Method and apparatus for transmitting binary data
By dividing the M-ASK constellation into M/2 two symbol sets and correlating it with probability, and using a finite binary source for symbol shaping, the problem of complexity of symbol probability shaping in non-Gaussian channels is solved, and high-performance data transmission is achieved.
Patent Information
- Application Number
- CN202280093895.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-04-13
- Filing Date
- 2022-12-16
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art is difficult to effectively implement the probability shaping of symbols in non-Gaussian channels, resulting in insufficient performance of the communication system and high complexity of the method for fitting the Maxwell-Boltzmann distribution.
The M-ASK constellation is adopted and divided into M/2 two symbol sets. Each set is associated with different probability. The probability shaping of symbols is realized through a finite binary source, which simplifies the shaping operation and is suitable for non-Gaussian channels.
High-performance data transmission in non-Gaussian channels is realized, reducing the complexity of shaping operations and the number of binary sources, and improving the channel capacity of the communication system.
Smart Images

Figure CN118975201A_ABST
Abstract
Description
Technical Field
[0001] At least one of the embodiments generally relates to a method for transmitting binary data using an M-ASK constellation (ASK is an abbreviation for "Amplitude Shift Keying") At least one of the embodiments also relates to a corresponding transmitter. Background Art
[0002] In a communication system, a transmitter is connected to a receiver via a communication channel (e.g., an optical fiber). The transmitter typically includes an encoder configured to encode input data (e.g., a bit stream) into symbols belonging to a finite set (called a constellation). One-dimensional ASK (abbreviation for "amplitude shift keying") and two-dimensional QAM (abbreviation for "quadrature amplitude keying") are examples of these constellations. Here, a one-dimensional or two-dimensional constellation means that the symbols take values in R or R2, respectively, where R is a set of real numbers.
[0003] These symbols are then transmitted to a receiver via a communication channel.The receiver comprises a decoder configured to decode the received symbols into output data.
[0004] Communication systems that send uniformly distributed symbols usually incur shaping losses. Therefore, in order to approach the channel capacity, it is known that the transmitter should process the input data to change the probability distribution of the transmitted symbols. More precisely, the input data is processed so that the transmitted symbols have a non-uniform probability distribution that adapts to the communication channel. This operation is called probability shaping and can provide energy savings (also known as shaping gain). For Gaussian channels, it is known that the Maxwell-Boltzmann distribution gives quasi-optimal performance. However, processing the input data to fit the Maxwell-Boltzmann distribution is complex to implement.
[0005] It is therefore desirable to find a method for transmitting binary data that is easy to implement, at the same time high performance, and also adaptable to non-Gaussian channels. Summary of the invention
[0006] At least one of the present embodiments is generally directed to a method of transmitting binary data in a transmitter using an M-ASK constellation, the M-ASK constellation being divided into M / 2 sets of two symbols, each set having an index i and a probability p of transmitting a first symbol of the set. i and the probability of sending the second symbol of the set is 1-p i The method comprises the following steps:
[0007] a) Obtain m-1 bits from a binary source, where m = log 2 M;
[0008] b) in response to the m-1 bits selecting a binary source from a plurality of binary sources, each binary source with index i being equal to p i The probability of the output bit being zero is associated with;
[0009] c) obtaining a bit from the selected source;
[0010] d) obtaining a symbol of an M-ASK constellation associated with a binary word formed by m-1 bits obtained from a binary source and one bit obtained from a selected source;
[0011] e) transmitting the obtained symbols to a receiver via a communication channel.
[0012] This transfer method is easier to implement than a transfer method that attempts to fit a Maxwell-Boltzmann distribution. More precisely, the above method allows the shaping operation to be implemented using a reduced number of binary sources.
[0013] In one embodiment, the given signature is a natural signature, and each set of index i includes the i-th symbol and the i+M / 2-th symbol of the M-ASK constellation, where i∈[1;M / 2].
[0014] In one embodiment, the given label is a Gray label, and each set of index i includes the i-th symbol and the M / 2+1-i-th symbol of the M-ASK constellation, where
[0015] In one embodiment, the m-1 bits obtained from the binary source are the least significant bits of the binary word and the bits obtained from the selected source are the most significant bits of the binary word.
[0016] In one embodiment, the binary source is an equiprobable binary source.
[0017] In one embodiment, selecting a binary source among a plurality of binary sources in response to the m-1 bit comprises:
[0018] - determining the decimal value of the binary sequence formed by the m-1 bits; and
[0019] - Select the binary source with index equal to the decimal value plus 1.
[0020] In one embodiment, the plurality of binary sources includes 2 m-1 A binary source.
[0021] In one embodiment, Multiple binary sources reduced to 2 m-2 binary sources, and selecting the binary source in response to the m-1 bits comprises:
[0022] - determining the decimal value D of the binary sequence formed by the m-1 bits; and
[0023] -exist In the case of , select the binary source with index equal to M / 2-D; and
[0024] - otherwise, select the binary source whose index is equal to the decimal value plus 1;
[0025] And in case a binary source with index equal to M / 2-D is selected, the bits obtained from the selected source are flipped before obtaining the symbols of the M-ASK constellation.
[0026] In one embodiment, Multiple binary sources reduced to 2 m-2 binary sources, and wherein selecting the binary source in response to the m-1 bits comprises:
[0027] - determining the decimal value D of the binary sequence formed by the m-1 bits; and
[0028] - Select the binary source with index equal to D+1mod(M / 2+1), where mod() is the modular operation.
[0029] In one embodiment, by The sequence of bits is matched by binary distribution to obtain each binary source of index i in multiple binary sources from equally probable binary sources, where S max is the number of sources in the multiple binary sources, H(p i ) indicates that the parameter p i The binary entropy of , r is an integer greater than or equal to 1.
[0030] In one embodiment, the method further comprises the following steps:
[0031] - encoding (m-1)*k bits obtained from an equally probable binary source using an error-correcting code into (m-1)*n bits r times, where n and k are integers;
[0032] - obtained from an equiprobable binary source The sequence of bits is matched by binary distribution to obtain each binary source of index i in multiple binary sources from equally probable binary sources, where S max is the number of sources in the plurality of binary sources, and H(p i ) indicates that the parameter p i The binary entropy of ; and
[0033] - Apply a) to e) to each of the r*n (m-1) bit sets.
[0034] In one embodiment, the method further comprises: obtaining the probability p from a table i and sending at least one index entry indicating the obtained probability to the receiver.
[0035] In one embodiment, the method further comprises: estimating the probability p from a predefined communication channel distribution i and sending the estimated probability to the receiver.
[0036] In one embodiment, the method further comprises: receiving a probability p from a receiver i .
[0037] At least one of the present embodiments is generally directed to a transmitter configured to transmit binary data using an M-ASK constellation divided into M / 2 sets of two symbols, each set having an index i and a probability p of transmitting the first symbol of the set. i and the probability of sending the second symbol of the set is 1-p i The transmitter includes at least one processor configured to:
[0038] a) Obtain m-1 bits from a binary source, where m = log 2 M;
[0039] b) in response to the m-1 bits selecting a binary source from a plurality of binary sources, each binary source with index i being equal to p i The probability of the output bit being zero is associated with;
[0040] c) obtaining a bit from the selected source;
[0041] d) obtaining a symbol of an M-ASK constellation associated with a binary word formed by m-1 bits obtained from a binary source and a bit obtained from a selected source;
[0042] e) transmitting the obtained symbols to a receiver via a communication channel.
[0043] The various embodiments disclosed with respect to the method are also applicable to the transmitter.
[0044] A computer program product is disclosed, comprising program code instructions loadable into a programmable device, the program code instructions causing the method according to any one of the aforementioned embodiments to be implemented, the program code instructions being executed by the programmable device.
[0045] A storage medium is disclosed, which stores a computer program including program code instructions, which, when read from the storage medium and executed by a programmable device, cause the method according to any one of the aforementioned embodiments to be implemented.
[0046] The characteristics of the invention will emerge more clearly on reading the following description of at least one example of implementation, which is made with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] [ Figure 1 ] Figure 1 A communication system according to certain embodiments is schematically illustrated.
[0048] [ Figure 2 ] Figure 2 The symbol of the 8-ASK constellation is depicted.
[0049] [ Figure 3A ] Figure 3A The principle of partitioning the M-ASK constellation into M / 2 sets of two symbols according to a specific embodiment is shown.
[0050] [ Figure 3B ] Figure 3B An alternative partitioning of the M-ASK constellation into M / 2 sets of two symbols is shown according to a specific embodiment.
[0051] [ Figure 4A ] Figure 4A A block diagram of a reshaping encoder is depicted in accordance with certain implementations.
[0052] [ Figure 4B ] Figure 4B Depicted is a block diagram of a reshaping encoder according to another specific implementation.
[0053] [ Figure 4C ] Figure 4C Depicted is a block diagram of a reshaping encoder according to another specific implementation.
[0054] [ Figure 5 ] Figure 5 A flow chart of a transmission method according to certain embodiments is depicted.
[0055] [ Figure 6 ] Figure 6 Depicted is a block diagram of a reshaping encoder according to another specific implementation.
[0056] [ Figure 7 ] Figure 7 Depicted is a block diagram of a reshaping encoder according to another specific implementation.
[0057] [ Figure 8 ] Figure 8A flow chart of a decoding method according to certain embodiments is depicted.
[0058] [ Fig. 9 ] Fig. 9 An example of a hardware architecture of a shaping encoder according to a specific embodiment is schematically shown.
[0059] [ Fig.10 ] Fig.10 An example of a hardware architecture of a decoding device according to a specific embodiment is schematically shown. DETAILED DESCRIPTION
[0060] Figure 1 A communication system 1 in which the present embodiment can be implemented is schematically shown. The communication system 1 comprises a transmitter 10 and a receiver 14 connected to each other via a communication channel 12. The transmitter 10 is fed with input data by at least one equally probable binary source S0 and outputs a given symbol alphabet. The input data is, for example, bits of an audio / video bitstream. In an exemplary embodiment, the alphabet is the M-ASK constellation, where M = 2 m , m is an integer. The symbols of the M-ASK constellation are defined as follows:
[0061]
[0062] Therefore, each symbol in this constellation can be represented by m = log 2 M-bit sequence representation. Figure 2 The symbols of an 8-ASK constellation are depicted, where the first symbol is -7 and the last symbol is 7. In the following, various embodiments are described with reference to ASK modulation. However, it will be understood that the present embodiments are not limited to ASK modulation. As an example, it can also be used with QAM modulation.
[0063] Reference Figure 1 , let X be a discrete random variable representing the symbol at the input of the communication channel 12, with probability distribution p(x i )=p(X i =x i ), Let p(y|x i ) is the channel distribution, such that for all x i , Gaussian channel Let Y be a random variable representing the communication channel output. In the case of a Gaussian channel, Y is defined as follows: Y = X + W, where W is Gaussian noise, e.g.
[0064] For a Gaussian channel, the signal-to-noise ratio (SNR) is defined as:
[0065]
[0066] Given any communication channel, let p * (x) is the distribution of input X that maximizes the mutual information (MI) for a given constellation
[0067]
[0068] Where P is the maximum average power. is called the channel capacity, where is maximized not only for the probability distribution but also for all possible inputs. In the following, discrete inputs are considered and only their distribution is optimized. The constellation (i.e., the set of element positions of the discrete input) is not an optimization variable.
[0069] is defined as the set of quasi-optimal distributions, i.e.,
[0070] if where ε is a quantity whose size depends on the requirements of the communication system.
[0071] The goal of probability shaping is to process the input so that its probability distribution maximizes or nearly maximizes the mutual information I(X;Y). In other words, the distribution of the input should be middle.
[0072] For M-ASK constellations, p(x) is often chosen to be the MB distribution (short for “Maxwell-Boltzmann distribution”). In fact, in this case, the performance obtained is close to that obtained with p * (x) The performance obtained (i.e., for small ε, ). However, implementing MB distribution is complicated.
[0073] According to a first specific implementation defined in order to limit the complexity in the case of M-ASK, the set of possible distributions is restricted to the set that can be expressed as follows:
[0074] For all in is a scaling constant used to make the sum of the probabilities equal to 1. We use p i Representation variables Therefore, now relative to the set {p i}, Optimizing mutual information:
[0075]
[0076] Therefore, according to the present principles, the M-ASK constellation is divided into M / 2 sets, each set with index i containing the i-th symbol and the i-th Symbol, for all i, such that In the following, "the i-th symbol" will be interpreted as the symbol of index i. The symbol " will be interpreted as an index Therefore, in the case of 8-ASK, the symbol set is defined as in the following table.
[0077] gather Index of a collection Probability value {-7,1} 1 <![CDATA[p 1 ]]> {-5,3} 2 <![CDATA[p 2 ]]> {-3,5} 3 <![CDATA[p 3 ]]> {-1,7} 4 <![CDATA[p 4 ]]>
[0078] The shaping encoder according to one embodiment is therefore configured to randomly select one of the M / 2 sets with equal probability (ie, α) i , so with probability p i Send set i The first symbol, with probability 1-p i The second symbol is sent. The selection of the set may involve an error correction code. For each set i , both the sender and the receiver know the probability p i .
[0079] According to a second specific implementation defined in order to limit the complexity in the case of M-ASK, the set of possible distributions is restricted to the set that can be expressed as follows:
[0080] For all and in is a scaling constant used to make the sum of the probabilities equal to 1. We use p i Representation variables Therefore, now relative to the set {p i}Optimize mutual information, and
[0081]
[0082] Therefore, according to the present principles, the M-ASK constellation is divided into M / 2 sets, each set with index i containing the i-th symbol and the i-th symbol, that is, the symbol and index of index i The sign of , for all i, such that and Therefore, in the case of 8-ASK, the symbol set is defined as in the following table.
[0083] gather Index of a collection Probability value {-7,-1} 1 <![CDATA[p 1 ]]> {-5,-3} 2 <![CDATA[p 2 ]]> {1,7} 3 <![CDATA[p 3 ]]> {3,5} 4 <![CDATA[p 4 ]]>
[0084] The shaping encoder according to one embodiment is therefore configured to randomly select one of the M / 2 sets with equal probability (ie, α) i , so with probability p i Send set i The first symbol, with probability 1-pi The second symbol is sent. The selection of the set may involve an error correction code. For each set i , both the sender and the receiver know the probability p i .
[0085] In a first embodiment, the probability value p is calculated offline for a given channel distribution. i For example, for a Gaussian distribution of noise in the channel, the probability value p is calculated offline once i , so it is not modified afterwards.
[0086] In a second embodiment, the probability value p is calculated offline for a given parameter range of the communication channel. i , and stored in a table shared between the transmitter 10 and the receiver 14. For example, if the channel distribution is Gaussian, the parameter may be the SNR of the communication channel. Thus, in the table, a predefined p i The value sets are associated with respective predefined SNR value ranges. Thus, the transmitter 10 selects the predefined p from the table in response to the current SNR value. i The value is set and the receiver 14 is informed accordingly.
[0087] In the third embodiment, the receiver 14 estimates the channel profile after receiving the pilot symbols. Then, the receiver 14 optimizes p i The receiver 14 estimates the channel distribution and sends the estimated channel distribution to the transmitter 10, which optimizes p i The value is fed back to the receiver 14. i The optimization step of the value set can be a linear search, that is, define the possible p i The selected p is a subset of the values and the mutual information associated with the estimated channel distribution is calculated. i The value is the value that maximizes the mutual information. In a variant, p is obtained by solving an optimization problem such as Eq. (1) i A set of values, where p i The following constraints apply to the values: i+M / 2 =1-p i .
[0088] like Figure 3A As depicted above, in the case of an M-ASK constellation where each symbol in the constellation is represented by a binary sequence using natural notation (the least significant bit is on the left), the shaping encoder 100 may be as follows: Figure 4A and Figure 4B As shown above, the most significant bit b m is an integer bit. The natural notation of the 8-ASK constellation is as follows
[0089] Table 1 provides.
[0090] symbol -7 -5 -3 -1 1 3 5 7 <![CDATA[Bit level 1 (b 1 )]]> 0 1 0 1 0 1 0 1 <![CDATA[Bit level 2 (b 2 )]]> 0 0 1 1 0 0 1 1 <![CDATA[Bit level 3(b 3 )]]> 0 0 0 0 1 1 1 1
[0091] Table 3
[0092] like Figure 3B As depicted above, in the case of an M-ASK constellation in which each symbol in the constellation is represented by a binary sequence using Gray notation (the least significant bit is on the left), the shaping encoder 100 may be as follows: Figure 4C Gray notation for the 8-ASK constellation is provided in Table 2A below.
[0093] symbol -7 -5 -3 -1 1 3 5 7 <![CDATA[Bit level 1 (b 1 )]]> 0 1 1 0 0 1 1 0 <![CDATA[Bit level 2 (b 2 )]]> 0 0 1 1 1 1 0 0 <![CDATA[Bit level 3(b 3 )]]> 0 0 0 0 1 1 1 1
[0094] Table 4A
[0095] In this case, the shaping bits are b m-1 , so the shaping bits are not the most significant bits.
[0096] In a variant, different Gray notations are used as shown in Table 2B. In Table 2B, the last two rows are reversed compared to Table 2A. In the latter case, the shaping encoder 100 may be as follows Figure 4A As described above, the most significant bit b m is an integer bit.
[0097] symbol -7 -5 -3 -1 1 3 5 7 <![CDATA[Bit level 1 (b 1 )]]> 0 1 1 0 0 1 1 0 <![CDATA[Bit level 2 (b 2 )]]> 0 0 0 0 1 1 1 1 <![CDATA[Bit level 3(b 3 )]]> 0 0 1 1 1 1 0 0
[0098] Table 2B
[0099] The shaping encoder 100 is a part of the transmitter 10. The shaping encoder 100 includes a switch 102 and a symbol mapper 104, which can use natural marking or Gray marking and is configured to implement such as Figure 5 The shaping method of which the flow chart is depicted above. Optionally, the shaping encoder 100 also includes an ECC (English abbreviation for "error correction code") module 106, which is also called a channel coding module. The ECC module 106 usually takes k*(m-1) bits as input and outputs n*(m-1) bits, where k and n are predefined integer values, for example k=500 and n=1000. The higher n*(m-1) is (for a given ratio k / n), the better the performance. On the other hand, the delay increases with n*(m-1).
[0100] exist Figure 4A and Figure 4C The shaping encoder 100 is composed of binary sources S0 to S2. m-1 Feed bits. Let S max +1 is a different binary source (i.e., including S 0 ), that is, S max =2 m-1Source S0 outputs bits with equal probability p(0)=p(1)=1 / 2. Each source Si(i∈[1;2 m-1 ]) is configured with probability p i The output bit is equal to 0, so with probability (1-p i ) outputs a bit equal to 1.
[0101] exist Figure 4B In the above, the shaping encoder 100 is fed with bits by a single binary source S0, which outputs bits with equal probability p(0)=p(1)=1 / 2. In this case, the sources S1 to S2 are obtained from this single source S0, for example using a binary DM (English abbreviation for "distribution matcher"). m-1 , that is, each source S1 to S2 m-1 A. For example, et al. disclose binary DM in a document entitled “Polar coded distribution matching” published in Electronic Express (Volume 55, Issue 9, Pages 537-539, 2019). However, it will be understood that the present embodiment is not limited to this method of obtaining sources S1 to S2. m -1 specific method.
[0102] exist Figure 4B In the case where the shaping encoder 100 does not include an ECC module, for each given source Si, DMi is generated by the source S0. As input and output bits, where S max =2 m-1 , the probability that each bit is equal to 0 is p i , r is a predefined integer value greater than or equal to 1, H(p i ) is a function with parameter p i Therefore, the output of each DMi is regarded as having a probability p i A source Si that outputs bits equal to 0.
[0103] exist Figure 4B The shaping encoder 100 includes an ECC module 106, which encodes k*(m-1) bits into n*(m-1) bits and repeats the operation r times. As input and output bits, where S max =2 m-1 , the probability that each bit is equal to 0 is p i .
[0104] Back to Figure 5, in step S300, (m-1) bits are obtained from source S0. In one embodiment (natural notation), these (m-1) bits are (b 1 ,b 2 ,...,b m-1 ), and forms the LSB bit of the symbol. Since the M-ASK constellation is divided into M / 2 constellations containing the i-th symbol and the A two-symbol set of symbols, So two symbols belonging to the same set have the same LSB bits. Figure 3A , belongs to set 2 Symbols -5 and 3 of have "10" as LSB bits. Thus, obtaining (m-1) bits from source S0 makes it possible to select a set among the M / 2 sets, and thus select a source Si.
[0105] In another embodiment (Gray notation of Table 2B), these (m-1) bits are (b 1 ,b 2 ,...,b m-1 ), and forms the LSB bit of the symbol. Since the M-ASK constellation is divided into M / 2 constellations containing the i-th symbol and the A two-symbol set of symbols, and So two symbols belonging to the same set have the same LSB bit.
[0106] In another embodiment (Gray notation of Table 2A), these (m-1) bits are bits (b 1 ,b 2 ,...,b m-2 ,b m ).
[0107] In step S302, in response to the (m-1) bit obtained from S0, the switch 102 switches between 2 m-1 In fact, among the possible binary sequences of (m-1) bits and 2 m-1 sets (thus, 2 m-1 In the case of natural or Gray marking of Table 2B, the index i of the selected set / source is equal to the (m-1)-bit binary sequence (b 1 ,b 2 ,...,b m-1 ) is the decimal value D(b 1 ,b 2 ,...,b m-1 ) plus 1, that is, 1+D(b 1 ,b2 ,...,b m-1 ). Each symbol in the constellation is represented by a binary sequence using natural notation, with the least significant bit on the left and the value Figure 3A Binary natural marking is shown above for the 8-ASK constellation. In the case of Gray marking (Table 2A), the index i of the selected set / source is equal to the binary sequence (b 1 ,b 2 ,...,b m-2 ,b m ) is the decimal value D(b 1 ,b 2 ,...,b m-2 ,b m ) plus 1.
[0108] In step S304, bits called shaping bits are obtained from the selected source. In the case of natural marking and Gray marking of Table 2B, the obtained bits are the MSBs of the symbols. m .exist Figure 2 In the case of Gray marking of A, the bit obtained is b m-1 Therefore, the (m-1)-bit control switch obtained at S300 selects one source of the output shaping bit (eg, the MSB in the case of natural marking or Gray marking of Table 2B).
[0109] In step S306, in the case of natural marking and Gray marking of Table 2B, the symbol mapper 104 thus generates a symbol to be transmitted corresponding to the binary word formed by the (m-1) bit obtained at S300 as the LSB and the selected bit obtained at S306 as the MSB. In the case of Gray marking of Table 2A, the symbol mapper 104 thus generates the (m-1) bit (b) obtained at S300 as the MSB. 1 ,b 2 ,...,b m-2 ,b m ) and the selected bit b obtained in S306 m-1 The binary word formed corresponds to the symbol to be transmitted. The binary word is thus (b 1 ,b 2 ,...,b m-1 ,b m ), with the LSB on the left.
[0110] In step S308, the obtained symbols are sent to the receiver 14 via the communication channel 12. Steps S300 to S308 may be repeated for the next (m-1) bit set.
[0111] exist Figure 4A or Figure 4C The shaping encoder 100 includes an ECC module 106. In the case where the ECC module 106 outputs n (m-1) bit sets, each (m-1) bit set is independently used to select a source Si, and thus a shaping bit, in step S302. In other words, steps S300 to S308 are repeated for each (m-1) bit set output by the ECC module 106.
[0112] An example is provided below where M = 8 and m = 3. In this case, the various sets are defined in the following table.
[0113] gather Index of a collection (m-1)LSB bits Selected Source {-7,1} 1 00 S1 {-5,3} 2 10 S2 {-3,5} 3 01 S3 {-1,7} 4 11 S4
[0114] Of the two bits "10" obtained from S0 at S300, the set of index 2 is selected at S302 and the switch 102 is therefore positioned to select source S2. In the case where the bit obtained from S2 is "1" at S304, the obtained binary word is "101", and is therefore mapped onto symbol 3 of the constellation at S306. Therefore, at S308, the symbol mapper 104 outputs symbol 3 and transmits it.
[0115] In another example, where the two bits obtained from S0 at S300 are "11", a set of index 4 is selected at 302. The switch is positioned to source S4. In the case where the bit obtained from S4 at S304 is "1", the symbol mapper outputs symbol 7 (corresponding to the binary word "111") at S306. In the case where the bit obtained from S4 at S304 is "0", the symbol mapper outputs symbol "-1" (corresponding to the binary word "110") at S306.
[0116] Another example is provided below where M = 8 and m = 3. In this case, various sets are defined in the following table.
[0117] gather Index of a collection <![CDATA[(m - 1 = 2) bits (b 1 b 3 )]]> Selected Source {-7,-1} 1 00 S1 {-5,-3} 2 10 S2 {3,5} 3 11 S3 {1,7} 4 01 S4
[0118] Of the two bits "10" obtained from S0 at S300, the set of index 2 is selected at S302 and the switch 102 is therefore positioned to select source S2. In the case where the bit obtained from S2 is "1" at S304, the binary word obtained is "110", and is therefore mapped to symbol -3 of the constellation at S306 (see Table 2 with Gray markings). Therefore, at S308, the symbol mapper 104 outputs symbol -3 and transmits it.
[0119] In another example where the two bits obtained from S0 at S300 are "01", the set of index 4 is selected at 302. The switch is positioned to source S4. In the case where the bit obtained from S4 at S304 is "1", the symbol mapper outputs symbol 1 (corresponding to binary word "011") at S306. In the case where the bit obtained from S4 at S304 is "0", the symbol mapper outputs symbol "7" (corresponding to binary word "001") at S306.
[0120] For some channels such as Gaussian channels, p * (x) is symmetric. Therefore,
[0121] Reference Figure 3A , p3 = 1-p2 and p4 = 1-p1. Considering this symmetry, Figure 4A The shaping encoder 100 may be as follows Figure 6 The above is further simplified, where the number of binary sources is divided by 2 and bit flipping is used when selecting some sources. In this case, S max =2 m-2 .
[0122] Assume b 1 ,b 2 ,...,b m-1 is the m-1 bits obtained from the binary source S0 at S300. If The switch selects the corresponding source in S302 as shown in FIG4, that is, selects the index D(b 1 ,b 2 ,...,b m-1 )+1, and the bit obtained from the selected source is not flipped at S304. The switch selects index in S302 source, and the bit obtained from the selected source is not flipped in S304.
[0123] An example of M=8 and m=3 is provided below. In this case, the various sets are represented in the table below. In this case, only 3 binary sources (S0, S1 and S2) are required instead of 5.
[0124] gather Index of a collection (m-1)LSB bits Selected Source {-7,1} 1 00 S1 {-5,3} 2 10 S2 {-3,5} 3 01 S2 (then flip) {-1,7} 4 11 S1 (then flip)
[0125] Of the two bits "10" obtained from S0 at S300, the set of index 2 is selected at S302 and the switch 102 is therefore positioned to select source S2. In the case where the bit obtained from S2 is "1" at S304, the obtained binary word is "101", and is therefore mapped onto symbol 3 of the constellation at S306. Therefore, at S308, the symbol mapper 104 outputs symbol 3 and transmits it.
[0126] In another example where the two bits obtained from S0 at S300 are "11", the set of index 4 is selected. At S302, the switch 102 is therefore positioned to select source S1. In the case where the bit obtained from S1 at S304 is equal to "1", it is flipped (i.e., changed to 0) by the flip module 108. Ultimately, the symbol mapper outputs the symbol "-1" (corresponding to the binary word "110"). In the case where the bit obtained from S1 at S304 is equal to "0", it is flipped (i.e., changed to 1) by the flip module 108. Ultimately, the symbol mapper outputs the symbol "7" (corresponding to the binary word "111"). Therefore, bit flipping is only applied when the sets of indexes 3 and 4 are selected.
[0127] In the variant ( Figure 6 In the embodiment (not shown above), the shaping encoder 100 is fed with bits by a single binary source S0 which outputs bits with equal probability p(0)=p(1)=1 / 2. In this case, according to Figure 4B In the same way as described above, sources S1 to S2 are obtained from the single source S0 using a binary DM (abbreviation for “distribution matcher”). m-2 , each source S1 to S2 m -2 Therefore, in the case where the shaping encoder does not include an ECC module, for each given source Si, DMi is generated by source S0. As input and output bits, where S max =2 m-2 .
[0128] In the case where the shaping encoder includes an ECC module that encodes k*(m-1) bits into n*(m-1) bits and repeats this operation r times, each DMi is generated by the source S0. As input and output bits, where S max =2 m-2 .
[0129] Reference Figure 3B (Gray notation in Table 2A), p1 = p3 and p2 = p4. Considering this symmetry, Figure 4CThe reshaped encoder 100 can be further simplified by dividing the number of binary sources by 2. In this case, the selected binary source is the one with an index equal to D(b 1 ,b 2 ,...,b m-2 ,b m )+1mod(M / 2+1).
[0130] In the case of Gray marking in Table 2B, Figure 4A The reshaped encoder 100 can be further simplified by dividing the number of binary sources by 2. In this case, the selected binary source is the one with an index equal to D(b 1 ,b 2 ,...,b m-1 )+1mod(M / 2+1).
[0131] To further reduce complexity, the number of different binary sources can be reduced by forcing adjacent symbols to have the same probability. For example, for 8-ASK, p 2 is set equal to p 1 . Symbols assigned the same pi select bits from the same binary source. Even in S max When it is smaller (e.g., equal to 2), In practice, for a Gaussian channel, in addition to S 0 Two additional binary sources beyond are sufficient to obtain most of the shaping gain. Therefore, near-optimal performance can be achieved with a smaller number of binary sources. Limiting the number of binary sources provides several advantages, such as:
[0132] Reducing the complexity (in terms of the number of operations required) of the shaping encoder and therefore the transmitter;
[0133] Reduce the complexity of probabilistic optimization;
[0134] · Reduce the time when the transmitter first signals p i signaling between a transmitter and a receiver; and
[0135] When both the transmitter and the receiver obtain p from the table i value, reduces the size of the table.
[0136] Typically, only a single binary source S0 is available. In one example, the binary sources S0 are equally probable. Therefore, Figure 4B Depicted above is obtaining additional binary sources from a single binary source S0. Figure 7 A reshaping encoder 100 is depicted according to a particular embodiment, wherein two additional binary sources are obtained from a single binary source S0. In this embodiment, S max= 2, and bits must be processed in packets (e.g., due to delays or system constraints). Let H(p i ) indicates that the parameter p i First, the source S 0 generate bits. The error correction code module 106 obtains (m-1)*k*r bits and encodes them. The error correction code module 106 outputs (m-1)*n*r bits. The two binary distribution matchers DM1 and DM2 process them in parallel. Bit and bit. DM1 output A sequence of bits where the probability that each bit is equal to 0 is p 1 , DM2 output A sequence of bits where the probability that each bit is equal to 0 is p 2 At the output of the error correction code module 106, each of the n sets of m-1 bits controls the switch 102 to select a shaping bit b at the output of the corresponding DM m . So for each of the n sets of m-1 bits, apply Figure 5 Depending on the selected source, further application may be made as follows Figure 6 The bit flip disclosed above.
[0137] The switch may request a bits and from the second DM where ∈ is a random quantity. In case one source has no more bits available, bits are selected from the other source.
[0138] Once you use about FIG. 4A to FIG. 7 Any of the aforementioned shaping encoders described above obtains M-ASK symbols, and the transmitter 10 transmits them through multiple dimensions such as polarization, time, frequency, and space. For example, the transmitter 10 may group two ASK symbols in a complex symbol, one for the real part and the other for the imaginary part. In this case, M-ASK modulation is also called QAM modulation.
[0139] In various embodiments, the shaping encoder 100 may optionally include an ECC module 106 for encoding the least significant bits. In an example, the ECC module is configured to encode the bits using multi-level polarity encoding. The principle of multi-level encoding is disclosed below.
[0140] Using the chain rule, the mutual information between the input X and output Y of a communication channel can be expressed as follows:
[0141]
[0142] Among them B iDenotes the random variable corresponding to the i-th bit of the tag under consideration.
[0143] A bit level is a i ; Y|B 1 ,…,B i-1 ) described by the channel. When using binary code to send information at this level i, the coding rate should be chosen to be equal to I(B i ; Y|B 1 ,…,B i-1 ) matches. In practice, the back-off applied depends on the code used. The communication channel can also be characterized by a log-likelihood ratio (LLR).
[0144] set up B i = 0 to obtain the symbol set of the constellation, and B i = 1, assuming B 1 =b 1 ,…,B i-1 =b i-1 Given a received symbol y, the LLR is defined as follows:
[0145]
[0146] Since the bits on the first level remain equally probable, we make p(B i =1|B 1 =b 1 ,…,B i-1 =b i-1 )=p(B i =0|B 1 =b 1 ,…,B i-1 =b i-1 )=0.5, and the mutual information is calculated as follows:
[0147]
[0148] Regarding the last bit level used for shaping, we make p(B m =1|B 1 =b 1 ,…,B m-1 =b m-1 )=p i And p(B m =0|B 1 =b 1 ,…,B m-1 =b m-1 )=1-p i , and the mutual information is calculated as follows:
[0149]
[0150] For the considered SNR range, the shaping bit b m No channel code is required for encoding. Mutual information I(Y; B m |B 1 ,……,B m-1 ) is equal to the entropy of the corresponding level, i.e., the communication channel is "clean". This is why the shaping bit b is used in the previously disclosed embodiment. m It is not encoded with a channel code. Therefore, Figure 7 The value of k on can be chosen as
[0151] Then, multi-level polar coding simply means that the individual bit levels are encoded at rate I(Y; B i |B 1 ,…,B i-1 ) is encoded with a polar code. The block length refers to the size of the polar code used.
[0152] At the receiver side, list decoding is used to decode polar codes. List decoding is disclosed in the document “List decoding of polar codes” published by Tal et al. in 2013 in IEEE Transactions on Information Theory (Vol. 61, No. 5, pp. 2213-2226). List decoders perform MAP decoding. Therefore, the decoder takes LLRs as input (see equation (2)), which requires knowledge of the input distribution to be calculated. In other words, the receiver needs to know p i to decode into a list.
[0153] Figure 8 A flow chart of a decoding method according to certain embodiments is depicted.
[0154] In step S400, the index i of the set (and therefore the source) selected at the transmitter side in step S302 is first obtained. If the selection of the set is done by the transmitter in blocks via an error correction code, the index i is obtained by channel decoding (ie, the index blocks of the set are decoded together).
[0155] In step S402, once the index i of the set is decoded, one of the two symbols of set i is detected based on the received y. Most channel decoders perform maximum a posteriori (MAP) decoding. The detection of the transmitted symbols in set i can also perform MAP detection. MAP decoding requires knowledge of the input distribution. In other words, the receiver needs to know p i for efficient decoding.
[0156] In some cases, it is sufficient to perform maximum likelihood (ML) detection in set i (because the reliability is high enough that the a priori probability does not need to be considered). As a result, the channel decoder can use only p i knowledge.
[0157] In step S404, an inverse mapping is applied to recover the bits from the detected symbols. This step is the inverse of step S306. In step S406, the inverse of step S302 is applied. Thus, the switch 102 is positioned to send the shaped bits (i.e., the MSB bits in the case of natural marking or Gray marking of Table 2B) to the appropriate source. Inverse binary DM may be applicable in the case where binary DM is used on the transmitter side.
[0158] In case bit flipping is applied on the transmitter side, bit flipping is applied on the receiver side by the flipping module 108 in step S406. The decoded bits are flipped, where Indicates the value of the decoded least significant bit.
[0159] Fig. 9 An example of a hardware architecture of a transmitter 10 according to a specific embodiment is schematically shown.
[0160] The transmitter 10 comprises, connected via a communication bus 110: a processor or CPU (short for “central processing unit”) 111; a random access memory RAM 112; a read-only memory ROM 113; a storage unit 114, such as a hard disk or a storage medium reader, such as an SD (short for “Secure Digital”) card reader; and at least one set of communication interfaces COM 115 enabling the transmitter 10 to send and receive data.
[0161] The processor 111 can execute instructions loaded into the RAM 112 from the ROM 113, from an external memory (e.g., an SD card), from a storage medium (e.g., an HDD), or from a communication network. When the transmitter 10 is powered on, the processor 111 can read the instructions from the RAM 112 and execute them. These instructions form a computer program that enables the processor 111 to implement the FIG. 4A to FIG. 7 The method described.
[0162] about FIG. 4A to FIG. 7 The described method may be implemented in software by executing an instruction set by a programmable machine such as a DSP (short for “digital signal processor”), a microcontroller or a GPU (short for “graphics processing unit”), or in hardware by a machine or dedicated components (chips or chipsets) such as an FPGA (short for “field programmable gate array”) or an ASIC (short for “application specific integrated circuit”). In general, the transmitter 10 includes a processor adapted and configured to implement the method. FIG. 4A to FIG. 7An electronic circuit for the described method.
[0163] Fig.10 An example of a hardware architecture of a receiver 14 according to a particular embodiment is schematically shown.
[0164] The receiver 14 comprises, connected via a communication bus 210: a processor or CPU (short for “central processing unit”) 201; a random access memory RAM 202; a read-only memory ROM 203; a storage unit 204, such as a hard disk or a storage medium reader, such as an SD (short for “Secure Digital”) card reader; and at least one set of communication interfaces COM 205 enabling the receiver 14 to send and receive data.
[0165] The processor 201 can execute instructions loaded into the RAM 202 from the ROM 203, from an external memory (e.g., an SD card), from a storage medium (e.g., an HDD), or from a communication network. When the receiver 14 is powered on, the processor 201 can read the instructions from the RAM 202 and execute them. These instructions form a computer program that enables the processor 201 to implement the Figure 8 The method described.
[0166] about Figure 8 The described method may be implemented in software by executing a set of instructions by a programmable machine such as a DSP (short for “digital signal processor”), a microcontroller or a GPU (short for “graphics processing unit”), or in hardware by a machine or dedicated components (chips or chipsets) such as an FPGA (short for “field programmable gate array”) or an ASIC (short for “application specific integrated circuit”). Typically, the receiver 14 includes a computer adapted and configured to implement the method. Figure 8 An electronic circuit for the described method.
Claims
1. A method for sending binary data in a transmitter using an M-ASK constellation, wherein the M-ASK constellation is divided into M / 2 sets of two symbols, each set of index i and the probability p of sending the first symbol of the set i and the probability 1-p of sending the second symbol of the set i associated, and among them, Each symbol of the M-ASK constellation is associated with a binary word defined with a given signature, the method comprising the following steps: a) Obtain m-1 bits from a binary source, where m=log2M; b) in response to the m-1 bits, selecting a binary source from a plurality of binary sources, each binary source with index i having a probability of outputting a bit zero equal to p i association; c) obtaining a bit from the selected source; d) obtaining a symbol of said M-ASK constellation associated with said binary word formed by said m-1 bits obtained from said binary source and said bits obtained from said selected source; e) Sending the obtained symbols to a receiver via a communication channel.
2. The method according to claim 1, wherein: The m-1 bits obtained from the binary source are the least significant bits of the binary word and the bits obtained from the selected source are the most significant bits of the binary word.
3. The method according to claim 1 or 2, wherein: The binary sources are equiprobable.
4. The method according to any one of claims 1 to 3, wherein: The step of selecting a binary source from a plurality of binary sources in response to the m-1 bit comprises: - determining the decimal value of the binary sequence formed by said m-1 bits; and - Selecting the binary source whose index is equal to the decimal value plus 1.
5. The method according to any one of claims 1 to 4, wherein: The plurality of binary sources include 2 m-1 A binary source.
6. The method according to any one of claims 1 to 5, wherein: The given signature is a natural signature, and wherein each set of index i includes the i-th symbol and the i+M / 2-th symbol of the M-ASK constellation, where i∈[1;M / 2].
7. The method according to claim 6, wherein: The number of binary sources is reduced to 2 m-2 binary sources, and wherein the step of selecting a binary source in response to the m-1 bits comprises: - determining the decimal value D of the binary sequence formed by said m-1 bits; and -exist In the case of , selecting the binary source with index equal to M / 2-D; and - otherwise, selecting the binary source whose index is equal to the decimal value plus 1; And wherein, in case of selecting said binary source with index equal to M / 2-D, said bits obtained from said selected source are flipped before obtaining the symbols of said M-ASK constellation.
8. The method according to any one of claims 1 to 5, wherein: The given signature is a Gray signature, and wherein each set of index i includes the i-th symbol and the M / 2+1-i-th symbol of the M-ASK constellation, wherein 9. The method according to claim 8, wherein: The number of binary sources is reduced to 2 m-2 binary sources, and wherein the step of selecting a binary source in response to the m-1 bits comprises: - determining the decimal value D of the binary sequence formed by said m-1 bits; and - Selecting said binary source with index equal to D+1 mod (M / 2+1), where mod() is the modulo operator.
10. The method according to any one of claims 1 to 9, wherein: Through The sequence of bits is subjected to binary distribution matching to obtain each binary source of index i in the plurality of binary sources from the binary source, where S max is the number of sources in the plurality of binary sources, H(p i ) indicates that the i The binary entropy of , and r is an integer greater than or equal to 1.
11. The method according to any one of claims 1 to 10, further comprising the steps of: - encoding (m-1)*k bits obtained from said binary source into (m-1)*n bits r times using an error correction code, where n and k are integers; - obtained from the binary source The sequence of bits is subjected to binary distribution matching to obtain each binary source of index i in the plurality of binary sources from the binary source, where S max is the number of sources in the plurality of binary sources, and H(p i ) indicates that the i The binary entropy of as well as - Apply a) to e) to each of the sets of n (m-1) bits.
12. The method according to claim 10 or 11, further comprising: The probability p is obtained from the table i And at least one index entry indicating the obtained probability is sent to the receiver.
13. The method according to claim 10 or 11, further comprising: The probability p is estimated from a predefined communication channel distribution i And the estimated probability is sent to the receiver.
14. The method according to claim 10 or 11, further comprising: The probability p is received from the receiver i .
15. A transmitter configured to transmit binary data using an M-ASK constellation, the M-ASK constellation being divided into M / 2 sets of two symbols, each set of index i being associated with a probability p of transmitting the first symbol of the set i and the probability 1-p of sending the second symbol of the set i associated, and among them, Each symbol of the M-ASK constellation is associated with a binary word defined using a given signature, the transmitter comprising at least one processor configured to: a) Obtain m-1 bits from a binary source, where m=log2M; b) in response to the m-1 bits, selecting a binary source from a plurality of binary sources, each binary source with index i having a probability of outputting a bit zero equal to p i association; c) obtaining a bit from the selected source; d) obtaining a symbol of said M-ASK constellation associated with said binary word formed by said m-1 bits obtained from said binary source and said bits obtained from said selected source; e) Sending the obtained symbols to a receiver via a communication channel.
16. The transmitter according to claim 15, wherein: The given signature is a natural signature, and wherein each set of index i includes the i-th symbol and the i+M / 2-th symbol of the M-ASK constellation, where i∈[1;M / 2].
17. The transmitter according to claim 15, wherein: The given signature is a Gray signature, and wherein each set of index i includes the i-th symbol and the M / 2+1-i-th symbol of the M-ASK constellation, wherein 18. The transmitter according to any one of claims 15 to 17, wherein: The m-1 bits obtained from the binary source are the least significant bits of the binary word and the bits obtained from the selected source are the most significant bits of the binary word.
19. A computer program product comprising program code instructions loadable in a programmable device, which program code instructions cause the implementation of the method according to any one of claims 1 to 14 when the program code instructions are executed by the programmable device.
20. A storage medium storing a computer program comprising program code instructions which, when read from the storage medium and executed by a programmable device, cause the method according to any one of claims 1 to 14 to be implemented.
