WO2004084476A1

WO2004084476A1 - A communications apparatus and method

Info

Publication number: WO2004084476A1
Application number: PCT/GB2004/001134
Authority: WO
Inventors: Nicholas Berryman; Mark Dominic Horgan; Bernhard Raaf
Original assignee: Siemens Aktiengesellschaft
Priority date: 2003-03-21
Filing date: 2004-03-17
Publication date: 2004-09-30

Abstract

An encoder comprises at least one recursive systematic convolutional encoder for, in use, producing at least parity stream and at least one rate modifier for modifying a rate of the at least one parity by applying at least one of puncturing, or puncturing and repetition. Regularities of the positions of the punctured bits within a repetition period of the encoder are avoided by, in the case of the application of puncturing, making the puncturing irregular with respect to the repetition period or, in the case of the application of puncturing and repetition, modifying repetition and puncturing rates in a complementary manner to avoid regularities of the positions of the punctured bits within the repetition period of the encoder.

Description

A COMMUNICATIONS APPARATUS AND METHOD

This invention relates to a digital communications apparatus and method particularly, but not exclusively, for use in a digital cellular communications network. In digital cellular communications networks, data is transmitted in digital, encoded form. The data is encoded to permit data recovery techniques to be used to engender some robustness to data loss due to the effects of interference and fading. There exists a first encoding called source encoding and a second encoding called channel encoding. One form of encoding data to increase robustness against errors involves the use of encoders called turbo encoders. Turbo encoders provide a very high error recovering capability in a noisy transmission channel. These comprise a structure of recursive systematic convolutional encoders (RSC) which produce two encoded outputs (known as "parity" outputs) and an uncoded output (original data, known as the "systematic" output). The first RSC encoder operates directly on the input data to generate the systematic stream and one ofthe parity streams. The second RSC encoder operates on an interleaved version ofthe input data to generate the second parity stream.

The parity data generated in each RSC encoder is an algebraic sum ofthe current input bit and previous input bits to that encoder. The parity data can be termed redundant in that it is additional data solely for the purpose of recovering the systematic data within the turbo decoder. The communication channel will have a particular capacity or data rate and, in order to match the data stream to the channel data rate, a process of rate matching is then undertaken. In the case that the output ofthe turbo encoder produces more combined systematic and parity bits than the channel capacity is able to transmit, the rate matching is carried out by process called "puncturing" in which bits are eliminated from the encoded parity streams in accordance with a predetermined pattern before the streams are assembled with the original signal and passed to the transmitter section. Fig. 1 shows a conventional turbo coder in which an input data stream Xt is input to the coder from a data source not shown. It will be shown that it results in three output streams: Xt, the same as the input stream, and two so-called parity encoded streams which are subjected to a puncturing process to create punctured streams Zk and Z'k. The streams are collected by a bit collector and passed to a transmitter section. It will be understood that the turbo encoder produces a combined output of three times the input data. Of this output, 1/3 is the input data Xt and 2/3 is parity data. The parity data introduces the redundancy to enable data recovery in the event of interference. To reduce the total amount of data to be transmitted to match the capacity on the channel, a puncturing process is carried out on the parity data.

The present invention arose by an inventive appreciation ofthe performance of such an encoder by the inventors in which the reason for performance drop off at particular encoding rates was observed. ("Encoding rate" refers to the ratio of input bits to the turbo encoder to bits actually transmitted across the air, following encoding & puncturing).

According to a first aspect ofthe present invention an encoder comprises at least one recursive systematic convolutional encoder for, in use, producing at least one parity stream; at least one rate modifier for modifying a rate ofthe at least one parity stream by applying at least one of puncturing, or puncturing and repetition, wherein regularities ofthe positions ofthe punctured bits within a repetition period ofthe encoder are avoided by, in the case ofthe application of puncturing, making the puncturing irregular with respect to the repetition period or, in the case ofthe application of puncturing and repetition, modifying repetition and puncturing rates in a complementary manner to avoid regularities ofthe positions ofthe punctured bits within the repetition period ofthe encoder.

According to a second aspect ofthe present invention a method of puncturing data provided by an encoder which encoder provides at least one parity stream comprises the steps of modifying a rate of the at least one parity stream by applying at least one of puncturing, or puncturing and repetition, wherein regularities ofthe positions ofthe punctured bits within a repetition period ofthe encoder are avoided by, in the case ofthe application of puncturing, making the puncturing irregular with respect to the repetition period or, in the case ofthe application of puncturing and repetition, modifying repetition and puncturing rates in a complementary manner to avoid regularities ofthe positions ofthe punctured bits within the repetition period of the encoder. The invention may be applied to one or more ofthe parity streams. One way in which the parity-encoded streams may be punctured is to use different puncturing rates for the streams, so that one parity stream will be longer than the other. Preferably, the puncturing rates are chosen such that overall rate is maintained at a level appropriate for the channel. One parity stream may be punctured at relatively high rate whilst the other is punctured at a relatively low rate. The difference in the rates may vary from a small difference to a relatively large difference and it may do this in a dynamic way or be predetermined. The rates may be periodically varied such that the higher rate is used on the stream that formally had the lower rate and vice versa. Preferably, the encoder comprises a rate modifier for modifying the at least one parity stream.

The parity streams may be considered to comprise a number of notional segments and the same puncturing rate may be applied to more than one ofthe segments or indeed all. Alternatively, different puncturing rates may be applied to different segments. In other words, a first rate may be applied to a first segment and a second rate applied to a second segment within the same parity stream.

Typically, at least one puncturing rate is applied per segment ofthe at least one parity stream.

Preferably, two or more puncturing rates are applied per segment ofthe at least one parity stream.

Preferably, one puncturing rate is applied to at least a first segment ofthe at least one parity stream and at least two puncturing rates are applied to at least a second segment ofthe at least one parity stream.

The rates are preferably chosen to avoid certain rates in particular problematic rates.

Problematic code rates result from puncturing a set number of bits within the bit stream such that if this puncturing were done in a uniform manner, i.e. puncture m bits out of every n, this would be synchronous with the inherent recursive period associated with the RSC encoder. The inherent recursive period ofthe RSC encoder is determined by the feedback structure ofthe RSC encoder, but can maximised to 2 ^Λq-1, where q is the number of memory elements within the RSC encoder. For example, a RSC encoder with 3 memory elements, with a feedback structure that maximises the repeat length, would have an inherent recursive period of 2^A3-1 = 7. We will define the inherent recursive period of the RSC as L.

Therefore if the puncturing rate to be applied is n/m, where m is equal to L (or a integer multiple of L), both n and m are integers, then this is said to be synchronous. Degradation in performance occurs not only when the puncturing pattern is generated using uniform puncturing with a puncturing rate synchronous with the inherent recursive period ofthe RSC encoder, but when the two rates are close, not necessarily identical.

Another alternative arrangement would maintain a constant puncturing rate on a first ofthe streams, at a non-problematic rate, and to use a varying puncturing on a second ofthe streams. Preferably, the varying rate avoids the particular problematic rates. The rate is also preferably varied to maintain the overall rate appropriate for the channel.

The puncturing rate may also be varied in a pseudo-random way on one or both of the parity streams .

Preferably, the puncturing is carried out in such a manner that the positions of the retained bits are pseudo-random.

In one example, in the case ofthe application of puncturing and repetition, the repetition rate is greater than the puncturing rate. Alternatively, the puncturing rate is greater than the repetition rate.

In accordance with a third aspect ofthe present invention, communications apparatus includes an encoder according to the first aspect.

A specific embodiment ofthe invention will now be described with reference to the accompanying drawing in which: Figure 1 is a prior art figure showing a turbo encoder;

Figure 2 shows in block diagram schematic form a communications apparatus in accordance with the invention;

Figure 3 shows in greater detail a transmit path ofthe apparatus of figure 2 and in particular more detail of a turbo encoder incorporated in the path; Figures 4 to 7 are explanatory figures; and,

Figure 8 illustrates HS-DSCH hybrid ARQ functionality. As is shown in Fig. 2, communications apparatus 1 is part of a 3^rd generation mobile communications system which includes a plurality of node Bs 2, one of which is shown. The node Bs connect via a node controller 3 to the rest ofthe network 4 and to the communications apparatus 1 by means of up and downlinks 5. The up and downlinks 5 include signalling up and down links as well as traffic channels. The protocol used for the links is in accordance with the 3^rd generation standard and is thus conventional in nature for cellular digital networks.

The communications apparatus 1 is hand-held user equipment although it may be mobile or stationary equipment. It includes an antenna 6 operably coupled to a receiver path and transmission path. The receiver path includes a receiver section 7, and a demodulator 8 of known form. The demodulator 8 demodulates incoming detected signals and passes the demodulated output to a channel decoder 9. The channel decoder 9 removes the channel coding and produces a decoded bit stream and passes this to a bit stream splitter 10. This produces an input stream and two parity streams which are passed to a turbo decoder 11. The turbo decoder 11 decodes the input stream using the parity streams to perform error correction and passes the decoded output to an output device 12 which includes a liquid crystal display and a speaker (although other types of devices for outputting information to a user could be used).

The transmission path will now be described starting from the input of information to be transmitted. Vision or sound is input by means ofthe input device 13. This includes a microphone, keypad and camera. The input device 13 creates a digital stream of data and passes this to a turbo encoder 14. The turbo encoder produces three streams, a first being the original input data stream and the other two streams being parity streams. The streams are formed into a single stream by the bit collector 15 and passed to a channel encoder 16. The channel encoder 16 passes the encoded bit stream to a modulator 17. The modulated output is passed to a transmitter section 18 and thence to the antenna 6 and via the links 5 to the node B 2.

Overall control ofthe various blocks is maintained by a controller 19 operating in accordance with a set of instructions held in memory 20. It will be appreciated that the controller 19 as well as much ofthe other blocks can be implemented on a microprocessor or a number of microprocessors.

Fig.3 shows the transmission path with greater detail ofthe turbo encoder 14 which includes an encoder 21, an input connected to the input device 13 and three output lines 23, 24 and 25. The first ofthe output lines line 23 is directly connected to a bit collector 15. The second line 24 is connected to a puncturer 27 and the third line 25 is connected to a further puncturer 28. The puncturers 27 and 28 are connected to, and controlled by, a puncturer controller 29 the function of which will be explained later. The output ofthe puncturers 27 and 28 are coupled to the bit collector 15 and the collected streams passed to the channel encoder 16.

In a conventional arrangement the puncturers 27 and 28 would be arranged to puncture the encoded streams on the lines 24 and 25 at the same or nearly the same near regular interval. A problem identified by the inventors with this, is that, for particular puncturing rates the puncturing is near synchronous with the inherent recursive period ofthe turbo coder's constituent RCS encoders, a significant energy increase in the required signal is then needed to overcome reduced performance at specific coding rates. Fig. 4 shows a graph ofthe degradation in performance of an encoder at specific rates. In this figure the y-axis represents the energy required in the signal to be decoded by a receiver for there to be a block error rate of 1 %. The x-axis relates to coding rates. It will be seen that the plot describes a generally upwardly curved profile which shows that the required energy increases as the coding rate increases to 1. However, the curve includes a number of significant peaks at particular coding rates which do not follow the general trend. The peaks are enclosed by circles 40 to 45. At these marked departures from the general trend ofthe curve it is necessary if the quality is to be maintained for the transmit power to be increased or the bit loss error rate will increase. These rates are termed as the problematic rates.

The present invention arose from an inventive realisation of this problem and the formulation of an inventive approach to at least alleviate the problem.

In the specific embodiment, the ways in which the puncturers 27 and 28 are controlled by the puncturer controller 29 are such that the identified problematic rates are avoided. It has been found that the problematic puncturing rates occur where n/m, where m is equal to, or a multiple of, L, the repetition period ofthe recursive systematic convolution. Assuming that the structure ofthe RSC encoders is chosen well, the repetition period is 2^Λq-l where q is the number of memory elements in each ofthe constituent encoders of encoder 21 and m and n are both integer values. The puncturer controller 29 is programmed to provide the puncturing rate in a number of ways. Each way is described as a separate embodiment although in some applications it may be possible to produce puncturing rates in the described ways in a single puncturer controller. In other words the approaches exemplified by the embodiments may be combined into a further single embodiment.

In a first embodiment ofthe invention and where the RSC encoder has a period of seven, the puncturing pattern provided by the puncturer controller 29 is as shown in Fig.5. The embodiment uses two puncturing rates. The first rate is the higher rate at 7/8 which means that 7 out of eight bits are punctured and a lower rate of 5/6 which means that 5 out of 6 bits are punctured. This compares with the rate that would have been used in the prior art of 6/7 that is to say 6 parity bits punctured out of 7. The prior art puncture algorithm is shown in the upper part 50 ofthe figure. It will be seen that each ofthe blocks for example 51, 52 and 53 represent bits. The bit 51 is retained and the bits 52 and 53 are kept in accordance with the key 54. In the lower part ofthe figure the puncturing algorithm according to this embodiment is depicted. A first segment 55 of this is punctured in accordance with the lower rate 5/6 and a second segment 56 is punctured at the higher rate 7/8. The rates switch at switching points 57 at 98 bit intervals. Enlargement 58 of part ofthe first segment shows bits 59 to 63 punctured and bit 64 kept. Enlargement 65 of part ofthe second segment shows bits 66 punctured and bit 67 retained.

The combination ofthe higher and lower rates matches overall the rate ofthe prior art approach- it can be seen that over the depicted sequences there are 14 kept bits for both the top and bottom sequences. However, this approach avoids the use of a rate of 6/7 which is one ofthe problematic rates for performance. Thus, the key to this embodiment is that a desired overall rate which is a problematic rate is achieved by puncturing the parity streams at higher and lower rates which combine to give the desired rate overall. In this way problematic puncturing rates are avoided. The same technique is used in the second parity stream but the puncture positions could be offset. Fig.6 shows a second embodiment. The desired rate is the same as in the first embodiment, that is to say the problematic rate of 6/7. The puncturing algorithm employed in this embodiment requires a first pass in which a high puncturing rate is employed to create a first set followed by a second pass in which a proportion ofthe punctured bits are re-inserted into the bit set. The re-inserted bits are shown arrowed in the figure and if we first ignore these and consider the first set it will be seen that there are twelve selected to be kept at a separation of seven. That is to say, every eighth bit is retained. To maintain the coding rate it is necessary to retain a further two bits and the arrowed bits 68, 69 are selected in the second pass. The other bits are eliminated in the puncturers 27 or 28. The same technique is used in the second parity stream but the puncture positions could be offset.

The last embodiment to be described is a refinement ofthe second embodiment. In the second embodiment the distribution ofthe bits to be retained is not a uniform distribution because the second pass bits arrowed are not placed with the same bit separation to the retained bits from the first pass. It will be seen that arrowed bit 68 is separated from its neighbouring retained bits from the first pass by three punctured bits. The arrowed bit 69 has a separation of two and four punctured bits between it and its immediate neighbouring bits retained by the first pass. The uneven separation of this embodiment is undesirable.

In the last embodiment illustrated in Fig. 7, it will be seen that the second pass bits to be retained are selected to maintain a uniform distribution of puncture (or kept bit) positions within the inherent recursive period. Both arrowed bit 70 and 71 are selected in the second pass to be separated from their neighbouring retained bits from the first pass by three punctured bits. By maintaining this uniform distribution the performance is enhanced because use of all positions within the inherent recursive period is ensured.

The algorithm used for puncturing the bits is based on that set out in the current 3 GPP specifications with certain modifications as set out below. The modified sections are labelled to the side as new section starts to new section ends.

eplus and eminus calculated as per 3 GPP specification, offset = int(eplus / repetition period) ; repetition period of 3GPP turbo coder is 2^Λ3-1 = 7

6plus_low ©plus OtlSet 6minus_low ^minus Ollset

Spiusjiigh = e_pius + offset 6minus_high ©_mj_nus + OtlSet

if puncturing is to be performed e = e_ini — initial error between current and desired puncturing ratio m = l - index of current bit count =0 ~ new variable mode = low — new variable

do while m

switch(mode) -new section starts-

case low: eplus = eplus_low eminus = eminus_low

case high: eplus = eplusjbigh eminus = eminus_high -new section ends-

e = e — e_mi_nus — update error if e <= 0 then — check if bit number m should be punctured set bit x_iwfn to δ where <5i_?{0, 1} e = e + e_pι_us — update error end if m = m + 1 — next bit count = count +1

switch(mode) -new section starts-

case low: if(count > 41) then count =0 mode = high end if case high: if(count>55) then count=0 mode = low end if -new section ends- end do Other embodiments will be apparent from the following description.

4.5.4 Hybrid ARQ for HS-DSCH:

The hybrid ARQ functionality matches the number of bits at the output ofthe channel coder to the total number of bits ofthe HS-PDSCH set to which the HS-DSCH is mapped. The hybrid ARQ functionality is controlled by the redundancy version (RV) parameters. The exact set of bits at the output ofthe hybrid ARQ functionality depends on the number of input bits, the number of output bits, and the RV parameters.

The hybrid ARQ functionality consists of two rate-matching stages and a virtual buffer as shown in Fig. 8.

The first rate matching stage matches the number of input bits to the virtual IR buffer, information about which is provided by higher layers. Note that, if the number of input bits does not exceed the virtual IR buffering capability, the first rate-matching stage is transparent. The second rate matching stage matches the number of bits after first rate matching stage to the number of physical channel bits available in the HS-PDSCH set in the TTI. The notation used is set out below.

N^πι : Number of bits in a transmission time interval before rate matching.

AN. : An intermediate calculation variable.

ANj ^! : If positive - number of bits to be repeated in each transmission time interval on TrCH i with transport format /.

If negative - number of bits to be punctured in each transmission time interval on TrCH i with transport format /.

N _PARITY • ' Number of bits to adjust the length of parity streams by, in HARQ 1^st stage rate-matching.

^N _data •' ^Total number of bits that are available for the HS-DSCH in a TTI.

eini Initial value of variable e in the rate matching pattern determination algorithm of subclause 4.2.7.5. e_pι_us Increment of variable e in the rate matching pattern determination algorithm of subclause 4.2.7.5.

e_m;_nus Decrement of variable e in the rate matching pattern determination algorithm of subclause 4.2.7.5.

b: Indicates systematic and parity bits

b=\ : Systematic bit. x_k in subclause 4.2.3.2.1.

b=2: 1^st parity bit (from the upper Turbo constituent encoder). z_λ in subcaluse 4.2.3.2.1.

b=3: 2^nd parity bit (from the lower Turbo constituent encoder). z'_k in subclause 4.2.3.2.1.

HARQ bit separation

The HARQ bit separation function shall be performed in the same way as bit separation for turbo encoded TrCHs

HARQ First Rate Matching Stage HARQ first stage rate matching for the HS-DSCH transport channel shall be done with the method described below with the following specific parameters. The maximum number of soft bits available in the virtual IR buffer is NIR which is signalled from higher layers for each HARQ process. The number of coded bits in a TTI before rate matching is N^TTI this is deduced from information signalled from higher layers and parameters signalled on the HS-SCCH for each TTI. Note that HARQ processing and physical layer storage occurs independently for each HARQ process currently active.

If NI is greater than or equal to N^TTI (i.e. all coded bits ofthe corresponding TTI can be stored) the first rate matching stage shall be transparent. This can, for example, be achieved by setting e_minus = 0. Note that no repetition is performed.

If NIR is smaller than N the parity bit streams are punctured as in 4.5.4.2.1 below by setting the rate matching parameter tXN™ = N_IR -N™ where the subscripts i and 1 refer to transport channel and transport format in the referenced sub-clause. Note the negative value is expected when the rate matching implements puncturing. Bits selected for puncturing which appear as δ in the algorithm in 4.2.7 above shall be discarded and not counted in the. totals for the streams through the virtual IR buffer.

4.5.4.2.1 HARQ First Rate Matching Parameters

If first stage puncturing is to be performed, the procedure below shall be applied. Index b is used to indicate systematic (b=l), 1^st parity (b=2), and 2^nd parity bit (b=3). The parameter AN_PARITY : is the change in length to the parity streams to circumvent problematic puncturing rates. The first stage rate matching delta is calculated as follows:

if puncturing is to be performed

while (a < 4)

■ - PR + - ^' 2 round ■

1 - a

λ — PR — round

= +l

δ 'P_DR_B = 98

else if {λ > 0) then ΔN PARITY _ΔN ^/ /2 - X,

a - 4

else

ΔN PARITY [ΔN, /2j-A.

a — A

end if

end while

ΔNf /2 ^V PARITY } 0 — Z

AN, ^'AN™ / 2 + Δ _pmτγ ,b = 3

a = 2 when b=2

a = 1 when b=3

If ΔN. is calculated as 0 for b=2 or b=3, then the following procedure and the rate matching algorithm of subclause 4.2.7.5 shall not be performed for the corresponding parity bit stream.

For each radio frame, the rate-matching pattern shall be calculated using the algorithm in subclause 4.2.7.5, where: i is as above,

^epius ~ aXA,

ax|ΔN_;-| 4.5.4.3 HARQ Second Rate Matching Stage

HARQ second stage rate matching for the HS-DSCH transport channel shall be done using one of two possible methods.

If second stage puncturing is to be performed, and the Composite Puncturing Rate of one ofthe parity streams, calculated using the formula below falls in any ofthe intervals [91/128,92/128], [217/256, 222/256], [231/256, 232/256], [237/256, 238/256] or [487/512, 488/512], then puncturing of that parity stream shall be done using the method described in 4.5.4.3.1. Note, puncturing using the method in 4.5.4.3.1 maybe carried out on one, both, or neither ofthe two parity streams, but never on the systematic parity stream.

Otherwise, second stage rate matching for the HS-DSCH transport channel shall be done with the general method described in 4.2.7.5 above with the following specific parameters. Bits selected for puncturing which appear as in the algorithm in 4.2.7.5 above shall be discarded and are not counted in the streams towards the bit collection. The parameters ofthe second rate matching stage depend on the value ofthe RV parameters s and r. The parameter s can take the value 0 or 1 to distinguish between transmissions that prioritise systematic bits (s = 1) and non systematic bits (s = 0). The parameter r (range 0 to r_max-l) changes the initial error variable e_ini in the case of puncturing. In case of repetition both parameters r and s change the initial error variable e_ini. The parameters X;, e_pι_us and e_minus are calculated as per table 10 below. Denote the number of bits before second rate matching as N_sys for the systematic bits, N_pi for the parity 1 bits, and N_P2 for the parity 2 bits, respectively. Denote the number of physical channels used for the HS-DSCH by P. N_data is the number of bits available to the HS-DSCH in one TTI and defined as N_data=PAiAN_data\, where N_datai is defined in [2]. The rate matching parameters are determined as follows. For N_daia < N_svs + N_pl + N_p2 , puncturing is performed in the second rate matching stage. The number of transmitted systematic bits in a transmission is N_{t iyi} = imn\N_m ,N_data j for a transmission that prioritises systematic bits and

N_{t ivs} - max{N_rfα,_a - [N_pl + N_p2

for a transmission that prioritises non systematic bits.

For N_data > N_m. + N_pl + N_p2 repetition is performed in the second rate matching stage.

A similar repetition rate in all bit streams is achieved by setting the number of

transmitted systematic bits to N, _svs =

The number of parity bits in a transmission is:

N_data -N_tt sys N, , = ma ^' N PARITY > and N, _p2 = N_data - (^ + N_t . ) for the

parity 1 and parity 2 bits, respectively.

Table 10 below summarizes the resulting parameter choice for the second rate matching stage.

Table 10: Parameters for HARQ second rate matching

The rate matching parameter β/,„- is calculated for each bit stream according to the RV parameters r and s using ^e, (r) = {{X, - [r ^■ e_plm I r_m J- 1) mod e_plui )+ 1 in the case of puncturing , i-AN_Λ„„ < N_Vi + N_pi + N_p2 , and ^e,_m (r) = {(A, - [(s + 2 - r) ^■ e_pl 1(2 ^■ r_ma )J- 1) mod e_pha }+ 1 for repetition, i.e.,

^N _{d a} > ^N _syi ^{+ N} _pi + ^N _p2 - Where re {0,l,-- -,r_max -1} and r_max is the total number of redundancy versions allowed by varying r as defined in 4.6.2. Note that r_max varies depending on the modulation mode, i.e. for 16QAM r_ma = 2 and for QPSK r_max = 4. Note: For the modulo operation the following clarification is used: the value of (x mod y) is strictly in the range of 0 to j -l (i.e. -1 mod 10 = 9).

4.5.4.3.1 HARQ Second Rate Matching Stage for problematic puncturing rates HARQ second stage rate matching for the HS-DSCH transport channel shall be done with the method described below, which separates each stream into segments and then applies the general method described in 4.2.7.5 above to each ofthe segments with specific parameters calculated below. Bits selected for puncturing which appear as r in the algorithm in 4.2.7.5 above shall be discarded and are not counted in the streams towards the bit collection.

4.5.4.3.1.1 Parity stream segmentation

The parity stream will be segmented into 3 segments, the first segment will be made up ofthe first X_segι bits ofthe parity stream; the second segment will be made up ofthe next X_seg2 bits, and the last segment will be made up ofthe remaining X_seg3 bits.

X X X

The first segment is denoted by '^■' ' '>² '^"' '•^Λi«¹ The second segment by *«.ΛS«I+I > ...»-gi+2 >-^x<,^χ _ieS2 .

And the final segment by ^Λ:'.Λ*ι+Λ.«2+ι .^Jf.,Λ.«ι+Λ«2+2»—^Λ:».Λr₁ Where ,,. = 42 - 1^ /98 ]

X_ieg2 = 56 ^• [X, 198j X_seg3 = X, mod 98 Note: If X; is less than 98 only the third segment exists, and if X; is a multiple of 98 then the third segment will be empty. For the case where a particular segment is empty, then of course no puncturing is performed on that non-existent segment.

4.5.4.3.1.2 Additional parameters for HARQ Second Rate Matching Stage for problematic puncturing rates

Parameters P, Ndata, N_sys, N_pl, N_p2, N_p>tl, and N_P;t2 are calculated as in 4.5.4.3 above, additional parameters are defined below. Denote the number of parity 1 bits before second rate matching after segmentation as

N_pι,_Segi, N_pl)Seg2, N_pι_;Seg3 for the parity 1 bits in the 1^st, 2^nd and 3^rd segment respectively.

Denote the number of parity 2 bits before second rate matching after segmentation as

N_p2>Segi, N_p2;Seg2, p _!se_g3 for the parity 1 bits in the 1^st, 2^nd and 3^rd segment respectively.

The rate matching parameters are determined as follows. For N_data ≤ N_sys + N_pl + N_p2 , puncturing is performed in the second rate matching stage.

The number of parity bits after segmentation is

^^₁ = 42 - 1^ /98]

N_pb,seg2 = 56 - [N_pb /9%} N_pl,_!Ses3 = N_pb mod98

The number of parity bits in each segment for parity 1 bits(b=2), and parity 2bits (b=3) is:

^•' * t,pb,seg3 t,pb ^V t,pb,seg\ ^{"■" ■}* * t,pb,seg2 )

The parameters X;, e_pι_us and e_rai_nus are calculated as per table 10a below. Table 10a: Parameters for dithered HARQ second rate matching

The rate matching parameter e_ιm is calculated for each segment of bit stream according to the RV parameters r and s using ^e„„ (^r) = fc - l^{r • e}pi I Αx J- 1) m° e_pha j+ 1 in the case of puncturing. Where r e {0, 1, • • ^• , r_max — l) and r_max is the total number of redundancy versions allowed by varying r as defined in 4.6.2. Note that r_max varies depending on the modulation mode, i.e. for 16QAM r_max = 2 and for QPSK r_ma = 4. Note: For the modulo operation the following clarification is used: the value of (x mod y) is strictly in the range of 0 to^-1 (i.e. -1 mod 10 = 9). The rate matching algorithm in 4.2.7.5 is called for each segment ofthe parity stream in turn.

4.5.4.3.1.3 Parity stream segment concatenation For both parity streams, after the rate matching algorithm in 4.2.7.5 has been called for each ofthe three segments, the three punctured segments should be concatenated together in their original order. The re-concatenated bit stream is denoted by:

^Xo,segl,l ' ^Xo,seg\,2 »^• "^Xo,segl,N_{l p} ^, ' ^Xo,seg2,l ' ^Xo,seg2,2 v ^■Xo,seg2,N_{l w2} ' ^Xo,seg3,\ ' ^Xo,ieg₃,2 v ^•x ₀,seg3,N

4.5.4.4 HARQ bit collection The HARQ bit collection is achieved using a rectangular interleaver of size N_row x N_col The number of rows and columns are determined from:

N_mw = 4 for 16QAM and N_row = 2 for QPSK

N¹ col , = N d,ata I ' N ^x row

where N_data is used as defined in 4.5.4.3. Data is written into the interleaver column by column, and read out ofthe interleaver column by column starting from the first column.

N_t,_sys is the number of transmitted systematic bits. Intermediate values N_r and N_c are calculated using:

N t,,sys

N.. = ∞d N_c = N_lΛ ,sy_βs -N ^J * _rr . N

N ^■i c,ol ' c.ol

If N_c=0 and N_r> 0, the systematic bits are written into rows 1...N_r.

Otherwise systematic bits are written into rows 1...N_r+1 in the first N_c columns and, if N_r > 0, also into rows 1...N_r in the remaining N_coι-N_c columns. The remaining space is filled with parity bits. The parity bits are written column wise into the remaining rows ofthe respective columns. Parity 1 and 2 bits are written in alternating order, starting with a parity 2 bit in the first available column with the lowest index number. If the two parity streams are of unequal length, Parity 1 and 2 bits shall be written alternately, again starting with a parity 2 bit, until the end ofthe shorter parity stream, then the remaining parity bits from the longer stream shall be written. In the case of 16QAM for each column the bits are read out ofthe interleaver in the order row 1 , row 2, row 3, row 4. In the case of QPSK for each column the bits are read out ofthe interleaver in the order rowl, row2.

Claims

1. An encoder comprising: at least one recursive systematic convolutional encoder for, in use, producing at least one parity stream; at least one rate modifier for modifying a rate ofthe at least one parity stream by applying at least one of puncturing, or puncturing and repetition, wherein regularities ofthe positions ofthe punctured bits within a repetition period ofthe encoder are avoided by, in the case ofthe application of puncturing, making the puncturing irregular with respect to the repetition period or, in the case ofthe application of puncturing and repetition, modifying repetition and puncturing rates in a complementary manner to avoid regularities ofthe positions ofthe punctured bits within the repetition period ofthe encoder.

2. An encoder as claimed in claim 1 comprising a rate modifier for modifying the at least one parity stream.

3. An encoder as claimed in claim 1 or claim 2, wherein at least one puncturing rate is applied per segment ofthe at least one parity stream.

4. An encoder as claimed in claim 2 or claim 3, wherein two or more puncturing rates are applied per segment ofthe at least one parity stream.

5. An encoder as claimed in claim 1 or claim 2, wherein one puncturing rate is applied to at least a first segment ofthe at least one parity stream and at least two puncturing rates are applied to at least a second segment ofthe at least one parity stream.

6. An encoder as claimed in any preceding claim wherein the puncturing is carried out in such a manner that the positions of retained bits are pseudo-random.

7. An encoder as claimed in claim 1 or claim 2, wherein, in the case ofthe application of puncturing and repetition, the repetition rate is greater than the puncturing rate.

8. An encoder as claimed in claim 1 or claim 2, wherein, in the case ofthe application of puncturing and repetition, the puncturing rate is greater than the repetition rate.

9. Communications apparatus including an encoder as claimed in any preceding claim.

10. A method of puncturing data provided by an encoder which encoder providing at least one parity stream comprising the steps of: modifying a rate of the at least one parity stream by applying at least one of puncturing, or puncturing and repetition, wherein regularities ofthe positions ofthe punctured bits within a repetition period ofthe encoder are avoided by, in the case of the application of puncturing, making the puncturing irregular with respect to the repetition period or, in the case ofthe application of puncturing and repetition, modifying repetition and puncturing rates in a complementary manner to avoid regularities of the positions of the punctured bits within the repetition period of the encoder.

11. A method as claimed in claim 10, wherein at least one puncturing rate is applied per segment ofthe at least one parity stream.

12. A method as claimed in claim 11 , wherein two or more puncturing rates are applied per segment ofthe at least one parity stream.

13. A method as claimed in claim 11, wherein a first puncturing rate is applied to a first segment ofthe at least one parity stream and a second puncturing rate is applied to a second segment ofthe at least one parity stream.

14. A method as claimed in any of claims 10 to 13, wherein the at least one parity stream is punctured such that bits in the stream to be retained have pseudo random positions in the stream.

15. A method as claimed in any of claims 10 to 14, wherein in the case ofthe application of puncturing and repetition, the repetition rate is greater than the puncturing rate.

16. A method as claimed in any of claims 10 to 15, wherein in the case ofthe application of puncturing and repetition, the puncturing rate is greater than the repetition rate.