WO2020083491A1 - A distribution matcher, a channel encoder and a method for encoding data bits or symbols - Google Patents
A distribution matcher, a channel encoder and a method for encoding data bits or symbols Download PDFInfo
- Publication number
- WO2020083491A1 WO2020083491A1 PCT/EP2018/079285 EP2018079285W WO2020083491A1 WO 2020083491 A1 WO2020083491 A1 WO 2020083491A1 EP 2018079285 W EP2018079285 W EP 2018079285W WO 2020083491 A1 WO2020083491 A1 WO 2020083491A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- state
- bits
- distribution matcher
- shaped
- channel encoder
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0041—Arrangements at the transmitter end
- H04L1/0042—Encoding specially adapted to other signal generation operation, e.g. in order to reduce transmit distortions, jitter, or to improve signal shape
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/32—Carrier systems characterised by combinations of two or more of the types covered by groups H04L27/02, H04L27/10, H04L27/18 or H04L27/26
- H04L27/34—Amplitude- and phase-modulated carrier systems, e.g. quadrature-amplitude modulated carrier systems
- H04L27/36—Modulator circuits; Transmitter circuits
Abstract
The invention relates to a state-dependent distribution matcher for a channel encoder for encoding data bits or symbols into shaped bits or symbols having a non-uniform distribution. The state-dependent distribution matcher comprises: a first binary distribution matcher configured to generate on the basis of a first subset of a plurality of data bits a first subsequence of shaped bits; and a second binary distribution matcher configured to generate on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits. The state-dependent distribution matcher is configured to generate a sequence of shaped bits on the basis of the first subsequence of shaped bits, on the basis of the second subsequence of shaped bits, and on the basis of a sequence of state bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit is equal to 0, or (ii) a respective shaped bit from the second subsequence, in case a corresponding state bit is equal to 1.
Description
DESCRIPTION
A DISTRIBUTION MATCHER, A CHANNEL ENCODER AND A METHOD FOR
ENCODING DATA BITS OR SYMBOLS
TECHNICAL FIELD
Generally, the present invention relates to the field of channel coding. More specifically, the present invention relates to a distribution matcher for a channel encoder, a channel encoder comprising such a distribution matcher and a method for encoding data bits or symbols.
BACKGROUND
The global society currently faces a rapid growth of data traffic in the internet, which will continue in the coming decades. This puts a tremendous pressure on telecommunication companies, which need innovations to provide the required digital link capacities.
Frequency bands for data transmission are a very expensive and restricted resource in wireless, fibre-optic and copper-cable communications. To mitigate this bandwidth limitation, higher-order modulation is usually required, where more than 1 bit is mapped to each real-dimensional time-frequency slot. Common higher-order modulation formats are quadrature amplitude modulation (QAM) and amplitude phase-shift keying (APSK).
The electric energy required to transmit one signal point is proportional to the square of its absolute value. Consequently, the outer points of a modulation format require significantly more energy than the inner points. The energy-efficiency of a transmission scheme can therefore be improved by transmitting the inner points more frequently than the outer points. Techniques for implementing such power-efficient non-uniform signalling are called probabilistic shaping.
Another challenge for modern communication systems is the support of different transmission rates, from which the best can be chosen, according to the current requirements. For example, a modem for transmission over optical fibres may need to work over different distances, so it is desirable to choose a higher transmission rate for short distances and a lower transmission rate for longer distances.
A promising approach to achieve rate-flexible modems is layered probabilistic shaping, where the overhead is split into a constant part for linear forward error correction (FEC) and a variable part, which is realized by a variable rate distribution matcher (DM). As an example, figure 1 shows a conventional communication system 100 with a distribution matcher (DM) 101 serving as a shaping device, followed by a FEC encoder 103. The communication system 100 further comprises a bit mapper 105 on a transmitter side and a corresponding demapper 107, FEC decoder 109 and inverse DM 1 1 1 on a receiver side.
A popular distribution matcher is the constant composition distribution matcher (CCDM) 201 , which is shown in figure 2 and has been successfully deployed to enable power- efficient signalling and rate-flexible transmission (see P. Schulte & G. Bocherer,“Constant composition distribution matching,” IEEE Trans. Inf. Theory, vol. 62, pp. 430, 2016). The CCDM 201 outputs sequences of fixed length, and each amplitude occurs a fixed number of times in each output sequence (“constant composition”). Figure 3 shows an example for a set of sequences of bits of length 5 with = 2 ones and n0 = 3 zeros.
As illustrated in figure 3, the CCDM 201 can enumerate these sequences in a
lexicographic order. The largest number k of input bits is chosen such that 2k is smaller or equal to the number of constant composition sequences. In this example, there are 10 distinct sequences of bits with two 1 s and three 0s, so k = 3 is chosen, since 2k = 8 < 10 and 24 = 16 > 10. Consequently, two sequences are not used, i.e., 10100 and 1 1000.
The CCDM 201 can now map length 3 input strings to distinct output strings. Probabilistic shaping and FEC can now be integrated by making use of the probabilistic amplitude shaping (PAS) architecture (see G. Bocherer, F. Steiner, and P. Schulte,“Bandwidth efficient and rate-matched low-density parity-check coded modulation,” IEEE Trans.
Commun., vol. 63, no. 12, pp. 4651 -4665, 2015). The bit-mapper maps the DM output bits to amplitudes, i.e., 0 to 1 and 1 to 3. The FEC encoder adds parity bits, which the bit- mapper maps to signs, i.e., 0 to -1 and 1 to 1 , which corresponds to an intermediate ASK signal with the values {-3, -1 ,1 , 3}. This is then demultiplexed to a 16-QAM signal. Figure 4 shows from left to right the amplitude distribution imposed by the distribution matcher 201 , the distribution of the intermediate ASK signal, and the distribution of the QAM signal.
When a 64-QAM is used, corresponding to an eight-ASK constellation
(-7, -5, -3, -1,1, 3, 5, 7} in each real dimension, a power-efficient distribution on four
different amplitudes {1,3, 5, 7} has to be generated by the distribution matcher. Even larger constellations have even more distinct amplitudes.
To generate the optimal power-efficient distribution of QAM signal points, it is enough to generate non-uniformly distributed amplitude sequences, separately for each real dimension. For 16-QAM with a 4-ASK constellation {-3, -1,1,3} in each real dimension, only the two distinct amplitudes {1,3} per real dimension exist, so that shaping can be implemented by a binary CCDM, i.e., a CCDM that outputs binary sequences with exactly n0 zeros and P-L ones. For constellations larger than 16-QAM, a non-binary CCDM is required to achieve the optimal distribution. For the exemplary 64-QAM the CCDM 201 needs to output amplitude sequences with exactly ^ones, n3 threes, n5 fives and n7 sevens. A non-binary CCDM is more complex than a binary CCDM (see Pikus, Marcin &Wen Xu, "Bit-level probabilistically shaped coded modulation," IEEE Communications Letters 21.9, 2017; Ramabadran, Tenkasi V., "A coding scheme for m-out-of-n codes" IEEE Trans. Comm., vol. 38, pp. 1 156, 1990).
Furthermore, longer output sequences are required to achieve optimal performance, which, in turn, increases the delay and requires more memory and arithmetic precision. The use of binary CCDMs for modulation formats beyond 16-QAM is therefore highly desirable.
The non-binary CCDM is more complex than the binary CCDM, since the internal arithmetic coder is required to directly manipulate 4ary amplitude distributions. To approximate a non-binary CCDM by binary CCDMs, product distribution matching (PDM) has been proposed as illustrated in figure 5, where the product distribution matcher 500 comprises first and a second binary CCDM 503, 505 as well as a de-multiplexer 501 and a multiplexer 507. More details about product distribution matching (PDM) can be found in the following documents: Pikus, Marcin & Wen Xu, "Bit-level probabilistically shaped coded modulation," IEEE Communications Letters 21 .9, 2017; Bocherer, Georg et. al., "High throughput probabilistic shaping with product distribution
matching," arXiv:1702.07510, 2017; F. Steiner, G. Bocherer, and P. Schulte,
“Approaching waterfilling capacity of parallel channels by higher order modulation and probabilistic amplitude shaping,” in Proc. Conf. Inf. Sciences and Systems (CISS), USA, 2018; and EP 3 306 821 A1 .
However, product distribution matching (PDM) has several problems: first of all, distributions generated by PDM must have power-of-two support sizes. In particular, PDM cannot be used for 36 QAM, where 3 (not a power of two!) amplitude values must be used in each real dimension. Secondly, distributions generated by PDM are the Kronecker product of two binary distributions and are therefore suboptimal in terms of power efficiency. Thirdly, the symbol sequences output by PDM are not constant composition. Consequently, PDMs are not non-binary CCDM compliant, which implies limited interoperability.
In light of the above, there is still a need for an improved distribution matcher and channel encoder comprising such a distribution matcher as well as a method allowing, in particular, for encoding a plurality of data bits or symbols into a plurality of shaped bits or symbols more efficiently.
SUMMARY
It is an object of the invention to provide an improved distribution matcher and channel encoder comprising such a distribution matcher as well as a method allowing, in particular, for encoding a plurality of data bits or symbols into shaped bits or symbols having a non-uniform distribution more efficiently.
The foregoing and other objects are achieved by the subject matter of the independent claims. Further implementation forms are apparent from the dependent claims, the description and the figures.
Generally, embodiments of the invention enable efficient distribution matching for more than two distinct amplitudes. Thereby, embodiments of the invention offer in particular the following advantages: a chained constant composition distribution matcher (CCDM) can implement a non-binary CCDM by using only binary component CCDMs; a chained CCDM can generate symbol distributions with arbitrary support sizes, in particular, distributions with support sizes that are not powers of two, in particular support sizes 3, 5, 6, 7; an extended chained CCDM achieves better performance than conventional product distribution matching (PDM), with binary component CCDMs of the same size.
More specifically, according to a first aspect the invention relates to a state-dependent distribution matcher for a channel encoder for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution, wherein the state-dependent distribution matcher comprises: a first binary distribution matcher configured to generate on the basis of a first subset of a plurality of data bits a first subsequence of shaped bits; and a second binary distribution matcher configured to generate on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits; wherein the state-dependent distribution matcher is configured to generate a sequence of shaped bits on the basis of the first subsequence of shaped bits provided by the first binary distribution matcher, on the basis of the second subsequence of shaped bits provided by the second binary distribution matcher, and on the basis of a sequence of state, i.e. state-defining bits by selecting (i) a respective shaped bit from the first subsequence, in case a
corresponding state bit of the sequence of state bits is equal to 0, or (ii) a respective shaped bit from the second subsequence, in case a corresponding state bit of the sequence of state bits is equal to 1.
Thus, an improved state-dependent distribution matcher is provided, allowing for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution more efficiently.
In a further possible implementation form of the first aspect, the state-dependent distribution matcher is further configured to partition a subset of the plurality of data bits into the first subset and the second subset of the plurality of data bits.
In a further possible implementation form of the first aspect, the first binary distribution matcher and/or the second binary distribution matcher is a constant composition distribution matcher.
According to a second aspect the invention relates to a channel encoder for encoding data bits or symbols into shaped bits or symbols having a non-uniform distribution, wherein the channel encoder comprises a first state-dependent distribution matcher according to the first aspect of the invention. The channel encoder is configured to generate a sequence of shaped symbols on the basis of the sequence of shaped bits and the sequence of state bits by mapping pairs or tuples of the shaped bits and the state bits to a plurality of symbols.
Thus, an improved channel encoder is provided, allowing for encoding data bits or symbols into shaped bits or symbols more efficiently.
In a further possible implementation form of the second aspect, the channel encoder further comprises a third binary distribution matcher configured to generate on the basis of a third subset of the plurality of data bits complimentary to the first subset and the second subset of the plurality of data bits the sequence of state bits.
In a further possible implementation form of the second aspect, the third binary distribution matcher is a constant composition distribution matcher.
In a further possible implementation form of the second aspect, the channel encoder further comprises a second state-dependent distribution matcher according to the first aspect of the invention and a third state-dependent distribution matcher according to the first aspect of the invention. The second state-dependent distribution matcher provides the first binary distribution matcher and the third state-dependent distribution matcher provides the second binary distribution matcher of a fourth state-dependent distribution matcher according to the first aspect of the invention.
In a further possible implementation form of the second aspect, the channel encoder is further configured to use the sequence of shaped bits provided by the third binary distribution matcher as the sequence of state bits for the fourth state-dependent distribution matcher.
In a further possible implementation form of the second aspect, the channel encoder is further configured to use the first subsequence of shaped bits provided by the first binary distribution matcher of the first state-dependent distribution matcher as the sequence of state bits for the second state-dependent distribution matcher and/or to use the second subsequence of shaped bits provided by the second binary distribution matcher of the first state-dependent distribution matcher as the sequence of state bits for the third state- dependent distribution matcher.
In a further possible implementation form of the second aspect, the channel encoder comprises a second state-dependent distribution matcher according to the first aspect of the invention, wherein the channel encoder is configured to use the sequence of shaped
bits generated by the first state-dependent distribution matcher as the sequence of state bits for the second state-dependent distribution matcher.
In a further possible implementation form of the second aspect, the channel encoder comprises a second state-dependent distribution matcher according to the first aspect of the invention and the first or second binary distribution matcher of the first state- dependent distribution matcher is provided by the second state-dependent distribution matcher.
In a further possible implementation form of the second aspect, pairs or tuples of the shaped bits and corresponding pairs or tuples of the sequence of state bits follow a desired joint distribution.
According to a third aspect the invention relates to a transmitter communication device comprising a channel encoder according to the second aspect of the invention for encoding a plurality of data bits for transmission to a receiver communication device.
Thus, an improved transmitter communication device is provided, allowing for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution efficiently.
According to a fourth aspect the invention relates to a corresponding method for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution. The method comprises: generating on the basis of a first subset of a plurality of data bits a first subsequence of shaped bits; generating on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits; and generating a sequence of shaped bits on the basis of the first subsequence of shaped bits, on the basis of the second subsequence of shaped bits, and on the basis of a sequence of state, i.e. state defining bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit of the sequence of state bits is equal to 0, or (ii) a respective shaped bit from the second subsequence in case a corresponding state bit of the sequence of state bits is equal to 1.
Thus, an improved method is provided, allowing for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution efficiently.
The invention can be implemented in hardware and/or software.
BRIEF DESCRIPTION OF THE DRAWINGS
Further embodiments of the invention will be described with respect to the following figures, wherein:
Fig. 1 shows a schematic diagram illustrating a communication system comprising an encoder, a communication channel and a decoder;
Fig. 2 shows a schematic diagram of a constant composition distribution matcher;
Fig. 3 shows a schematic diagram illustrating a set of sequences of length 5 bits output by a constant composition distribution matcher;
Fig. 4 shows a schematic diagram illustrating a distribution of amplitudes, a distribution of intermediate amplitude-shift-keying signals and a distribution of quadrature-amplitude- modulation signals output by a distribution matcher;
Fig. 5 shows a schematic diagram illustrating a product distribution matching scheme;
Fig. 6 shows a schematic diagram illustrating a channel encoder according to an embodiment comprising a distribution matcher according to an embodiment;
Fig. 7 shows a table of representation of amplitudes by binary labels implemented by a channel encoder according to an embodiment;
Fig. 8 shows a table illustrating an amplitude distribution of a 64-QAM constellation implemented by a channel encoder according to an embodiment;
Fig. 9 shows a table illustrating number of occurrences for two-bit levels
B2 implemented by a channel encoder according to an embodiment;
Fig. 10 shows a table illustrating an amplitude distribution of a 36-QAM constellation implemented by a channel encoder according to an embodiment;
Fig. 1 1 shows a table illustrating number of occurrences for two-bit levels Bx and B2 implemented by a channel encoder according to an embodiment;
Fig. 12 shows a table illustrating chained probabilities for two-bit levels Bt and B2 implemented by a channel encoder according to an embodiment;
Fig. 13 shows a table illustrating appending a second sequence of bits B2 to a first sequence of bits B1 implemented by a channel encoder according to an embodiment;
Fig. 14 shows a table illustrating appending a third sequence of bits B3 to a first sequence of bits B1 implemented by a channel encoder according to an embodiment;
Fig. 15 shows a schematic diagram illustrating a channel encoder according to an embodiment comprising a state-dependent distribution matcher according to an embodiment; and
Figure 16 shows a schematic diagram illustrating a method for encoding data bits or symbols into shaped bits or symbols according to an embodiment.
In the various figures, identical reference signs will be used for identical or at least functionally equivalent features.
DETAILED DESCRIPTION OF EMBODIMENTS
In the following description, reference is made to the accompanying drawings, which form part of the disclosure, and in which are shown, by way of illustration, specific aspects in which the present invention may be placed. It is understood that other aspects may be utilized, and structural or logical changes may be made without departing from the scope of the present invention. The following detailed description, therefore, is not to be taken in a limiting sense, as the scope of the present invention is defined by the appended claims.
For instance, it is understood that a disclosure in connection with a described method may also hold true for a corresponding device or system configured to perform the method and vice versa. For example, if a specific method step is described, a corresponding device may include a unit to perform the described method step, even if such unit is not explicitly described or illustrated in the figures. Further, it is understood that the features of the
various exemplary aspects described herein may be combined with each other, unless specifically noted otherwise.
As will be described in more detail in the following, embodiments of the invention relate to a state-dependent distribution matcher and a channel encoder comprising such a state- dependent distribution matcher, which is configured to encode data bits or symbols into shaped bits or symbols and can enable efficient distribution matching when more than two distinct amplitudes are to be used.
Figure 6 shows a schematic diagram of a channel encoder 600 according to an embodiment. The channel encoder 600 is configured to encode data bits or symbols into shaped bits or symbols having a non-uniform distribution. As will be described in more detail below, the channel encoder 600 comprises as a kind of basic processing cell a state-dependent distribution matcher 61 1 a.
By way of example, in the following the basic functioning of the channel encoder 600 will be described in more detail in the context of a 64-QAM (quadrature amplitude
modulation), where the distribution matcher (DM) outputs sequences with 4 distinct amplitudes {1 , 3, 5, 7}. The person skilled in the art, however, will appreciate that the channel encoder 600 can be advantageously operated using different kinds of
modulations as well.
As illustrated in figure 6, the input data bits provided to the channel encoder 600 are de multiplexed into three different bit streams or subsets by a first demultiplexer or bit- mapper 601 and a second de-multiplexer 602. Each bit stream is the input of a“binary” CCDM 603a-c. The binary output of the first and second CCDM 603b, 603c are combined to a single output bit stream by a multiplexer 605, which uses the output of the third CCDM 603a as state information. The output of the third CCDM 603a and the output of the first multiplexer 605 are then combined by a second multiplexer or bit mapper 607, which outputs a sequence of shaped amplitudes.
More specifically, the channel encoder 600 according to an embodiment comprises a first state-dependent distribution matcher 61 1 a configured to partition a subset of a plurality of data bits into a first subset and a second subset of the plurality of data bits. The first state- dependent distribution matcher 61 1 a comprises the first binary distribution matcher 603b configured to generate a first subsequence of shaped bits on the basis of a first subset of
the plurality of data bits; and the second binary distribution matcher 603c configured to generate a second subsequence of shaped bits on the basis of a second subset of the plurality of data bits.
The first state-dependent distribution matcher 61 1 a is configured, for instance by means of the multiplexers 605, illustrated in figure 6, to generate a sequence of shaped bits on the basis of the first subsequence of shaped bits provided by the first binary distribution matcher 603b, on the basis of the second subsequence of shaped bits provided by the second binary distribution matcher 603c, and on the basis of a sequence of state, i.e. state-defining bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit of the sequence of state bits is equal to 0, or (ii) a respective shaped bit from the second subsequence, in case a corresponding state bit of the sequence of state bits is equal to 1 .
As will be described in more detail further below, the state-dependent distribution matcher 61 1 a, which in the exemplary embodiment of figure 6 comprises the first and second binary distribution matcher 603b, 603c as well as the de-multiplexer 602 and the“state- dependent” multiplexer 605 (as illustrated by the dashed box in figure 6), defines a kind of basic processing cell. In embodiments of the invention, the channel encoder 600 can comprise one or more than one implementations of this basic processing cell, i.e. the state-dependent distribution matcher 61 1 a also in a nested arrangement, as will be described in more detail in the context of figure 15.
In an embodiment, the channel encoder 600 can generate a sequence of shaped symbols on the basis of the sequence of shaped bits and the sequence of state bits by mapping pairs or tuples of the shaped bits and the state bits to a plurality of symbols.
Furthermore, as illustrated in figure 6, the channel encoder 600 can further comprise a third binary distribution matcher 603a configured to generate on the basis of a third subset of the plurality of data bits complimentary to the first subset and the second subset of the plurality of data bits the sequence of state bits.
In an embodiment, the first binary distribution matcher 603b, the second binary distribution matcher 603c and/or the third binary distribution matcher 603a can be a constant composition distribution matcher.
At the receiver side, the chained CCDM architecture follows by the processing in a reverse direction to the processing direction shown in figure 6, namely starting at the shaped amplitudes and ending at the data bits. The multiplexers 605, 607 can be replaced by their inverses and the binary distribution matchers 603a-c can be replaced by the corresponding binary distribution de-matchers.
By way of example, in the following the operation of the channel encoder 600 will be described in more detail considering a normalized 64-QAM constellation with signal point amplitudes being 1, 3, 5, 7 in each real dimension and transmission of length n = 10 sequences with 40% of the signal points with amplitude 1, 30% with amplitude 3, 20% with amplitude 5 and 10% with amplitude 7 in each real dimension. By way of example, each transmitted amplitude sequence is a permutation of the sequence 1111333557. The channel encoder 600 according to the present invention can index all permutations of this sequence. Each amplitude can thus be represented by a binary label according to table 700 shown in figure 7.
The sequence of amplitudes can now be represented by the sequence of bit labels:
(00) (00) (00) (00) (01) (01) (01) (10) (10) (11). The number of amplitude and bit label occurrences are listed in table 800 illustrated in figure 8.
Considering that BI, B2 represent the two bits in a bit-label, an alternative representation of the sequences is via the length n = 10 and the probabilities:
Based on the chain rule of probabilities, PBlB2 (b i&2) = PB1 (^I)PB2 |B1 (^2 ^I)> the sequences can be indexed in a different way. The number of occurrences of 0s and 1 s in the chained probabilities are summarised in table 900 illustrated in figure 9.
Table 900 shown in figure 9 can be used to index the sequences as follows, which is implemented in the channel encoder 600 according to an embodiment of the invention (referred to as a chained CCDM): first, the three binary CCDMs 603a-c are used to generate one length 10 sequence b1 with distribution PBi , one length 7 sequence b2 with distribution PB2\Bl=0 and one length 3 sequence b3 with distribution PB2 \Bl=i, e.g., b1 = 0101000100, b2 = 1001001, b3 = 001.
Secondly, entry-wise b2 can be appended to bt at the positions where the bt entries are equal to zero, for example, (01)(1)(00)(1)(00)(01)(00)(1)(00)(01). Also, entry-wise b3 can be appended to b1 at the positions where the b1 entries are equal to one, for example, (01)(10)(00)(10)(00)(01)(00)(11)(00)(01).
Finally, the bit pairs are mapped to amplitudes:
(01)(10)(00)(10)(00)(01)(00)(11)(00)(01) are mapped to 3515131713. This result shows that the desired distribution of the amplitudes can indeed be obtained: four 1 s, three 3s, two 5s and one 7.
According to an embodiment, exemplary implementations of the multiplexers 605, 607 of the channel encoder 600 shown in figure 6 can be realized as follows:
The first multiplexer 605 can be implemented according to the following algorithm:
receive inputs of DMBi : b = b lb 2...b n (the sequence of state bits B ), DMB2
output a sequence of bit 2, b2 = b2 1 ... b2 n,
initialize by i = 1 ,j = l, k = 1, and
repeat until k = n, wherein if bl k = 0: b2 k <- c0 , j
+ 1, /c <- /c + 1 and wherein if bl k = 1 : b2 k <- cl i ; i <- i + 1, k <- k + 1.
The second multiplexer 607 can be implemented according to the following algorithm: receive inputs of DMBi : b = b lb 2 - ..b n (the sequence of state bits B ) and the output of the first multiplexer 605: b2 = b2 1b 2.. J n (the sequence of bit B2 );
output sequence of amplitudes an = ax a2 ... an;
initialize by k = 1; and
repeat until k = n, wherein ak <- f(bl kb2 k), k <- k + 1, where / implements the mapping: 00 to 1 , 01 to 3, 10 to 5 and 1 1 to 7.
Furthermore, the input lengths k, k0, k t in figure 6 can be calculated as follows: the third binary distribution matcher DMBI 603a can index length 10 sequences with 7 zeros and 3 ones. There are 7 ) =—
= 120 such sequences, so the number of input bits for the third binary distribution matcher DMBI 603a can be k = [log2 120J = 6.
The first binary distribution matcher DMB2 |Bi =0 603b can index length 7 sequences with 4 zeros and 3 ones. There are = = 35 of such sequences, so the number of input bits for the first binary distribution matcher DMB2 |Bi =0 603b can be k0 = [log2 35J = 4.
The second binary distribution matcher DMB2 |Bi =1 603c can index length 3 sequences with 2 zeros and 1 one. There are (2) =
= 3 of such sequences, so the number of input bits for the second binary distribution matcher DMB2 |Bi =1 603c can be k = [log2 3J = 1. The total input length of the channel encoder, i.e. chained CCDM 600 in this example is k + ko + k = 11 bits.
The first demultiplexer 601 a of the channel encoder 600 shown in figure 6 is configured to partition a length k + k0 + k 1 binary sequence into three sequences of lengths k, k0, and kx, respectively. Any suitable demultiplexing strategy can be applied with the chained CCDM architecture. An exemplary embodiment of the demultiplexer 601 can be implemented as follows: the input sequence of the channel encoder, i.e. chained CCDM 600 is bk total = b1b2 - bktotal with /ctotal = k + k0 + kt = 11; the input sequence for the third binary distribution matcher DMBI 603a is b1b2 ... bk = b^^^b^b^^, the input sequence for the first binary distribution matcher DMB2 |Bi= 0 603b is bk+1bk+2 ... bk+ko = b7b b9bw and the input sequence for the second binary distribution matcher DMB2 |Bi =1 603c is
^/ +/ Q + 1 bk--k +2 bk--k o+/ ^ bxx.
According to a further embodiment, if only three amplitudes per real dimension are used (corresponding to a 36-QAM constellation), this is equivalent to a 64-QAM constellation where amplitude 7 is not used, as shown in table 1000 illustrated in figure 10. The chained probabilities for the two bit-levels B1 and B2 are summarized in table 1 100 in figure 1 1. It is to be noted that, for the positions where Bt = 1, a DM is not necessary, since bit-level B2 is deterministically equal to 1.
According to a further embodiment the channel encoder 600 can be implemented as follows: two binary CCDMs are used to generate one length 10 sequence b1 with distribution PBi and one length 8 sequence b2 with distribution PB2 |Bl=0- Entry-wise b2 is appended to
at the positions where the b1 entries are equal to zero, and entry-wise 0s can be appended to b1 at the positions where the b1 entries are equal to zero (this realizes the distribution
Finally, the bit pairs are mapped to amplitudes.
According to a further embodiment, if the chained probabilities for the two bit-levels Bx and B2 follow table 1200 of figure 12 and if all binary component CCDMs are allowed to use output length 12, an extended chained CCDM provided by the channel encoder 600 according to an embodiment can be realized as follows:
The third binary CCDM 603a is run for PBi three times, which produces a b1 sequence of length 3 x 12 = 36 with 24 positions equal to 0 and 12 positions equal to 1 , for example b1 = 000000101110000010001101000010110001.
The first binary CCDM 603b is run for RB2 \B1 =O twice and this produces a sequence of length 2 x 12 = 24, corresponding to the number of positions equal to 0 in sequence bl t for example b2 = 010110001010010100001101.
The second binary CCDM 603c is run for PB2 |Bi=1 once and this produces a sequence of length 1 x 12 = 12, corresponding to the number of positions equal to 1 in sequence bl t for example b3 = 001001011011.
Moreover, entry-wise b2 is appended to b1 at the positions where the b1 entries are equal to 0, as shown in table 1300 of figure 13, and entry-wise b3 can be appended to b1 at the positions where the b1 entries are equal to 1 , as shown in table 1400 of figure 14. Finally, the bit pairs are mapped to amplitudes.
The embodiments of the channel encoder 600 described above are configured to operate with two-bit levels. However, further embodiments of the channel encoder 600 can be extended to an arbitrary number of bit-levels guided by the chain rule for probabilities, as will be explained in more detail in the following.
For m bit levels, considering PBm as the desired distribution of m-tuples, the probability of a specific m tuple bm can be written as:
In this representation, level 1 requires one binary distribution PBi and level 2 requires two binary distributions:
, as already discussed above. As for level 3,
it requires four binary distributions and level 4 eight binary distributions, and so forth. The four binary distributions for level 3 are shown below:
PB3 \B1B2 (' | 00)> PB3 \B1B2 (' |01)> B3 \B^B2 ( | 10) , ¾3 |B1B2 (· | H) ·
According to an embodiment, the channel encoder 600 generates each of these binary distributions by a binary distribution matcher. The outputs are then combined by multiplexers that take as third input the positions of 0s and 1 s of higher levels. Figure 15 shows an embodiment of the channel encoder 600 for three bit-levels, which is realized by repeatedly using the same basic building block, i.e. the state-dependent distribution matcher 61 1 a shown in figure 6. In comparison, the basic building block 61 1 a of figure 6 is used for bit level 2.
As already described above, according to an embodiment the basic building block, i.e. the state-dependent distribution matcher 61 1 a of the channel encoder 600 can implement the following processing steps: first, demultiplexing into two-bit sequences (e.g. using the de multiplexer 602); secondly, passing each sequence through a binary distribution matcher 603b, 603c; thirdly, combining the two shaped output sequences by the multiplexer 605 that takes as third input a sequence of state bits that selects entry-wise between the two shaped sequences during multiplexing.
A further embodiment of the channel encoder 600 comprising more than one state- dependent distribution matcher 61 1 a is shown in figure 15, where each implementation of the state-dependent distribution matcher 61 1 a-d is indicated by a dashed box. As will be appreciated from figure 15, for bit-level three, the basic building block of a higher layer or level of the channel encoder 600, for instance the state-dependent distribution matcher 61 1 d, can contain two basic building blocks of a lower layer or level of the channel encoder 600, for instance the state-dependent distribution matcher 61 1 b and 61 1 c. In other words, the two binary distribution matchers of the state-dependent distribution matcher 61 1 d of a higher level are themselves realized by a basic building block each, i.e. a state-dependent distribution matcher 61 1 b, c at a lower level. Thus, in the embodiment of the channel encoder 600 shown in figure 15 at least some of the state-dependent distribution matchers 61 1 a-d are provided in a nested arrangement.
Thus, according to an embodiment, the channel encoder 600 further comprises a second state-dependent distribution matcher 61 1 b-d and the channel encoder 600 is configured to
use the sequence of shaped bits generated by the first state-dependent distribution matcher 61 1 a as the sequence of state bits for the second state-dependent distribution matcher 61 1 b-d. Furthermore, the first binary distribution matcher 603b or the second binary distribution matcher 603c of the first state-dependent distribution matcher 61 1 a can be provided by the second state-dependent distribution matcher 61 1 b-d.
According to a further embodiment, the channel encoder 600 can further comprise a second state-dependent distribution matcher 61 1 b and a third state-dependent distribution matcher 61 1c, wherein the second state-dependent distribution matcher 61 1 b provides the first binary distribution matcher and the third state-dependent distribution matcher 61 1 c provides the second binary distribution matcher of a fourth state-dependent distribution matcher 61 1 d, as illustrated in figure 15.
Moreover, the channel encoder 600 can be configured to use the sequence of shaped bits provided by the third binary distribution matcher 603a as the sequence of state bits for the fourth state-dependent distribution matcher 61 1 d, as illustrated in figure 15.
In the embodiment shown in figure 15, the channel encoder 600 is further configured to use the first subsequence of shaped bits provided by the first binary distribution matcher 603b of the first state-dependent distribution matcher 61 1 a as the sequence of state bits for the second state-dependent distribution matcher 61 1 b and to use the second subsequence of shaped bits provided by the second binary distribution matcher 603c of the first state-dependent distribution matcher 61 1 a as the sequence of state bits for the third state-dependent distribution matcher 61 1 c.
In an embodiment, pairs or tuples of the shaped bits and corresponding pairs or tuples of the sequence of state bits follow a desired joint distribution.
Figure 16 shows a schematic diagram illustrating a corresponding method 1600 for encoding data bits or symbols into shaped bits or symbols having a non-uniform desired distribution according to an embodiment.
The method 1600 comprises the following steps: a step 1601 of generating on the basis of a first subset of a plurality of data bits a first subsequence of shaped bits; a step 1603 of generating on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits; and a step 1605 of generating a sequence of shaped bits on
the basis of the first subsequence of shaped bits, on the basis of the second subsequence of shaped bits, and on the basis of a sequence of state, i.e. state-defining, bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit of the sequence of state bits is equal to 0, or (ii) a respective shaped bit from the second subsequence in case a corresponding state bit of the sequence of state bits is equal to 1 .
While a particular feature or aspect of the disclosure may have been disclosed with respect to only one of several implementations or embodiments, such feature or aspect may be combined with one or more other features or aspects of the other implementations or embodiments as may be desired and advantageous for any given or particular application. Furthermore, to the extent that the terms "include", "have", "with", or other variants thereof are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term "comprise". Also, the terms "exemplary", "for example" and "e.g." are merely meant as an example, rather than the best or optimal. The terms "coupled" and "connected", along with derivatives may have been used. It should be understood that these terms may have been used to indicate that two elements cooperate or interact with each other regardless whether they are in direct physical or electrical contact, or they are not in direct contact with each other.
Although specific aspects have been illustrated and described herein, it will be
appreciated by those of ordinary skill in the art that a variety of alternate and/or equivalent implementations may be substituted for the specific aspects shown and described without departing from the scope of the present disclosure. This application is intended to cover any adaptations or variations of the specific aspects discussed herein.
Although the elements in the following claims are recited in a particular sequence with corresponding labeling, unless the claim recitations otherwise imply a particular sequence for implementing some or all of those elements, those elements are not necessarily intended to be limited to being implemented in that particular sequence.
Many alternatives, modifications, and variations will be apparent to those skilled in the art in light of the above teachings. Of course, those skilled in the art readily recognize that there are numerous applications of the invention beyond those described herein. While the present invention has been described with reference to one or more particular embodiments, those skilled in the art recognize that many changes may be made thereto
without departing from the scope of the present invention. It is therefore to be understood that within the scope of the appended claims and their equivalents, the invention may be practiced otherwise than as specifically described herein.
Claims
1. A state-dependent distribution matcher (61 1 a-d) for a channel encoder (600) for encoding data bits or symbols into shaped bits or symbols having a non-uniform distribution, wherein the state-dependent distribution matcher (61 1 a-d) comprises: a first binary distribution matcher (603b) configured to generate on the basis of a first subset of a plurality of data bits a first subsequence of shaped bits; and a second binary distribution matcher (603c) configured to generate on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits; wherein the state-dependent distribution matcher (61 1 a-d) is configured to generate a sequence of shaped bits on the basis of the first subsequence of shaped bits, on the basis of the second subsequence of shaped bits, and on the basis of a sequence of state bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit is equal to 0, or (ii) a respective shaped bit from the second subsequence, in case a corresponding state bit is equal to 1 .
2. The state-dependent distribution matcher (61 1 a-d) of claim 1 , wherein the state- dependent distribution matcher (61 1 a-d) is further configured to partition a subset of the plurality of data bits into the first subset and the second subset of the plurality of data bits.
3. The state-dependent distribution matcher (61 1 a-d) of claim 1 or 2, wherein the first binary distribution matcher (603b) and/or the second binary distribution matcher (603c) is a constant composition distribution matcher.
4. A channel encoder (600) for encoding data bits or symbols into shaped bits or symbols having a non-uniform distribution, wherein the channel encoder (600) comprises a first state-dependent distribution matcher (61 1 a) according to any one of claims 1 to 3 and wherein the channel encoder (600) is configured to generate a sequence of shaped symbols on the basis of the sequence of shaped bits and the sequence of state bits by mapping pairs or tuples of the shaped bits and the state bits to a plurality of symbols.
5. The channel encoder (600) of claim 4, wherein the channel encoder (600) further comprises a third binary distribution matcher (603a) configured to generate on the basis of a third subset of the plurality of data bits the sequence of state bits.
6. The channel encoder (600) of claim 5, wherein the third binary distribution matcher (603a) is a constant composition distribution matcher.
7. The channel encoder (600) of any one of claim 5 or 6, wherein the channel encoder (600) further comprises a second state-dependent distribution matcher (61 1 b) according to any one of claims 1 to 3 and a third state-dependent distribution matcher (61 1c) according to any one of claims 1 to 3, wherein the second state-dependent distribution matcher (61 1 b) provides the first binary distribution matcher and the third state-dependent distribution matcher (61 1c) provides the second binary distribution matcher of a fourth state-dependent distribution matcher (61 1 d) according to any one of claims 1 to 3.
8. The channel encoder (600) of claim 7, wherein the channel encoder (600) is further configured to use the sequence of shaped bits provided by the third binary distribution matcher (603a) as the sequence of state bits for the fourth state-dependent distribution matcher (61 1 d).
9. The channel encoder (600) of claim 7 or 8, wherein the channel encoder (600) is further configured to use the first subsequence of shaped bits provided by the first binary distribution matcher (603b) of the first state-dependent distribution matcher (61 1 a) as the sequence of state bits for the second state-dependent distribution matcher (61 1 b) and/or to use the second subsequence of shaped bits provided by the second binary distribution matcher (603c) of the first state-dependent distribution matcher (61 1 a) as the sequence of state bits for the third state-dependent distribution matcher (61 1c).
10. The channel encoder (600) of claim 4, wherein the channel encoder (600) comprises a second state-dependent distribution matcher (61 1 b-d) according to any one of claims 1 to 3, wherein the channel encoder (600) is configured to use the sequence of shaped bits generated by the first state-dependent distribution matcher (61 1 a) as the sequence of state bits for the second state-dependent distribution matcher (61 1 b-d).
1 1. The channel encoder (600) of claim 4, wherein the channel encoder (600) comprises a second state-dependent distribution matcher (61 1 b-d) according to any one of claims 1 to 3 and wherein the first or second binary distribution matcher (603b, 603c) of the first state-dependent distribution matcher (61 1 a) is provided by the second state- dependent distribution matcher (61 1 b-d).
12. The channel encoder (600) of any one of claims 4 to 1 1 , wherein pairs or tuples of the shaped bits and corresponding pairs or tuples of the sequence of state bits follow a joint distribution.
13. A transmitter communication device comprising a channel encoder (600) according to any one of claims 4 to 12 for encoding a plurality of data bits for transmission to a receiver communication device.
14. A method (1600) for encoding data bits or symbols into shaped bits or symbols having a non-uniform distribution, wherein the method (1600) comprises: generating (1601 ) on the basis of a first subset of a plurality of data bits a first
subsequence of shaped bits; generating (1603) on the basis of a second subset of the plurality of data bits a second subsequence of shaped bits; and generating (1605) a sequence of shaped bits on the basis of the first subsequence of shaped bits, on the basis of the second subsequence of shaped bits, and on the basis of a sequence of state bits by selecting (i) a respective shaped bit from the first subsequence, in case a corresponding state bit is equal to 0, or (ii) a respective shaped bit from the second subsequence in case a corresponding state bit is equal to 1.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201880098677.XA CN113169809B (en) | 2018-10-25 | 2018-10-25 | Distributed matcher, channel encoder and method for encoding data bits or symbols |
| PCT/EP2018/079285 WO2020083491A1 (en) | 2018-10-25 | 2018-10-25 | A distribution matcher, a channel encoder and a method for encoding data bits or symbols |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/EP2018/079285 WO2020083491A1 (en) | 2018-10-25 | 2018-10-25 | A distribution matcher, a channel encoder and a method for encoding data bits or symbols |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020083491A1 true WO2020083491A1 (en) | 2020-04-30 |
Family
ID=64049228
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/EP2018/079285 WO2020083491A1 (en) | 2018-10-25 | 2018-10-25 | A distribution matcher, a channel encoder and a method for encoding data bits or symbols |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN113169809B (en) |
| WO (1) | WO2020083491A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2801163C1 (en) * | 2019-10-25 | 2023-08-02 | Хуавэй Текнолоджиз Ко., Лтд. | Device for multilevel coding |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3306821A1 (en) | 2016-10-05 | 2018-04-11 | Technische Universität München | Methods of converting or reconverting a data signal and method and system for data transmission and/or data reception |
| US10069519B1 (en) * | 2018-01-23 | 2018-09-04 | Mitsubishi Electric Research Laboratories, Inc. | Partition based distribution matcher for probabilistic constellation shaping |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018077443A1 (en) * | 2016-10-31 | 2018-05-03 | Huawei Technologies Duesseldorf Gmbh | Error detection using symbol distribution in a system with distribution matching and probabilistic amplitude shaping |
| CN110140330B (en) * | 2017-01-02 | 2021-08-13 | 杜塞尔多夫华为技术有限公司 | Apparatus and method for shaping probability distribution of data sequence |
-
2018
- 2018-10-25 WO PCT/EP2018/079285 patent/WO2020083491A1/en active Application Filing
- 2018-10-25 CN CN201880098677.XA patent/CN113169809B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3306821A1 (en) | 2016-10-05 | 2018-04-11 | Technische Universität München | Methods of converting or reconverting a data signal and method and system for data transmission and/or data reception |
| US10069519B1 (en) * | 2018-01-23 | 2018-09-04 | Mitsubishi Electric Research Laboratories, Inc. | Partition based distribution matcher for probabilistic constellation shaping |
Non-Patent Citations (7)
| Title |
|---|
| BOCHERER, GEORG: "High throughput probabilistic shaping with product distribution matching", ARXIV:1702.07510, 2017 |
| F. STEINER; G. BOCHERER; P. SCHULTE: "Approaching waterfilling capacity of parallel channels by higher order modulation and probabilistic amplitude shaping", PROC. CONF. INF. SCIENCES AND SYSTEMS (CISS, 2018 |
| G. BOCHERER; F. STEINER; P. SCHULTE: "Bandwidth efficient and rate-matched low-density parity-check coded modulation", IEEE TRANS. COMMUN., vol. 63, no. 12, 2015, pages 4651 - 4665, XP011593618, DOI: doi:10.1109/TCOMM.2015.2494016 |
| MARCIN PIKUS ET AL: "Bit-Level Probabilistically Shaped Coded Modulation", IEEE COMMUNICATIONS LETTERS., vol. 21, no. 9, 1 September 2017 (2017-09-01), US, pages 1929 - 1932, XP055528044, ISSN: 1089-7798, DOI: 10.1109/LCOMM.2017.2704596 * |
| P. SCHULTE; G. BOCHERER: "Constant composition distribution matching", IEEE TRANS. INF. THEORY, vol. 62, 2016, pages 430, XP011594649, DOI: doi:10.1109/TIT.2015.2499181 |
| PIKUS, MARCIN; WEN XU: "Bit-level probabilistically shaped coded modulation", IEEE COMMUNICATIONS LETTERS, vol. 21.9, 2017 |
| RAMABADRAN; TENKASI V.: "A coding scheme for m-out-of-n codes", IEEE TRANS. COMM., vol. 38, 1990, pages 1156, XP000162507, DOI: doi:10.1109/26.58748 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2801163C1 (en) * | 2019-10-25 | 2023-08-02 | Хуавэй Текнолоджиз Ко., Лтд. | Device for multilevel coding |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113169809A (en) | 2021-07-23 |
| CN113169809B (en) | 2023-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6872296B2 (en) | Communication systems and methods for communicating bit symbols | |
| US10735135B2 (en) | Variable spectral efficiency optical modulation schemes | |
| US20170054483A1 (en) | System and method for designing and using multidimensional constellations | |
| US20190173617A1 (en) | Iterative multi-level equalization and decoding | |
| US10411808B2 (en) | Probabilistically shaped multi-level pulse modulation with gray code mapping | |
| EP3306821B1 (en) | Methods of converting or reconverting a data signal and method and system for data transmission and/or data reception | |
| US10951455B2 (en) | Methods of converting or reconverting a data signal and method and system for data transmission and/or data reception | |
| CN113632397B (en) | Short block length distribution matching algorithm | |
| KR100924526B1 (en) | Coded modulation for partially coherent systems | |
| US10523480B1 (en) | K-bit enumerative sphere shaping of multidimensional constellations | |
| US20120033752A1 (en) | Wireless communication device and wireless communication method | |
| EP0880838B1 (en) | Method of communicating according to a trellis code chosen from a fixed set of baseband signal points | |
| CN112422195B (en) | Design method and application of multidimensional signal modulation circuit based on generalized mutual information | |
| EP0459524B1 (en) | Error control coding arrangement for communication using convolutionally encoded QAM or PSK | |
| WO2016121341A1 (en) | Optical transmitter, optical communication system, and optical communication method | |
| Benedetto et al. | Geometrically uniform TCM codes over groups based on L/spl times/MPSK constellations | |
| WO2020052431A1 (en) | Information modulation-demodulation method and apparatus | |
| CN1799214A (en) | Bit swapping for different interleaving depths | |
| WO2020083491A1 (en) | A distribution matcher, a channel encoder and a method for encoding data bits or symbols | |
| Yao et al. | Polar coded modulation via hybrid bit labeling | |
| US9774420B1 (en) | Reed-solomon coding for 40GBASE-T ethernet | |
| KR20060106041A (en) | A multilevel codes transceiver using semi-bicm scheme | |
| Lo et al. | High dimensional circular trellis-coded modulation | |
| JP3053283B2 (en) | Coded multi-level modulator | |
| Mahesh et al. | Noisy Index Coding with Quadrature Amplitude Modulation (QAM) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18795495 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 18795495 Country of ref document: EP Kind code of ref document: A1 |