. For example, the channel information bits can indicate a number of the codes signaled for the subsequent data transmission which is greater than the number of codes which the mobile station supports or which are currently configured in the base station. Such an assignment indicates that the channel information bits were received incorrectly and the corresponding HS-SCCH can consequently be discarded. The same applies to the modulation scheme bit if it indicates an unsupported modulation type. Based on this objective function, the following puncturing pattern appears to be particularly suitable. For better readability, the positions at which the bit is punctured are marked with 0, positions at which there is no puncturing are marked with 1: 011101101011111111111111111111111111111100011101

Simplification of the decoding on the receiver side As already explained, the proposed simplification of the rate adjustment, in particular on the receiver side, for example in a mobile station, is a great advantage due to the lower complexity of the decoding. Differences in the decoding, how it is currently carried out and how it is done of the invention are set out below.

Figure 5

FIG. 5 shows an exemplary implementation in the receiver device, as is required in the current specification (release 99). The transmitted data TD is received via the air interface AI (Air Interface). These transmitted data TD are demodulated in the demodulator demodulator. After demodulation, this data is fed directly to a bit error count unit. On the other hand, this data is linked to the masking data, for example by an XOR connection or a multiplication. The masking data are generated in the mobile station from the identification number of the mobile station UE ID, which is encoded and then subjected to a rate adjustment (RM2). This is followed by the link with the demodulated, transmitted data TD. The rate adjustment RM2 of the masking data is required to match the bit lengths of the masking data. Adaptation data to the bit length of the received data TD.

The rate adjustment RM1 ^{"1 is} undone for the linked signal before decoding Dec. This data is decoded and re-encoded to check whether the information was intended for the respective receiving mobile station and subjected to a further rate adjustment RM1 before it is repeated The result of this new combination also flows into the bit error counting unit, in which the detection of the errors is based on a processing of 40 bits, that is as many bits as via the air interface AI per HS-SCCH Frame (HS-SCCH subframe), which consists of three so-called slots or time slots, are transmitted.

Figure 6

FIG. 6 shows two exemplary implementations that can be used with a rate adjustment carried out according to the invention.

In the figure above (DEC_40), bit error detection in the bit error count unit Bit Error Count is also based on 40 bits. Due to the same rate adjustment pattern used in the transmitter for identification data ID and user data LD, the rate adjustment only takes place together with the transmitted data TD received via the air interface, immediately before the bit error counting unit Bit Error Count. In this way, compared to the prior art, a rate adjustment is saved, namely that, as can be seen from FIG. 5, of the masking data before it is linked to the received data.

The following steps are shown in detail in FIG. 6 in the example above (dec 40):

The transmitted data TD are received via the air interface AI. Become a demod after a demodulation process the data is divided and flows on the one hand in a first

Branch directly into a bit error count unit. In the other branch, the rate adjustment RM ^{"1 is} undone and subsequently linked to the masking data, which are generated by coding the mobile station identification number. In contrast to the implementation shown in FIG. 5, no rate adjustment of the masking data is required because the rate adjustment the transferred data was undone before the link was created.

The linked data is decoded in a decoding process Dec. On the one hand, the required data is then available, on the other hand, this data is subjected to coding in a further coding process and linked again with the masking data. This is done for the purpose of

Error detection in the bit error counting unit Bit Error Count, into which this data flows after the reconnection and a rate adjustment process RM. In summary, a rate adjustment is saved compared to the implementation shown in FIG. 5. This is made possible by using a common rate adjustment pattern for user data LD and masking data ID in the transmitter. If different rate adjustment patterns were used, the common rate adjustment in the rate adjustment unit RM in FIG. 6 in front of the bit error counter would not lead to the original signal, for example.

Even clearer improvements, namely the saving of two rate adjustments, are achieved in the implementation shown in the lower image of FIG. 6.

In the image below (dec_48), the detection of bit errors is based on 48 bits. In this case, you only have to undo the rate adjustment. A further rate adjustment is no longer required.

The following steps are carried out in detail in the lower representation in FIG. 6: The transmitted data TD are received via the air interface AI. Subsequently, the rate adjustment RM ^{"1 is} canceled, which is necessary because, on the one hand, the data is passed in a first branch directly to the bit error counting unit Bit Error Count, in which the bit error detection takes place on the basis of 48 bits.

On the other hand, in a second branch, this data is linked to the masking data generated in the mobile station from the mobile station identification number UE ID. After the linking and a subsequent decoding Dec, the required data are then available. Analogous to the example above, the data is subjected to a coding coding for the subsequent error detection and then linked to the masking data. In contrast to the example above in FIG. 6, a rate adjustment after the linking is not necessary since there are 48 bits, on the basis of which the error detection also takes place.

Thus, no rate adjustment is required in this implementation.

In general, the detection can be supported by further criteria or based on entirely different criteria. These criteria are explained below: Since the probability of false alarms is also an essential factor for the detection of HS-SCCH, Part 1, a combination, for example, that is to say a logical AND combination of the bit error counting unit, can be carried out using a so-called status criterion. This state criterion assumes that data is only available if the best decoding metric occurs in a decoding state, which element is a predefined state set. Decoding metric is a measure of the probability of decoding.

A so-called Viterbi decoder is often used to determine the best decoding metric from a data set. The following explanations therefore relate to one Viterbi decoders, but can be adapted accordingly to others

Apply or generalize decoding algorithms. In general, a Viterbi decoder compares the received data record with all possible data records, that is to say the predefined set of states, and then selects the data sequence which has the greatest probability of agreeing with the data sequence actually transmitted. The Viterbi decoder or decoder is widely used for decoding convolutionally coded data. It calculates a set of states for each bit position, the number of which depends on the so-called influence length Le. Each state then corresponds to a possible value of the last Le bits under consideration. A metric is calculated for each state, which represents a measure of the probability that this state or the associated bit sequence is the one that is actually transmitted.

As already described, so-called tail bits, which generally have the value 0, are inserted in the coding process after the actual data. As a result, at the end of the decoding, assuming ideal decoding, the zero state, in which the last Le bits are assumed to be 0, should be the most likely. In the simplest case, the state set for the final state, in which it is assumed that the data packet is actually intended for the mobile station, only contains the zero state. Then the criterion is also called a zero state criterion. The mobile station then only assumes that the data packet is intended for it if the zero state actually appears as that at the end of the decoding

Most likely turns out. Otherwise the mobile station assumes that the packet is not intended for it. This criterion serves in particular to reduce the probability of false alarms. In other words, the mobile station compares a priori information with this criterion, namely the known fact that the tail or end bits have the value 0, with the value of these bits, which is determined by a decoding tion without using this knowledge. A match is then seen as an indication that the entire coding is consistent and therefore correct.

However, the use of the zero state criterion leads to a higher probability of non-detection than when using the rate adaptation pattern of the data, which punctuates many bits particularly in the area of the _s bits at the end, which belong to the tail bits a rate adjustment pattern that punctures less or none of the last bits in the data block. In order to avoid this deterioration, there are basically two possibilities: 1. The zero state criterion is expanded to a best state criterion and the presence of data is assumed as soon as the best decoding metric occurs in a state which is in a predetermined set of states M is included, for example, in addition to state 0 (that is ... 00), states 1 (that is ... 01) and 2 (that is

... 10) of the total of 256 possible states of the Viterbi decoder with influence length 9. However, this procedure can lead to a slightly higher probability of false alarms. For the individual case, it must be weighed up whether a slightly increased probability of false alarms or a slightly higher probability of non-detection is more favorable. 2. A rate adjustment pattern specially optimized for the two above-mentioned detection criteria is used. Such an optimized rate adjustment pattern is characterized by the fact that it punctures as little as possible in the area of the end or tail bits. In extreme cases, the entire tail area is not punctured. With 8 tail bits and a coding rate R = 1/3 coding, this results in 3 * 8 = 24 bits, which are not punctured. This tail area is characterized by the fact that it depends on fewer and fewer data bits, particularly with the last bits. As a result, the bits in the tail area are increasingly correlated. On the one hand then these bits are less helpful for decoding the data by a decoder. On the other hand, these bits are then better suited for the detection of differences in a mask, ie the identification data. Bits that do not contain any information at all would be ideal for this, but on the other hand this would represent a waste of bits in the decoding.

A good compromise is not to puncture the bits in the tail area and to perform the puncturing completely outside of this area.

For this second option, simulations were created that were used to search for rate adjustment patterns by optimizing the target function. Since the search for suitable rate adjustment patterns would be too complex, taking into account all possibilities, the search was limited to selected areas taking into account the constraints shown above. Shortening 6 bits f ASλ from 48 bits would already result in possibilities.

The following puncturing patterns obtained by the above search have been found to be particularly suitable. For better readability, positions where a bit is shortened are marked with 0, positions where no bit is shortened with 1.

Rate adjustment pattern A:

000110111011010101111111111111111111111111111111 Rate Adjustment Pattern B:

000111111011010101101111111111111111111111111111

In both samples it can be clearly seen that no cuts (puncturing) are made in the end area.

Under other system requirements or channel properties, puncturing in the tail area or puncturing affects only less strongly in the beginning area of the tail area

Detection probability.

For the search for such rate adjustment patterns, the secondary conditions are then modified or the selected areas are expanded. For example, in the case of the HS-SCCH, the computer-assisted search could also be extended to the first 28 bits, that is to say partly also to an area in which the tail bits already have an influence, in order to find a good puncturing pattern there.

Other uses

The common use of rate adjustment patterns has been explained in particular for the HS-SCCH, but is not limited to this. Masking of the useful data is also used in other control channels, as a result of which the invention can be used in various configurations. In principle, there are further applications for all channels in which different data streams are linked together for transmission and a rate adjustment is required.

Finally, with reference to the drawings, it should also be mentioned that the term, pattern is also used for rate adjustment patterns; Physical channel mapping ^x also means the mapping to the physical data channel. XOR is used as a shorthand for an exclusive OR combination.

Abbreviations

HSDPA High Speed Downlink Packet Access HS-DSCH High Speed Downlink Shared Channel (data)

HS-SCCH High Speed Shared Control Channel (control information, signaling)

UE User Equipment (designation for a UMTS mobile station) UE ID Identification number of the mobile station Sources: [1] Rl-02-0605, "Coding and Rate Matching for HS-SCCH", TSG RAN WG1 Meeting # 25, Paris, April 9-12, 2002.

The cited document is a document from 3GPP (Third Generation Gartnership Project), address: ETSI, Mobile Competency Center, 650, route des Lucioles, 06921 Sophia-Antipolis Cedex and is cited here in the format used by this organization.

Claims

claims

1. method for rate adjustment to a specified data rate,

in which the data, the data rate of which is adjusted, is composed of at least two coded data streams, and

- In which the rate adaptation of the data rates of the at least two data streams to the specified data rate takes place using a rate adaptation pattern, via which the scheme for repeating or puncturing individual data sequences from the data streams is established and which is the same for the at least two data streams.

2. The method as claimed in claim 1, in which the data are formed by linking the at least two data streams, and the rate adjustment is carried out before or after the linkage of the data streams.

3. The method of claim 2, wherein the link is a bitwise link.

, Method according to one of the preceding claims, in which the coding of at least one data stream takes place by means of convolutional coding.

5. The method as claimed in one of the preceding claims, in which the fixed data rate is predetermined by the data rate of a physical channel used by at least a first transmitting / receiving unit and a second transmitting / receiving unit in a communication system, and - In which the at least two data streams on the one hand from user data data (LD) and on the other hand identification data

(ID) for identifying the second communication device are formed, - in which the user data (LD) and the identification data (ID) are coded separately from one another,

- and the coding (C_LD, C_ID) takes place in such a way that the same bit rate results for the useful data LD and identification data ID after the coding process, and - the rate adaptation to the bit rate defined for the physical channel using the for useful data (LD) and identification data (ID) of the same rate adjustment pattern.

6. The method as claimed in one of the preceding claims, in which the coding process supplies a bit sequence from bits 1 to n in a fixed time window, as a result of which the rate is fixed, and the rate adjustment is carried out by means of a rate adjustment pattern by which individual bits from this bit sequence be punctured. ^{* ~}

7. The method according to any one of the preceding claims, wherein the physical channel is the High Speed Shared Contro.1 Channel (HS-SCCH).

8. The method according to any one of the preceding claims, wherein the identification data is the identification number of a transmitting / receiving unit.

9. The method of claim 7 and 8, wherein the rate adjustment is carried out with a rate adjustment pattern by which in the bit sequence consisting of n = 48 bits, the bits at positions 1, 2, 4, 8, 42, 45, 47, 48 are punctured.

10. The method according to claim 7 and 8, in which in the bit sequence consisting of n = 48 bits, the bits at positions 1, 7, 13, 19, 25, 31, 37, 43 are punctured.

11. The method according to claim 10, wherein the position of the bits to be punctured is shifted by an integer k, where 0 <k <= 5 applies.

12. The method according to claim 4 or 7 and 8, in which a fixed bit sequence is added to the user data (LD) before coding, the coding for user data (LD) is carried out and a link before or after the rate adjustment with the coded identification data (ID) takes place, the rate adjustment using a rate adjustment pattern, by which only or preferably bits are punctured which do not carry information about bits in these defined bit sequences.