KR20070090900A - Method for reducing ambiguity levels of transmitted symbols - Google Patents

Abstract

The present invention is directed to a transmitter and method for transmitting data in a digital communication system, the method comprising generating an original symbol by mapping the bits of the original bit sequence using a modulation constellation, generating at least one counter part symbol from the original symbol or from at least one counter part bit sequence generated from the original bit sequence where a combination of the original symbol and the at least one counter part symbol forms a quasi pilot symbol.

Description

How to reduce the level of multiplicity of transmission symbols {METHOD FOR REDUCING AMBIGUITY LEVELS OF TRANSMITTED SYMBOLS}

The present invention relates to a digital communication system. The present invention is particularly applicable to a communication system for transmitting data through a time varying or frequency changing channel such as a mobile communication system or satellite communication. The present invention is particularly applicable to a communication system for transmitting data through a channel in which noise or interference effects occur.

For transmission over a remote or wireless link, digital data is modulated onto one or more carriers. Conventionally, various modulation schemes, such as amplitude shift keying (ASK), phase shift keying (PSK), and quadrature amplitude modulation (QAM), are used. Known. In all the above-described modulation schemes, for example, a signal modulated with respect to voltage or electric field strength can be expressed as follows.

A bit sequence or data word is represented by a symbol having a complex value A for a predetermined time interval (symbol duration), where

Represents the instantaneous amplitude of the modulated signal,

Represents the instantaneous phase of the modulated signal. An assignment between a bit value combination and a complex value (modulation state) is called a mapping. In general, in a data word consisting of a b-bit bit sequence, a 2 ^b bit sequence is mapped to a 2 ^b complex value.

As the actual transmission channels distort the modulated signal due to phase shift and attenuation, and the channels add noise to the signal, an error occurs in the received data after demodulation. The likelihood of error increases as the data rate increases, i.e., as the number of modulation states increases and the symbol duration decreases. In order to cope with such an error, redundancy can be added to the data, thereby recognizing and correcting a symbol in which an error has occurred. A more economical approach is given by a method of repeating only the transmission of uncorrectable error data, for example hybrid automatic repeat request (HARQ) and incremental redundancy.

In the conventional approach to repeatedly transmitting data conventionally, the same mapping as applied to the first transmission is used again for retransmission. Thus, the complex value representing the repeated data word is equal to the complex value of the original data word. This will be referred to as "simple mapping".

EP 1 293 059 B1 discloses a method for reordering digitally modulated symbols to improve the average reliability of all bits. This can be accomplished by changing the mapping rule of the bits onto the modulation symbols. This patent focuses on the reordering of retransmitted data words in an ARQ system.

WO2004 / 036817 and WO2004 / 036818 disclose a method of achieving a reliability averaging effect for a system in which the original data word and the repeated data word are transmitted over different branches or in combination with an ARQ system. have.

The methods and mechanisms described in the foregoing patent publications will be referred to simply as "Constellation Rearrangement" or "CoRe".

The main difference between wired and wireless communication systems is the operation of the physical channel through which information is transmitted. The wireless or mobile channel essentially changes over time and / or frequency. For good performance in most modern wireless communication systems, the demodulation of data symbols at the receiver requires an accurate estimation of the channel, which is a channel variable that includes perception of the gain of the channel, phase shift, or both characteristics. Mainly measured by. To facilitate this, several kinds of pilot symbols are usually inserted between or within the data symbol streams, which have a certain phase value and / or apparent magnitude that can be used to determine channel variables. This information is then used for correction measures such as adaptive filtering.

The communication channel may experience noise or interference effects. This effect also affects the transmission of the pilot symbols described above. Even if the size and phase characteristics of the channel do not change, the receiver may incorrectly estimate the channel due to noise or interference. For simplicity, this application refers to noise and interference effects as just noise, and it will be apparent to those skilled in the art that what is described below regarding noise can also be applied to interference.

"Decision-feedback demodulation" is an iterative process that uses initial rough channel estimation (or none at all) to demodulate data symbols. After demodulation, preferably after decoding, the obtained information is fed back to the channel estimator for improved estimation resulting from the data symbols. It must be ensured that this process introduces delays and requires a lot of computation in each iteration stage, as well as greatly depends on the quality of the initial rough channel estimate due to the feedback loop. Such procedures are described, for example, in Lutz H.-J., published in IEEE Transactions on Communications, Volume: 49, July 2001, Pages: 1176-1184. Lampe and Rebert Schober's "Iterative Decision-Feedback Differential Demodulation of Bit-Interleaved Coded MDPSK for Flat Rayleigh Fading Channels".

In general, the magnitude and / or phase is unknown before demodulation, so the data symbols themselves cannot be used accurately for channel estimation. The receiver must determine the transmitted symbol based on the received signal before channel estimation is possible. Since there may be an error in recognizing the symbol, ambiguity is introduced in the channel estimation. This behavior can be seen in FIG. 1 and is further detailed in Table 1 to show the number of multiplicity used in the different digital modulation schemes.

It can be readily seen from Table 1 that the performance of the iterative decision-feedback demodulation scheme will greatly depend on the number of multiplicity used in the modulation scheme. Misleading assumptions about the passed symbol lead to incorrect results in channel estimation. In particular, in modulation schemes with a large number of modulation states, errors are likely to occur in symbols due to unavoidable noise. Incorrect channel estimation then results in inaccurate correction, resulting in more errors in the received symbols. Therefore, there is a need to improve the reliability of channel estimation in related fields.

The above-mentioned prior art only shows the aspect of reordering the mapping or averaging the average bit reliability of the bits mapped onto one digital symbol by bit operations before mapping. This has a good effect if the time / frequency variation or noisy channel is known very accurately, but if the coherence time / frequency is relatively small compared to the data packet, the receiver improves the perception of the time / frequency variation channel. Not only does it provide a means to do this, but it does not provide any means to improve the recognition of noisy channels at the receiver.

Summary of the Invention

Accordingly, it is an object of the present invention to provide a method for improving the reliability of channel estimation in a digital transmission system.

It is another object of the present invention to provide a transmitter in a digital communication system to improve the reliability of channel estimation. A special object of the present invention is to completely eliminate phase multiplicity after combining the original symbols with the retransmitted symbols representing the same data.

This object is achieved by defining a special method of mapping repeated data words onto a signal constellation point. The rearranged arrangement pattern is selected to reduce the number of multiplicity when the original data symbol and the repeated data symbol are combined. That is, the number of different results that can be obtained by adding a complex value or vector to the complex plane representing the constellation point of initial transmission and retransmission of the same data word is smaller than the number of original constellation points or modulation states. The number of phase diversity is further reduced to one by using only a subset of all possible modulation states depending on the modulation (mapping) scheme applied for the original symbol and the corresponding symbol (ie, phase diversity is completely eliminated). This subset is chosen such that the complex value (modulation state) representing all modulation symbols included in the subset is in the half plane of the complex plane or in the sub plane of the half plane. For convenience and clarity, this subset is referred to as "phase ambiguity one subset" or simply "PAO subset".

Reducing magnitude and eliminating phase diversity promotes better channel estimation that is less dependent or independent of the actual data symbols transmitted.

To reduce the number of size multiplicity,

1. Determine the magnitude and phase values for each array point in the original array. This can be represented by a complex value.

2. For each array point in the original array, determine one or more complex replica values to meet the following conditions.

a. The coherent combination of original and duplicate complex values for all data words will have a reduced number of magnitude levels compared to the original array.

b. The average transmit power of the duplicate array is equal to the average transmit power of the original array. (Optional)

To eliminate phase multiplicity, proceed as follows.

1. Determine the magnitude and phase values for each array point in the original array. This can be represented by a complex value.

2. For each array point in the original array, determine one or more complex replication values that meet the following conditions.

a. The coherent combination of the original complex value and the duplicate complex value for at least some or each of all data words will have a reduced number of magnitude levels compared to the original arrangement.

b. The average transmit power of the duplicate array is equal to the average transmit power of the original array. (Optional)

3. Select a PAO subset of modulation symbols (array points) to be used for transmission from the original array so that complex values representing all modulation symbols contained within the PAO subset are located in one half plane of the complex plane, The boundary passes through the complex origin 0 + j0 and, for each symbol in the PAO subset, ensure that each duplicate complex value according to condition 2 is contained within the same half plane.

Step b is optional in both cases because it is not necessary for the reduction of versatility. However, step b has the advantage of providing a constant transmit power on the channel through the transmitted and retransmitted signals.

Of course, there must be a one-to-one correspondence for each data word between the original and duplicate arrays. Therefore, the relationship between the original array and the array point in each replica array is obvious but can be arbitrary. In addition, all replica arrays have the same number of array points (different modulation states, differently assigned complex values) as the original array.

A replication array can be created and a PAO subset selected by the following method.

1. Divide the complex plane into two nonoverlapping adjacent subplanes, each containing an array point half.

2. For each subplane, find the point of the average complex value of all constellation points in that subplane.

3. For each subplane, find the copy arrangement by mirroring the constellation point of each subplane to be close to the point of the average complex value.

4. Select a symbol contained in one of the two subplanes as the PAO subset of the symbol to be used for transmission.

Step 4 is not necessary if all possible modulation states of the modulation scheme are already at least within the half plane of the complex plane. This is the case with total modulation, for example, 8-ASK shown in FIG.

In any case, as each system adds noise and distortion to the transmitted signal, the mentioned mirroring is not required to be mathematically accurate but is desirable. In a real system, close mirroring would be sufficient. Proximity means that the distance between the actual constellation point and the ideal mirrored position is less than half of the distance to the closest constellation point representing another value of the data word. Such close mirroring can be advantageously used to represent a fixed point of complex values where the reduced accuracy of the number of fixed points does not represent a mathematically correct solution.

If a constant average transmit power condition is not required, the following more general method may be applied.

1. Divide the complex plane into two nonoverlapping adjacent subplanes, each containing an array point half.

2. For each subplane, find the axis of symmetry for at least some constellation points within that subplane.

3. For each subplane, find the copy arrangement by mirroring the constellation points of that subplane to approximate one predefined point on the axis of symmetry within each subplane.

4. Select one of the two subplanes as the PAO subset of symbols to be used for transmission.

Again, step 4 is not necessary if all possible modulation states of the modulation scheme are already at least within the half plane of the complex plane.

Those skilled in the art will understand that these steps require very simple geometry or calculus techniques.

For arrangements symmetrical to at least one arbitrary axis in the complex plane, preferably the division into two half planes is made for that axis of symmetry which does not contain any signal point. Each axis is used for an array symmetric to the real or imaginary axis, otherwise the axis of symmetry will move up and down.

It is clear that this method can result in duplicate arrangements that differ in shape from the original arrangement if the arrangements are not point symmetrical to the mirroring points in each subplane. This is especially true if the original arrangement represents any mixed ASK / PSK modulation or PSK that is different from QAM. Maintaining the shape of the original arrangement may be advantageous for implementing a receiver's demodulator (LLR calculator), which will not be discussed in further detail here.

In order to maintain the same shape for the cloned array as in the original array, steps 1 to 4 of the cloned array generation should be changed as follows.

1. Divide the complex plane into two nonoverlapping adjacent subplanes, each containing an array point half.

2. Create a duplicate array so that the number of duplicate arrays is one less than the number of subpoint array points.

3. For each subplane in each replica array, change the data word mapping to the array point so that each data word in the original array and the replica array maps exactly to each array point.

4. Select one of the two subplanes as the PAO subset of symbols to be used for transmission.

It will be appreciated that for the same form of original and replica arrays, the complex value representing a symbol included in the PAO subset of the original array is the same as the complex value representing a symbol located within the same half plane of the replica array.

For certain modulation schemes, simultaneously reducing both magnitude and phase diversity is not necessarily required for demodulation. For example, in the PSK scheme, all data information is included in the phase angle of the modulation symbol, but the size does not matter at all. For PSK, the following procedure can be applied to obtain a copy arrangement that eliminates phase multiplicity.

1. Divide the complex plane into two nonoverlapping adjacent subplanes, each containing the same number of constellation points.

2. For each subplane, determine the axis of symmetry for the position of at least some constellation points within that subplane.

3. Obtain a copy arrangement by mirroring the array point of each subplane on the axis of symmetry of this subplane.

4. Select one of the two subplanes as the PAO subset of symbols to be used for transmission.

The mapping of words using the original array, ie the mapping of data words onto complex values according to the original array, results in the original array symbol or simply the original symbol. Similarly, the mapping of data words using duplicate arrays, ie the mapping of data words onto complex values according to the duplicate array, results in duplicate array symbols or simply duplicate symbols.

In an alternative of the invention, the object is achieved by using the same mapping of a plurality of bits (constituting a data word) to a modulation symbol and by using predetermined bit manipulation on each of the plurality of bits for retransmission. To achieve. In a similar manner, the selection of a PAO subset of symbols to be used for transmission is made by replacing at least one of the bits in one word (multiple bits) mapped to a modulation symbol with a fixed value, for example 0 or 1.

According to one aspect of the present invention, a method of transmitting data in a digital communication system comprises the steps of: a) selecting a subset of all available modulation states in a predetermined modulation scheme to be used for transmission, and b) a first plurality; Transmitting a first symbol representing a bit of a first transmission step having a first modulation state contained in the subset, and c) transmitting an additional symbol representing the first plurality of bits, wherein the additional At least one additional transmission step 1206 with each additional symbol state included in the subset. The addition of the complex value associated with the first and the additional modulation state yields the same phase of the complex result for each combination of bit values included in the plurality of bits.

According to another aspect of the invention, a computer readable storage medium stores program instructions which, when executed in a processor of a transmitter of a digital communication system, cause the transmitter to perform the method according to the aspect.

According to another aspect of the invention, a transmitter for a digital communication system is implemented to perform the method of one aspect.

According to another aspect of the invention, a base station for a mobile communication system comprises a transmitter according to the preceding aspect.

According to another aspect of the invention, a mobile station for a mobile communication system comprises a transmitter as defined in the foregoing aspect.

According to another aspect of the invention, a method of receiving data in a digital communication system comprises: a) first and second receiving steps of receiving first and second symbols representing a first plurality of bits, and b) at least said Calculating a likelihood value from the received first and second symbols for a subset of a first plurality of bits; and c) obtaining a possible value for at least one predetermined bit of the first plurality of bits. Determining a value representing an unknown bit value.

According to another aspect of the invention, a computer readable storage medium stores program instructions that, when executed in a processor of a receiver of a digital communication system, cause the receiver to perform the method of the preceding aspect.

According to another aspect of the invention, a receiver of a digital communication system is implemented to perform the method of the aforementioned aspect.

According to another aspect of the invention, a base station for a mobile communication system comprises a receiver as defined above.

According to another aspect of the invention, a mobile station for a mobile communication system comprises a receiver as defined above.

According to another aspect, the present invention provides a method for transmitting data in a transmitter and a digital communication system, the method transmitting a first symbol representing a first plurality of bits, the symbol having a first modulation state Transmitting a first symbol and an additional symbol representing the first plurality of bits, each of the additional symbols having at least one additional transmission state having an additional modulation state, wherein at least one of the at least one of the additional symbols By combining a parameter with at least one parameter of the first symbol, the number of different possible parameter states after the combination is smaller than the number of different parameter states before the combination.

According to another aspect, the present invention provides a method for transmitting data in a transmitter and a digital communication system, the method comprising generating an original symbol by mapping bits of an original bit sequence using a modulation arrangement; Generating at least one duplicate symbol from at least one duplicate bit sequence generated from a bit sequence or from the original symbol, and combining the original symbol and the at least one duplicate symbol to form a quasi-pilot symbol (quasi-). pilot symbols).

According to another aspect, the present invention provides a method of receiving data in a receiver and a digital communication system, the method comprising receiving a first and at least one additional symbol and at least one parameter of the first symbol And obtaining at least one combination of at least one parameter of the at least one additional symbol, and using the at least one combination to obtain an estimate of a communication channel parameter.

The accompanying drawings are intended to explain the principles of the invention and constitute a part of the detailed description of the invention. The drawings will not be construed as limiting the invention to examples illustrating and describing how the invention is made and used. Further features and advantages will become apparent from the following more detailed description of the invention as set forth in the accompanying drawings.

1 shows an overview of various digital modulation mapping arrangements;

2 shows an example of the original data word and repeated data word position for data word number 10 of 16-QAM;

3 shows an example of original data words and repeated data word positions for data word numbers 14 and 39 of 64-QAM;

4 shows the effect of the method described above when applied to QPSK modulation;

5 shows an example of two alternative mappings for original 8-PSK modulation;

6 shows an example of two alternative mappings for the original 16-PSK modulation;

7 and 8 illustrate two alternatives for improving the reliability of channel estimation in case of 8-PSK modulation;

9 shows an example of eight mappings for 16-PSK modulation;

10 (a)-(c) show examples of coherent combination results of the same data word values using two, four or eight different mappings of FIG. 9, respectively;

11 shows an example of a one-dimensional frame structure for pilot and data symbols;

12 illustrates the steps of a data transmission method in a digital communication system;

13 shows an example of a transmitter chain,

14 (a) to (c) show an example of combining original and replica mapping into super-mapping for original 8-PSK modulation;

15 (a) to (c) show an example of combining original and replica mapping into super-mapping for original 16-QAM modulation,

FIG. 16 shows an example of original and replica mapping at 16-QAM resulting in four different combination result values similar to the QPSK modulation state;

17 shows an example of original and duplicate 4-bit sequences in 16-QAM,

18 illustrates steps of a method for improving reliability in estimating transport channel characteristics;

19 illustrates the steps of determining which bits to be inverted and which bits to be replaced by a fixed value to be retransmitted to the PSK;

20 is a diagram illustrating an example in which bits are inverted and retransmitted to 8-PSK;

21 shows the steps of determining which bits to be inverted and which bits to be replaced by a fixed value to be retransmitted to the ASK;

FIG. 22 is a diagram illustrating an example of bit inversion to 8-ASK and retransmission; FIG.

FIG. 23 illustrates the steps of determining which bits to be inverted and which bits to be replaced by a fixed value to be retransmitted to the mixed ASK / PSK.

24 is a diagram illustrating an example in which bits are inverted and retransmitted to 4-ASK / 4-PSK;

FIG. 25 illustrates a 4-ASK part of the modulation scheme of FIG. 24;

FIG. 26 shows a 4-PSK portion of the modulation scheme of FIG. 24;

27 illustrates the steps of determining which bits to be inverted and which bits to be replaced by a fixed value to be retransmitted to square QAM;

FIG. 28 is a diagram illustrating an example in which bits are inverted and retransmitted to 16-QAM; FIG.

FIG. 29 shows an in-phase portion of the modulation scheme of FIG. 28;

30 shows an orthogonal portion of the modulation scheme of FIG. 28;

31 to 34 show examples of non-uniform square QAMs,

35 shows an example of a transmitter chain,

36 illustrates an exemplary structure of a base station;

37 illustrates an exemplary structure of a mobile station;

FIG. 38 illustrates the combination and reversal case of a product resulting in a QPSK-equivalent multivariability situation for the original 4-ASK / 4-PSK;

FIG. 39 illustrates the combination and reversal case of a product resulting in a QPSK-equivalent multivariability situation for the original 16-square-QAM,

40 shows a half plane and half plane bits in the original QPSK according to the present invention;

FIG. 41 shows a half plane and half plane bits in the original 8-PSK in accordance with the present invention; FIG.

42 illustrates a half-plane and half-plane bit in the original 16-QAM according to the present invention;

43 and 44 show examples of half planes in 8-PSK and QPSK;

45 illustrates an exemplary receiver structure;

46 (a) and (b) show a simplified structure of original and duplicate symbol generation and their joint interpretation into a quasi-pilot

47 illustrates a conventional OFDM frame structure including pilot symbols, shared control symbols, and shared data symbols;

48 to 56 illustrate different non-consumable possibilities of how pseudo-pilot symbols are located in an OFDM frame,

57 is a diagram illustrating a process of multiplying a pseudo-pilot component having a spreading code in an element direction;

58 is a diagram illustrating a process of multiplying a pseudo-pilot symbol having a spreading code in the pseudo-pilot direction,

59 is a diagram illustrating a process of spreading a pseudo-pilot component having a spreading code in an element direction;

60 shows an exaggeration of spreading a pseudo-pilot symbol having a spreading code in the pseudo-pilot direction,

61 is a view illustrating a process of constantly phase shifting a pseudo-pilot component in an element direction;

62 shows an example of QPSK representing an original arrangement and a duplicate arrangement when the power combination and phase combination are each determined to be one level;

63 shows an example of 8-PSK representing an original arrangement and a duplicate arrangement when the power combination and phase combination are each determined to be one level;

64 shows an example of 16-QAM representing an original arrangement and a duplicate arrangement when the power combination and phase combination are each determined to be one level;

65 illustrates the use of a different modulation scheme depending on whether a pseudo-pilot symbol or a simple data symbol is used.

66 shows using the same modulation scheme for original symbol, duplicate symbol and simple data symbol;

67 is a flow diagram related to a method for obtaining one or more replica arrangements from an original arrangement when considering power combinations;

68 is a flow diagram of a method for obtaining one or more duplicate arrangements from an original arrangement when considering size combinations;

69 is a flow diagram of a method for obtaining one or more replica arrangements from an original arrangement when considering phase combinations;

FIG. 70 is a diagram showing an example of 4-ASK / 4-PSK representing an original arrangement and a duplicate arrangement when the size combination and the phase combination are each determined to be one level.

In all the figures showing the mapping or arrangement, the points are identified by numeric markers. This labeling refers to any given data word or bit sequence associated with the communication, and the marking itself is fixed but used to represent any data word, so that the sequential markings are binary, octal, decimal, It will be apparent to those skilled in the art that sequential bit sequences should not be represented in connection with hexadecimal notation or other numeric representation.

2 shows an example of transmission using a 16-QAM modulation scheme. According to Table 1, a data transmission symbol carries four bits. In the method described here, these four bits are transmitted twice.

1. Use first array mapping 201 for a four bit original data word.

2. Use the first different arrangement 202 for a four bit repeated data word.

Without loss of universality, the following assumes that the average transmit power of the array is equal to one. The values given in the figures refer to this situation. It will be clear to those skilled in the art how to properly adjust the values when the average transmit power is different from one. It is also apparent how to obtain the transmit power value of the digitally modulated symbol such that the average transmit power of all digitally modulated symbols is 1 or any other value.

To obtain the demodulation arrangement 202 from the original arrangement 201, the complex plane is divided into two non-overlapping adjacent subplanes 204 and 205 along the virtual axis 203. For the arrangement of FIG. 2, the virtual axis is the axis of symmetry. Diagonal line 206 may also be used, but it is advantageous to select a dividing line for two sub-planes where no arrangement point is located thereon. Next, the axis of symmetry for the two sub planes is determined. In the case of FIG. 2, the real axis 207 is the axis of symmetry for the two sub planes. In order to obtain reduced multiplicity after combining the originally transmitted data word with its repeated, the position of the constellation point in the replica array is mirrored to the axis of symmetry from the original constellation point, that is, points 208 and 209 on the real axis 207. Should be. As the subplanes 204 and 205 are divided, all constellation points belonging to the sub plane 204 must be mirrored with respect to the point 208, while all constellation points belonging to the sub plane 205 have to be mirrored with the point 209. Should be mirrored. In order to achieve the same average transmit power in transmission and retransmission, these mirroring points 208 and 209 must be equal to the average of all complex values in each subplane.

In FIG. 2, the constellation point or modulation state of the original mapping and replica array is noticeable for the word number 10. FIG.

In order to eliminate phase multiplicity after fully combining the original and duplicate symbols, one of the subplanes 204 and 205 is selected as the PAO subset to be used for transmission. If subplane 204 is selected, constellation points (modulation states) 9-16 are used for transmission and retransmission. Conversely, if subplane 205 is selected as the PAO subset, constellation points 1-8 are used to transmit and retransmit.

43 and 44 show an example in which division into adjacent sub-planes (here, semi-planes) is possible without overlapping the complex planes. The modulation state on one of the subplanes 4301 or 4302, 4303 or 4304, 4401 or 4402, 4403 or 4404, 4405 or 4406, 4407 or 4408 may be selected as a subset of the modulation states to be used for transmission. The anti-plane 4305 or 4306 is not recommended as the modulation state is located on the line dividing the sub plane.

Since the PAO subset contains only some of the constellation points available in the original modulation scheme, the data to be transmitted must be suitable for the reduced channel capacity. Assuming that the PAO subset contains exactly half of the constellation points available in the original modulation scheme, this is done for example by:

Distributing data bits over a larger number of modulation symbols (e.g. transmitting 3 words with 4 bits on 4 symbols rather than 3 each)

Invalidating one bit per transmitted symbol

Alternatively, higher order modulation schemes can be used, for example 32-QAM instead of 16-QAM.

3 shows a first mapping 301 and an additional mapping 302 for 64-QAM. Here, the complex plane is again divided into two non-overlapping adjacent sub planes along the virtual axis 303. Then, for the second mapping, each constellation point is mirrored to the average complex values 304 and 305 in the same subplane from the original position in the first array to the subplane to which the constellation belongs. The right or left half plane is selected and used for transmission. That is, the PAO subset includes constellation points 1 to 32 or constellation points 33 to 64.

(410)를 유도하는 두 벡터를 더하는 것과 동일하다. 도 4(b)~(d)는 각각 워드 수 "2", "3" 그리고 "4"에 대한 코히어런트 조합을 나타낸다. 도면은 다의성의 개수가 BPSK 변조에서와 유사하게 하나의 크기 레벨과 2개의 위상 레벨(410 및 411)로 감소하는 것을 보여준다. 이는 쉽고 명백하게 전송 채널의 감쇠를 결정하도록 하며, 그 위상이 -π와 +π 사이에서 천이하도록 한다. 사용된 배열점이 단지 "1"과 "2" 또는 "3"과 "4"로 감소함으로 인해 조합 결과가 각각 단독으로 점(410) 또는 점(411)으로 추가 감소할 수 있게 된다.The result of combining the original transmission and the repeated (replicated) transmission of the same arbitrary word is shown in FIG. 4 for the example of QPSK. To obtain a second or additional or replica mapping 402 from the first or original mapping 401, the complex plane is divided into two non-overlapping adjacent sub planes 404 and 405 along the virtual axis 403. In each subplane, the array points are mirrored with respect to the mean values 406 and 407, respectively. The word number "1" is represented by the vector 408 in the first mapping and by the vector 409 in the second mapping. Each vector has a length of 1 as the average carrier size is defined to be 1. Coherent combination of symbols is a mistake

Equivalent to adding two vectors leading to 410. 4 (b)-(d) show coherent combinations for word numbers “2”, “3” and “4”, respectively. The figure shows that the number of multiplicity decreases to one magnitude level and two phase levels 410 and 411, similarly to BPSK modulation. This allows for easy and obvious determination of the attenuation of the transmission channel and its phase shifting between -π and + π. The constellation points used are only reduced to "1" and "2" or "3" and "4" so that the combination result can be further reduced to point 410 or point 411 alone, respectively.

The principle described in connection with FIG. 4 for an example of QPSK can be applied to all QAM arrays in a similar manner, with coherent combinations resulting in one single value independent of the number of modulation states or array points.

If you do not need to keep the original array in shape for a duplicate (or second or additional) array, you can always find a single clone array that fulfills the requirement to completely eliminate multiplicity. An example of such a situation is shown in FIG. 5, where the original (first) mapping follows the 8-PSK scheme. In order to achieve BPSK-equivalence after the coherent combination of the two mappings, the complex plane is divided into adjacent subplanes 504 and 505 that do not overlap along the virtual axis 503. Each constellation point in each subplane is mirrored with respect to the average complex value 506 or 507. For example, the constellation point for word number "1" is mirrored with respect to point 507 versus position 508. The replica (second) mapping 502 is a mixed ASK / PSK arrangement. Again, subplane 504 (points 5-8) or subplane 505 (points 1-4) are selected as a subset of PAOs to be used for both transmission and retransmission to eliminate BPSK-equivalent residual phase diversity. .

6 shows a similar situation for the case where the original (first) mapping is the 16-PSK scheme. If the multiplicity would be removed, the replication (second) mapping would be quite irregular.

If the original command must be maintained in a replica array, more than one replica array will be required to eliminate phase multiplicity. This is especially true for PSK modulation with more than four signal constellation points. The result of the same coherent combination as the example of the replication arrangement for 8-PSK is in FIG. 7, the example for the replication arrangement is in FIG. 9, and the result of the coherent combination for 16-PSK is in FIG. 10 (a). It is shown in (c). As can be seen in the figure, the constellation point or modulation state for all retransmissions is located in the same subplane as the constellation point for the original transmission. Therefore, the constellation points included in one of the subplanes 706, 707, 804, 805, etc. may be selected as the PAO subset to be used for transmission.

Returning to FIG. 7, the complex plane of the original arrangement (first mapping) 701 is divided by two virtual non-overlapping adjacent subplanes 706 and 707 by the virtual axis 705. Within each subplane, if a given data word is mapped to an array point, the order is changed so that once the same word is correctly specified to each position (array point) of that subplane within all mappings 701-704. As a result, the coherent combination of all four transfers of the same word becomes the same value independent of the word value. In FIG. 7, the word number “1” is represented by the vector 708 in the first mapping 701, the vector 709 in the second mapping 702, and the vector 710 in the third mapping 703. In the mapping 704, it is represented by the vector 711. As the same vector is added to all word values in a different order, the result 712 becomes approximately 2.6131 for all word values specified in the right halfplane. Similarly, the actual value of approximately -2.6131 is the resultant value 713 for all values specified in the left half plane. As a result, the multiplicity can be completely eliminated by mapping the word four times to the array point and selecting the modulation state 1-4 or modulation state 5-8 to be used for transmission.

If only phase multiplicity has to be removed for the PSK scheme, then (point (1-4) or point (5-8) (Fig. 8) for 8-PSK, point (1-8) or point (9) for 16-PSK. -16) (in the case of 0 or 180 degrees for the actual axis if Fig. 9 is selected as the PAO subset), as in Fig. 8 or 10 (a) or 10 (b), which already shows only one phase level. It may be sufficient to use only one copy array that results in the same combined result.

In FIG. 8, the complex plane is divided into adjacent sub planes 804 and 805 without overlapping along the virtual axis 803. Instead of mirroring each array position to a point in the first mapping 801, to obtain a position within the second mapping 802, the position is in the actual axis 806, an axis symmetric to the two sub-planes. Mirrored about. The combination of the first (original) transmission of the word number " 1 " and the repeated transmission is the sum of the vector 807 and the vector 808, which is the actual value of approximately 0.7654 at point 812. The same is true for the word number "4". When combining the vector 809 and the vector 810 for the word number "2" or "3", the result is approximately 1.8478 at the point 811.

Even if the multiplicity of size is greater than 1, such a situation will greatly improve the channel estimation capability as the exact size may not be required in the PSK demodulation process.

9 shows eight different mappings for 16-PSK. If only the first and second mappings are combined, four result values are possible for any half of the real axis, as shown in FIG. 10 (a) (four magnitude levels). When the first four mappings are combined, two result values occur for each possible PAO subset, as shown in Figure 10 (b) (two magnitude levels). Only when all eight mappings are combined, the multiplicity is completely eliminated when the set of constellation points used is reduced to constellation points on the right half plane or to constellation points on the left half plane.

The procedure described in the present invention can be interpreted as rearranging the rules for mapping a word (plural bits) to an array point between the original and repeated versions of the word. Therefore, this method is called "Repetition Rearrangement" or hereinafter simply "Rearm".

Not all words in one frame need to be transmitted using an iterative reordering approach as described herein. If the channel is changing slowly, only a few rearm words will be sufficient to create a good channel estimation for the receiver. As a result, other data words may use other known methods, such as transmission without repetition, simple mapping iterations, or constellation rearrangement (CoRe) iterations. The latter is the preferred solution in repetitive situations as the receiver mainly provides fewer bit error rates. Such a repeat alternative is shown in FIG. 11. Data frame 1101 contains data transmitted as conventional in this case with array reordering. In contrast, data frame 1102 contains only data transmitted according to the method presented herein. Data frame 1103 includes data transmitted according to both methods. The data word 1104 transmitted using the first (original) mapping is repeated into the data word 1105 according to the second mapping as described in detail above. The same applies to data word 1106 that is retransmitted to data word 1107.

In order to inform the receiver which part of the data frame follows which repetition strategy, the amount and position of the ReRe data symbols may be explicitly known as additional signals in the control channel or by some parameter from the transmitter to the receiver.

For an optional channel, it is advantageous for the original symbol and its duplicate symbol to be transmitted in contiguous locations within the time frame, since the benefits of repeat reordering depend on the same channel state as possible for the original and duplicate symbols. Alternatively, it would be possible to transmit the original and duplicate symbols simultaneously on different frequency channels of the FDMA system or on different code channels of the CDMA system. It should be apparent to those skilled in the art that these alternatives may be combined. For example, in an OFDM system, the original and duplicate symbols may be sent on adjacent subcarriers, within adjacent time slots, or both. The latter possibility is particularly applicable when there are several duplicate symbols to be sent with the same original symbol, for example three duplicate symbols for 8-PSK. The first duplicate symbol may be sent in adjacent time slots in the same subcarrier as the original symbol, the second duplicate symbol may be sent in the same time slot in subcarriers adjacent to the original symbol and the third duplicate symbol may be sent in the original It may be transmitted in adjacent subcarriers in time slots adjacent to the symbol.

The examples shown in the figure show a mapping arrangement that results in a combined signal point that will lie on the right axis of the graph that primarily represents the actual axis within the complex signal plane. It should be apparent to those skilled in the art that other mappings may be defined that reduce the number of multiplicity without resulting in a combined signal point on the real axis. For example, it is very easy to define a QAM mapping that results in a signal point on a virtual axis. It is likewise possible to define a mapping for the PSK resulting in a straight point inclined at an angle to the actual axis. Any of those mappings may be selected and implemented by the system designer and do not directly affect technical concepts as long as the present invention is contemplated.

This description focuses on modulation arrays that require coherent demodulation. The algorithm described is thus designed to coherently combine the original constellation point and the rearranged constellation point as well. However, it should be evident that the design algorithms and combinations can be easily modified to suit non-coherent methods. For example, for ASK, simple non-coherent detection of carrier size should be possible and the scalar value should be added to the combination.

In the foregoing detailed description, two nonoverlapping, adjacent sub-planes have always been used. As another example of multi-dividing into sub-planes, it can be divided into four non-overlapping adjacent quadrants, each quadrant similar to the quadrant of the complex plane. The replication arrangement for the first mapping shown in FIG. 9 may be the third mapping in the same figure. In such a case, the modulation state contained in one of the four quadrants should be chosen as a PAO subset of the modulation state to be used for all transmissions, for example, a number (1-4, 5-8, 9-12 or 13-16). something to do.

The original and replica mappings with four nonoverlapping contiguous quadrants for 16-QAM are shown in FIGS. 16A-B, respectively. Here again only the modulation states contained in one of the four quadrants can be selected as the PAO subset to be used for all transmissions. Combining the original symbol with the retransmission symbol may be one of the points 1601-1604, depending on the selected PAO subset.

Another criterion when choosing a replication mapping is that the coherent combination should never be the origin of the complex plane. This is simply because the receiver cannot extract any information about the channel state from the combined signal point of complex value zero.

 In another alternative, only a subset of all possible modulation states or a subset of all existing data word values will follow the method described above. This method also reduces the versatility in determining the transmission channel characteristics.

This description assumes that the original data word and the repeated data word each have the same b-bit bit sequence. For simplicity, the mapping is assumed to map b bits to one complex value. Therefore, the original array has a complex value of 2 ^b distinct saphenous arrangement has a complex value of 2 ^b. The original array and one or more copy arrays can be summarized as "super-arrays." This super-array can then represent a "super-mapping" summarizing the original mapping and one or more replica mappings. In such a case, control information representing the original mapping and the replica mapping should be included in the super-mapping or super-array.

A control word is added to each data word. The control word assumes a specific value for each transfer, for example "1" for the first transfer of data words and "2" for the second transfer of the same data word. Super-mapping maps different values of chained control words and data words to modulation states or super-array points. Thus, different mappings from data word values to modulation states are obtained for different values of the control word. If super-mapping is organized in an appropriate manner, different mappings from data word values to modulation states will show the characteristics described above.

Figure 14 (a) shows the original arrangement for 8-PSK as an example, and Figure 14 (b) shows the relevant replication arrangement. For example, constellation point 1401 represents symbol "1" in the first transmission and constellation point 1402 represents the same symbol in the second transmission or retransmission.

It will be appreciated that the difference is limited to different constellation point labels compared to the arrangement shown in FIG. 5. This difference is only a matter of convenience and one skilled in the art will realize that it is a matter of determining whether to number the symbols from 1 to 8 or 0 to 7. From the arrangements shown in Figs. 14 (a) and 14 (b), it contains the array points from both arrays and indicates that the array point was created using the original mapping or the replica mapping, and the major "0" in the label. Or by adding " 1 " to obtain the super-array shown in Fig. 14 (c). Thus, in Figure 14 (c) all points having a label beginning with "0" correspond to the original constellation point and the corresponding mapping and all points with a label beginning with "1" are duplicate arrays and corresponding mappings. Corresponds to

Figure 15 (a) shows the original arrangement for 16-QAM, for example, and Figure 15 (b) shows the replication arrangement associated with it. It can be seen that the difference compared to the arrangement shown in Fig. 2 is limited to different arrangement point labels for the same reasons as described above with respect to Figs. 14 (a)-(c). From the arrangements shown in Figures 15 (a) and 15 (b), it includes the array points from both arrays and indicates that the array point was created using the original mapping or the replica mapping, and the major "0" in the label. Or by adding " 1 " to obtain the super-array shown in Fig. 15 (c). Since the positions of the constellation points are the same and the original and duplicate arrangements only change according to the labeling, each arrangement point in Fig. 15 (c) should indicate two labels. For example, constellation point 1501 represents a value "1" in the first transmission and a value "4" in the second transmission or retransmission. The super-array thus represents the values "01" and "14". Similarly, point 1502 represents "4" in the first transmission and "1" in the second transmission. In the super-array of Fig. 15C, the values " 04 " and " 11 "

All labels beginning with "0" are equivalent to the original constellation point and the corresponding mappings and labels, and all labels beginning with "1" are equivalent to the duplicate array and corresponding mappings and labels.

It will be appreciated that this super-mapping and super-array are entirely similar to the so-called "set partitioning" approach known to those skilled in the art of trellis-coded modulation. An example of this is G. Ungerboeck's "Trellis-coded modulation with redundant signal sets Part I: Introduction," published in IEEE Communications Magazine, Volume: 25, Issue: 2, Feb 1987, Pages: 5-11 and 12-21. "And" Trellis-coded modulation with redundant signal sets Part II: State of the art ".

12 shows a flow diagram for a method that can be used to reduce multiplicity for data symbols in a digital communication system. The method consists of mapping generation step 1201, transmission step 1205 and one or more retransmission steps 1206.

First, a first mapping is created at step 1202. This mapping can be randomly generated according to a particular algorithm or by simply reading from a table stored at the transmitter using this method. This table may be further received from another entity, such as a mobile station or base station to which transmission is directed.

Next, in step 1208, the appropriate PAO set to be used for transmission from all modulation states is selected according to the rules given above. Alternatively this step may be performed after step 1204. Subsequently, in a further step 1203, a second mapping is generated according to one of the algorithms described above. Step 1204 asks if more mappings should be created. In this case, the loop returns to step 1203. Otherwise, the method proceeds to step 1209. The generated mapping can be stored in a table for later use. Therefore, generation step 1201 is not necessarily required for each transmission session or even each transmitted data word. It is also possible to store all used mappings during the generation of the transmitter, for example by firmware download or to receive all mappings from another and store them in a table in memory.

In step 1209, the transmitted data is adapted to the reduced transfer capacity, for example by reordering the bits into a larger number of words or invalidating the bits. In step 1205, the symbol is transmitted according to the first mapping representing the data word. The same data word is transmitted again as a retransmission symbol in step 1206 according to the second mapping generated in step 1203. Step 1207 asks if there are more mappings as data words should be sent. If this is the case, the method repeats steps 1206 and 1207 again. If no additional mapping exists, the method ends this data word transfer. All transfers of the same data word should be advantageously transferred in close proximity in time, but other data words may be transferred in between.

In FIG. 13 the transmitter 1300 is shown to be used for transmitting data according to the method described above.

At the transmitter 1300, the information bit stream to be transmitted is encoded at the encoder 1301. The encoded bit stream is interleaved in the random bit interleaver 1302. In the S / P unit 1303, the bit groups are combined into data words. The number of bits to be combined depends on the number of modulation states available. For example, for 16-QAM, ld 16 = 4 bits are combined into one data word, and for 64-QAM, ld 64 = 6 bits are combined into one data word. In iterator 1304, the data word is repeated for retransmission. The ratio of repeat factor to data word to be repeated depends on the particular version of the method. The generated word is passed to the mapper 1305. The mapper 1305 may operate according to different modes. In simple mapping and thin first mode, the mapper uses only one word-to-array point mapping to map a non-repeatable word or repeated word to a complex symbol. In the array rearrangement mode, the mapper 1305 applies the array rearrangement described in the prior art by applying different mappings to the words produced by the iterator 1304. In the third mode, the mapper 1305 applies the method described herein to the word generated by the iterator 1304. The mapper 1305 is controlled by the mapping control unit 1306 which selects the mapping mode to be applied to the word. When the third mode is selected, the mapper 1305 receives mapping information from the mapping control unit 1306, which may include a memory 1307 for storing a table containing the mapping information. . The mapping control unit 1306 may, in the third mapping mode, perform second and additional mappings (ie, replication mappings or replication arrangements) for retransmissions derived from the first mapping used for the first transmission in accordance with the rules defined above. It is further implemented to select. The mapping may be calculated at runtime according to the rules described above. Alternatively, the mapping may be read from a table in memory 1307 that stores the mapping in advance, depending on the communication signal design.

Various mapping modes as described above may alternatively be used depending on the information provided by the receiving unit or the network. Also, various mapping modes may alternatively be used within a single frame, depending on the predetermined pattern with the frame 1103 shown in FIG. Information about such a pattern and information about the mapping used may be communicated to the receiving unit via control data transmitter 1308 and transmission channel 1312. In addition, the iteration control unit 1309 controls the iteration factor of the iterator 1304 in accordance with the requirements of the mapping control unit 1306. For example, in the third mapping mode, the repetition control unit 1309 receives information about the number of repetitions required for the selected mapping from the mapping control unit 1306.

After mapping, pilot data is added and the frames are combined in pilot / data frame generation unit 1310 before the information is modulated onto a carrier in modulator 1311. The modulated signal is delivered to the receiver via channel 1312.

According to certain implementations, the transmitter 1300 may further include a unit such as an IF stage, mixer, power amplifier or antenna. In view of signal flow, such additional units may be included in channel 1312 as they may add noise to the signal or cause phase shifting or attenuation in the signal.

Units 1301 to 1311 may be implemented within dedicated hardware or digital signal processors. In this case, the processor executes instructions read from a computer readable storage medium such as read-only memory, electrically erasable read-only memory, or flash memory. Implement the method described here. Such instructions may be further stored on another computer readable medium, such as a magnetic disk, optical disk, or magnetic tape, and downloaded to the device prior to use. It is also possible to implement mixed hardware and software.

Alternatively, the present invention may be implemented by one mapping a word (a plurality of bits) into a modulation state with additional bit manipulation steps.

As an example, assume transmission using a 16-QAM modulation scheme as can be seen in FIGS. 17 and 28. According to Table 1, such data symbols carry four bits. In the method described here, these four bits are transmitted twice.

1. Use 16-QAM mapping for the original sequence (4 bits).

2. Use the same 16-QAM mapping for the replication sequence (4 bits).

In general, for any modulation scheme other than pure ASK, the required bit manipulation step is to replace at least one bit with a fixed value to select a subplane according to the method outlined above. As an example, this is illustrated in FIG. 17 for the gray mapping in which the original bit sequence 1010 and the copy sequence 1100 are highlighted. Each 4-bit sequence is mapped to a modulation state of 16-QAM. Since the applied mapping is a gray mapping, the nearest neighbor always has only one bit value different. For example, the modulation state 1701 is assigned to the bit sequence "0000". The four nearest neighbors 1702-1705 are assigned to the bit sequences "0001", "0010", "0100" and "1000".

Each of the four bit sequences is associated with an additional bit sequence obtained by bit inversion as described below. In addition, at least one bit that must be properly selected in both the original bit sequence and the duplicate bit sequence is replaced by a fixed value, for example 0 or 1. Combining the first symbol resulting from the first bit sequence with the additional symbol resulting from the additional bit sequence, phase multiplicity is removed and one of the two possible vector sum results (1706 or 1707) is obtained, which is the fixed of the replaced bits. Depends on the value One or more of these bits carrying the fixed value due to the effect of reducing phase multiplicity to one are referred to as PAO bits.

The flowchart of FIG. 18 illustrates the steps necessary to remove phase multiplicity in transport channel estimation.

At step 1801, a first sequence or a plurality of bits are received. The number of bits in one sequence depends on the number of different modulation states in the applied modulation scheme. For example, for 64-QAM each sequence contains ld 64 = 6 bits. For 8-PSK each of the plurality of bits includes ld 8 = 3 bits.

In step 1802, one or more bits included in the received plurality of bits are replaced by a fixed value. This corresponds to the selection of a PAO subset of modulation states to be used for the transmission described above.

Obviously, if one of these bits is replaced by a fixed value, that bit generally loses the ability to transmit information. Therefore, each of the PAO bits used in the same sequence of multiple bits reduces the number of different modulation states available by a factor of two. For example, if one of the six bits defining a 64-QAM modulation symbol is replaced by a fixed value, only 0.5 * 64 = 32 remaining modulation symbols of the 64 modulation symbols will be generated, which is the remaining five Depends on the bit value of the bit. One bit may be represented by the remaining 50% modulation symbol for the first fixed value of the bit by the first half plane of the complex plane, and the remaining 50% modulation symbol is complex for the second fixed value of the bit. Can be represented by a second half plane of the plane, and if the first and second half planes are adjacent without overlapping each other and the modulation symbol set is separated such that the boundary between the first half plane and the second half plane contains a complex origin 0 + j0, Bits are referred to as "half-plane bits". Examples for QPSK, 8-PSK and 16-QAM are shown in FIGS. 40-42, respectively. In the example on the left, the half plane bits 4001, 4101 and 4201 select vertically separated half planes 4002, 4102, 4202 or 4003, 4103, 4203 depending on their fixed values. In the example on the right, half-plane bits 4004, 4104 and 4204 select vertically separated half-planes 4005, 4105, 4205 or 4006, 4106, 4206 depending on their fixed values.

In step 1803, the first plurality of bits are mapped to one modulation state according to a predefined gray mapping that maps the bit sequence to the modulation state. In step 1804 the first bit sequence is transmitted by modulating the carrier according to the modulation state assigned to the bit sequence in gray mapping.

In each retransmission, a subset of the bits included in the bit sequence is determined for inversion at step 1805. Decision step 1805 may be accomplished, for example, by executing a decision algorithm, receiving data from a peer entity, or simply reading data from memory. In step 1806, an additional plurality of bits are obtained by taking the first plurality of bits from step 1801 and inverting those bits according to one of the inversion rules determined in step 1805. This additional bit sequence is mapped to the modulated state at step 1807 according to the same gray mapping used at step 1803. As will be described further below, the bit replaced with a fixed value in step 1802 is a modulation state in which an additional plurality of bits are mapped in step 1807 to the modulation state selected through the bit operation in step 1802. It is selected to be included in the PAO subset. In step 1808 the first sequence is retransmitted by transmitting the additional sequence obtained in step 1806, ie by modulating the carrier according to the modulation state obtained in step 1807.

Step 1809 asks whether to retransmit the same first bit sequence. If retransmission is needed, the method returns to box 1805. If not, the method ends and transmission and retransmission of the first bit sequence are completed.

As discussed above, in decision 1805 one inversion rule is selected to obtain an additional bit sequence. This inversion rule can be represented as a subset of bits to be inverted. Depending on the mapping method chosen, there may be one or several inversion rules needed to reduce the versatility to the required target level. Decision step 1805 should preferably select one of those rules for each retransmission so that each inversion rule can be determined once for a given first plurality of bits. The half-plane bits selected for use in reducing one phase multiplicity (i.e., according to the PAO bits defined above) cannot be selected as bits to be inverted in the copy sequence, and vice versa. In the following, the determination of the inversion rule to be selected in step 1805 and the appropriate PAO bit selection in step 1802 will be described in more detail with reference to different modulation schemes.

The following algorithm shown in FIG. 19 may be applied to PSK modulation using gray mapping.

Let n be the number of bits mapped to one PSK symbol (step 1901).

Select n-1 bits to be inversion candidates from n bits (step 1902).

Inversion Rule: Determine the bits to be inverted by getting all possible combinations using all n-1 bits from 1 out of the selected n-1 bits (step 1903).

Obtain an n-1 duplicate bit sequence from the original bit sequence by inverting the bits from the combination obtained earlier.

One antiplane bit that is not selected for inversion is a PAO bit, i.e. a half plane bit to be replaced with a fixed value (step 1904).

For example, the arrangement shown in FIG. 20 will be described.

3 bits are mapped to one symbol using 8-PSK. -> n = 3

-First and third bits are selected as inversion candidates.

Inversion rule: Invert only the first bit, only the third bit, and both the first and third bits.

The half-plane bits are first and second bits.

Since the first bit is used to generate a duplicate array in the inversion rule, the second bit is selected as the PAO bit and replaced by a fixed value of zero or one.

The modulation state 2001 is assigned to the bit sequence "000". By applying the inversion rule, the bit sequences " 100 ", " 001 " and " 101 " to which the modulation states 2002-2004 are to be assigned are obtained. The symbols are combined by adding a vector 2005-2008 representing the complex value of the carrier for this modulation state. The result is a point 2009 for the fixed PAO bit value 0 and a point 2010 for the fixed PAO bit value 1. Therefore, the result can have only one magnitude value and one phase value.

For all manners that require PSK, at least in part, (e.g., n-PSK, n-ASK / m-PSK, n-QAM, as outlined above), at least part of the information is included in the phase of the information symbol. The number of multiplicity can be completely eliminated.

As shown in FIG. 22, the following algorithm shown in FIG. 21 may be applied to ASK modulation in which the transmission powers of symbols are classified in ascending or descending order according to gray coding.

Let n be the number of bits mapped to one ASK symbol (step 2101).

Inversion rule: Invert exactly one bit carrying the same bit value for exactly 0.5 * 2 ⁿ = 2 ^n-1 symbols at the lowest full power (step 2102).

Obtain a duplicate sequence by applying an inversion rule to the original bit sequence.

It will be appreciated by those skilled in the art that the same inversion bit may alternatively be identified as the bit carrying the same bit value for exactly 0.5 * 2 ⁿ = 2 ^n-1 symbols at the highest transmit power.

As an example, 8-ASK-modulation with the mapping of FIG. 22 is considered. In Fig. 22, bars 2201, 2202 and 2203 indicate that bits 1, 2 and 3 each have a value of "1". The assumed bit order is b ₁ b ₂ b ₃ .

Using 8-ASK, 3 bits are mapped to one symbol. -> n = 3

The bit carrying the same value for the smallest transmit power symbol of exactly 0.5 * 2 ³ = 4 is the second bit b ₂ , which is 1 for such a symbol.

Inversion rule: Invert the second bit b ₂ .

Original bit sequence in gray coding: 011, 010, 110, 111, 101, 100, 000, 001

Replica sequence to invert the second b ₂ : 001, 000, 100, 101, 111, 110, 010, 011.

Modulation state 2204 is assigned to bit sequence " 011. " By inverting the second bit according to the inversion rule, a copy sequence " 001 " is obtained. The modulation state 2205 is assigned to the copy sequence "001". The symbols are combined by adding vectors 2206 and 2207 representing the complex values of the modulation states 2204 and 2205. By calculating the result of the combination of all the first bit sequences having a copy sequence, it becomes apparent that the result is always a point 2208. In this case, therefore, no versatility remains in determining the transport channel characteristics.

It is not necessary to replace the bits with a fixed value for pure ASK modulation because all modulation states are in one halfplane and the inversion procedure outlined above can completely eliminate the multiplicity.

As shown in Fig. 24, for mixed ASK / PSK modulation in which bits can be separated into bits carrying gray coded ASK information and bits carrying gray coded PSK information (" star QAM "), such bits Should be handled separately in accordance with the PSK or ASK rules mentioned above. The algorithm accordingly is shown in the flowchart of FIG.

Separate ASK / PSK modulation into separate ASK and PSK portions (step 2301).

Determine separate inversion rules for the ASK and PSK parts according to the algorithms passed.

Determine which ASK / PSK bits correspond to inversion rule bits from the ASK portion (step 2302) and the PSK portion (step 2303).

PSK anti-planar bits not selected for inversion in the PSK portion are selected as PAO bits to be replaced with fixed values (step 2304).

Determine the ASK / PSK inversion rule by combining all ASK / PSK inversion rule bits from 1 (step 2305).

Obtain all duplicate sequences by inverting bits according to the determined ASK / PSK inversion rule.

As an example, consider the star-QAM of FIG.

Using 4-ASK / 4-PSK as shown in FIG. 24, the first two bits 2401 and 2402 are mapped as PSK and the last two bits 2403 and 2404 are ASK-> nASK = 2, nPSK Mapped as = 2.

ASK part (see FIG. 25):

The bits that carry the same value for the smallest transmit power symbol of 0.5 * 2 ² = 2 are the first bit 2403 with zero for those bits.

Inversion rule: Invert the first ASK bit 2403.

Original ASK bit sequence in gray coding: 00, 01, 11, 10

Duplicate sequence to invert first bit 2403: 10, 11, 01, 00

PSK part (see FIG. 26)

The second bit 2402 is selected for inversion.

Inversion rule: Invert the second PSK bit 2403.

Original bit sequence in gray coding: 00, 01, 11, 10

Duplicate sequence to invert second bit 2403: 01, 00, 10, 11

Determining the ASK / PSK inversion rule bits:

The first bit of the ASK portion 2403 is the third bit of the ASK / PSK portion.

The second bit of PSK portion 2402 is the second bit of ASK / PSK portion.

The anti-planar bits in the PSK portion are the first and second PSK bits.

Since the second PSK bit has been selected for inversion, the first bit 2401 of the PSK portion is selected as the PAO bit to be replaced with a fixed value of zero or one.

Determine the ASK / PSK reversal rule.

Inversion rule: Invert only the second ASK / PSK bits 2402, only the second ASK / PSK bits 2403, or the second and third ASK / PSK bits 2402, 2403.

The modulation state 2405 is assigned to the bit sequence "0010". PSK sub-sequence is "00" and ASK sub-sequence is "10". In accordance with the above rules, there is one bit 2402 determined to be inverted from the PSK sub-sequence and one bit 2403 determined for inversion from the ASK sequence. Thus, there are three duplicate bit sequences. Only bit 2402 is inverted to produce " 0110 " to which modulation state 2406 is assigned. Only bit 2403 is inverted to produce " 0000 " to which modulation state 2407 is assigned. Both bits 2402 and 2403 are inverted to produce "0100" corresponding to modulation state 2408. If all the symbols are combined by summing the vectors 2411-2414 representing the corresponding complex values, the result is point 2409. If this calculation is made for all possible combinations of values in the bit sequence, the combined result is a point 2409 for a fixed PAO bit value 0 for bit 2401, and a fixed PAO bit for bit 2401. The value 1 is the point 2410. Thus, multiplicity is completely eliminated.

A special method of mixed ASK / PSK modulation is to combine two orthogonal gray coded m-ASK / 2-PSK modulation. This mixed array is sometimes called "square QAM" and hereinafter simply called sq-QAM. Instead of processing two ASK / PSK modulations separately, a more effective method is introduced here with reference to FIGS. 27 and 28.

Split sq-QAM into two orthogonal m-ASK / 2-PSK modulations, hereinafter referred to as AP1 and AP2 (step 2701).

AP1 inversion rule: The bits to be inverted are bits with the same bit value for exactly m / 2 symbols with the smallest transmit power of the m-ASK portion. This is technically identical to the m symbol of m-ASK / 2-PSK with the smallest transmit power.

AP2 Inversion Rule: The bits to be inverted are bits that carry 2-PSK partial information (step 2703).

Determine which bits in sq-QAM correspond to separate AP1 and AP2 inversion bits.

Obtain the sq-QAM inversion rule by combining the AP1 and AP2 inversion rules for the corresponding QAM bits (step 2704).

Select a bit that carries the PSK information (ie, half-plane bit) of AP1, which will be the PAO bit, that is to be replaced with a fixed value (step 2705).

Obtain the sq-QAM replication sequence by applying the sq-QAM reversal rule.

Those skilled in the art will appreciate that the same inversion bit for AP1 may alternatively be recognized as a bit carrying the same bit value for exactly m / 2 symbols with the highest transmit power of the m-ASK portion.

It should be noted that as in the examples of FIGS. 28 and 31-34, the in-phase component may be selected as AP1 or AP2 with quadrature components, each of which is different for the arrangement. This does not cause a difference in the multiplicity reduction effect. In one case, the result of the combination has an actual value and in other cases a hypothetical value. In the case of any square-QAM, the half-plane bit selected from AP1 to the PAO bit is also a square QAM half-plane bit, in particular an in-phase or co-phase half-plane 4202, which depends on its value, as can be seen in FIG. It will be further appreciated that it may be a half plane bit 4201 or 4204 representing 4203, 4205 or 4206.

Also, two components that are orthogonal to each other but not parallel to either the real axis or the virtual axis may be selected as AP1 and AP2, respectively.

Yes:

Using 16-sq-QAM as in FIG. 28, AP1 is defined as 2-ASK / 2-PSK in FIG. 29 and AP2 is defined as 2-ASK / 2-PSK in FIG.

AP1:

The bits that carry the same value for the smallest transmit power symbol 2901 or 2902 of exactly m / 2 = 1 of the ASK portion are the second ASK / PSK bits 2803, which are zero for those symbols (Fig. 29).

Inversion rule AP1: Invert the second ASK / PSK bit 2803.

-Harmonization of AP1 and AP2 inversion rule bits for the original QAM bits (see Figure 28)

The second ASK / PSK bit 2803 from AP1 corresponds to the third QAM bit.

The first ASK / PSK bit 2802 from AP2 corresponds to the second QAM bit.

Find the 16-sq-QAM inversion rule: Invert the second and third sq-QAM bit modes.

Select the phase bit 2801 (= half-plane bit) of AP1 as a PAO bit, i.e., a bit to be replaced with a fixed value of 0 or 1 (see FIG. 29). This bit corresponds to the first QAM bit, which defines an in-phase half plane.

Original sq-QAM bit sequence:

0000, 0001, 0011, 0010, 0100, 0101, 0111, 0110 or 1100, 1101, 1111, 1110, 1000, 1001, 1011, 1010

A duplicate sq-QAM sequence that inverts the second and third bits:

0110, 0111 0101, 0100, 0010, 0011, 0001, 0000 or 1010, 1011, 1001, 1000, 1110, 1111, 1101, 1100

As an example, the first bit as the PAO bit is set to a fixed value "1". The modulation state 2805 is assigned to the bit sequence " 1011 ". The replica " 1101 " is obtained by inverting the second and third bits and is associated with the modulation state 2806. The combination of the two symbols is achieved by adding vectors 2807 and 2808 representing the complex value of the modulation state. The result is point 2809. By repeating this calculation for all possible combinations of values in the bit sequence, every bit sequence having a fixed value of 1 for bit 2801 produces a result of a combination equal to point 2809 and a fixed value of 0 for bit 2801. All bit sequences having the same combination result as point 2810. Thus, in both cases, multiplicity is eliminated.

Sometimes it should be noted that the term "square QAM" strictly applies only for QAM mappings where the distance between the nearest neighboring points is equal for all points in the array. However, those skilled in the art will appreciate that the algorithms presented herein can also be applied to QAM mappings where this feature is valid only for a subset of points. Examples are the heterogeneous 16-QAM and 64-QAM arrangements used for the DVBs shown in FIGS. 31 to 34. In this arrangement, the real and imaginary axes are axes that are symmetric about the constellation point representing the complex value of the modulation state. Therefore, the term "square QAM" is used here in a broad concept and is not limited to the arrangement as shown in Figures 28 and 31-34.

Those skilled in the art will appreciate that a communication system or apparatus may employ different methods to actually determine the inversion rule. In one embodiment the inversion rule is obtained by executing the algorithm described in the present invention. In a preferred embodiment the inversion rule is determined for each modulation scheme used in the communication system or device and stored in a memory or look-up table to quickly obtain the inversion rule. In another preferred embodiment, the inversion rule is encoded in a hardware or software module, where step 1805 controls which hardware or software module to select during transmission.

Some algorithms will generate more than one replication sequence or inversion rule. This means that more than one bit sequence repetition is necessary for an optimized multilevel reduction, that is, a bit sequence must be transmitted three or more times. If this is not required in terms of system capacity, one of the replication sequence / inversion rules should be chosen. Unoptimized size reduction or phase elimination alone may be considered sufficient. Thus, fewer than optimal replication sequences may be sufficient.

The algorithms described so far have assumed that an optimal multilevel level reduction is obtained by combining the complex values of the first and additional plurality of bits mapped to the modulation state. However, it may be desirable or sufficient to define a sub-optimized size multiplicity level reduction as a target. For example, it may be desirable to reduce the multiplicity to four ASK-equivalent levels, meaning four magnitude levels and one phase level. Channel estimation for this is generally inferior to the situation of single result complex values, but may be advantageous in terms of demodulated LLR values for data bits transmitted in multiple bits or in terms of reducing transmission capacity loss.

The algorithm provided in the ASK results in exactly one magnitude level given the modulation state of exactly 2 ^n-1 with the lowest transmit power in step 2102 and n bits per sequence (compare FIG. 21 and FIG. 22). This algorithm can be extended to any number of target size levels with power of two. Let ^2k be the target number of the size level. Then, the procedure for finding the inversion rule proceeds as follows.

Determine the bit for inversion that has the same first value for the 2 ^nk−1 modulation state with the lowest transmit power and the value opposite to the first value for the next modulation state with the next highest transmit power value.

Or as already described, alternatively:

Determine the bit for inversion that has the same first value for the 2 ^nk−1 modulation state with the highest transmit power and the value opposite to the first value for the next modulation state with the next lowest transmit power value.

The same strategy as described above for k = 0 and as in block 2102 of FIG. 21 is obtained. There is no possible magnitude level reduction for k = n. Thus k may preferably have an integer value ranging from 0 to n-1.

For example, applying k = 1 to the array of FIG. 22 where n = 3, the two array points " 011 " and " 010 " have the same bit values b ₁ = 0 and b ₂ = 1. However, since b ₂ is 1 for not only the two lowest transmit power points but also the four lowest transmit power points, "the same first value is obtained for the 2 ^nk-1 modulation state with the lowest transmit power, and the next highest transmit. The next modulation state with a power value does not fulfill the requirement to have a value opposite to the first value. Therefore, bit b ₁ is determined as the bit to be inverted in the inversion rule.

For the PSK modulation scheme, an inversion rule set is obtained. By selecting only a subset of these inversion rules the multiplicity in phase can already be reduced. In the example of FIG. 20, the inversion to the first bit alone results in two phase levels after the combination, and the combination of symbol 2001 and symbol 2002 and the combination of symbol 2003 and symbol 2004 are two different points. But both are on the virtual axis and share the same phase level. Overall this inversion rule results in a combination of two phase levels and two magnitude levels equivalent to 2-ASK / 2-PSK. Similarly only the inversion of the third bit results in a QPSK-equivalent combination. The symbol 2001 combined with the symbol 2003 has the same magnitude level as the symbol 2002 combined with the symbol 2004. In total, the inversion of only the third bit results in a combination of one magnitude level and four phase levels.

In these cases to completely eliminate phase multiplicity, the number of half-plane bits that should be used as PAO bits with a fixed value depends on the number of phase levels that only inversion rules can achieve. If the result achieved by the inversion rule includes two phase levels, then it is sufficient to set one half-plane bit to the PAO bit. If the result achieved by the inversion rule includes four phase levels, then it is required to specify two half-plane bits as PAO bits. In general, the number of PAO bits needed for the removal of phase multiplicity is a dual logarithm (log to base 2) of the number of phase levels obtained from the inversion rule used. It will be appreciated that the fixed bit value of the first PAO bit and the fixed bit value of the second PAO bit may be selected independently. Of course, the more PAO bits used, the greater the loss of transmission capacity.

Obviously the above strategy for reducing the magnitude or phase level for ASK and PSK can also be applied to mixed ASK / PSK. In the example of FIG. 38, the 4-ASK portion is modified by inverting the first ASK bit to reduce the number of magnitude levels from four to one. As the 4-PSK portion is not modified, the only inversion rule is usually the inversion of 4-ASK / 4-PSK bit number 3, which is the same as 4-ASK bit number 1. The combination results in one magnitude and four phase levels equivalent to QPSK.

As an example, vector 3801 represents the constellation point for bit sequence " 0010. " The first ASK bit becomes the third bit in the sequence. The inversion rule therefore determines to invert the third bit, which produces the bit sequence "0000" represented by the vector 3802. The combination of the two transmissions produces a value 3803. Another possible combination result for other values of the bit sequence is the values 3804, 3805 and 3806. To completely eliminate versatility, both the first and second bits must be fixed. Depending on the combination of these fixed values, one of the combination results 3803, 3804, 3805, and 3806 is obtained.

For square-QAM or sq-QAM, sub-optimized multiplicity level reduction can be achieved if the AP1 or AP2 reversal rule is modified. As outlined above, for a combination of one magnitude and two phase levels, the AP1 inversion rule is equivalent to reducing the versatility for the m-ASK portion and the AP2 inversion rule reduces the versatility for the 2-PSK portion. Same as For sub-optimal combinations with more than one size level, the reduction for the m-ASK portion of AP1 must follow the extended algorithm outlined above to reduce the n size level of ASK to 2 ^k size levels. For sub-optimal combinations with more than two phase levels, the AP2 reversal rule for reducing the 2-PSK portion is for reducing the m-ASK portion of AP2 to 2 ^k as outlined in the extended algorithm above. It must be replaced with a reverse rule. Of course, it should be mentioned that the k value for AP1 may be different from the k value for AP2. See the discussion above for the number of PAO bits required to completely remove the idolatry.

In the example of FIG. 39, the combination of one magnitude level and four phase levels is shown as follows.

Apply the AP1 inversion rule to the 2-ASK portion to invert the second bit 2803 among the two AP1 ASK / PSK modulation bits (compare with FIG. 29).

A modified AP2 inversion rule is applied to the 2-ASK portion to invert the second bit 2804 of the two AP2 ASK / PSK modulation bits (compare with FIG. 30).

Resulting inversion rule: Invert the third and fourth 16-sq-QAM bits b ₃ and b ₄ corresponding to the second bits AP1 and AP2, respectively.

To completely eliminate phase diversity, the two half-plane bits b ₁ and b ₂ of 16-QAM must be fixed.

As an example, the bit sequence “0010” is represented by the vector 3901. The AP1 inversion rule determines the third bit b ₃ (the second bit of b ₁ and b ₃ ) of the bit sequence as the bit to be inverted. The AP2 inversion rule is the bit to be inverted as the fourth bit b ₄ (b ₂ and b ₄ 2nd bit). The resulting bit sequence for the second transmission (or retransmission) is 0001 "represented by the vector 3902 in the complex plane of the modulation state. The combination of the two modulation states achieved by the addition of the vectors 3901 and 3902. Produces a complex point 3403. Similarly, for the bit sequence " 0011 " represented by the vector 3904, the bit sequence for the second transmission is " 0000 " represented by the vector 3905. Both values The combination of s again produces a complex value 3403. Another possible combination result for other bit sequences is the points 3906, 3907, and 3908. To completely remove phase polymorphism, the semi-planar bits (ie, the first two bits) ) Should be fixed as the PAO bit, which would mean selecting one of four quadrants as a PAO subset of the modulation state to be used for transmission.

The original arrangement may be different than shown in the examples. However, the procedure outlined above can still be used as long as the bit sequence mapping conforms to the gray coding / mapping power.

As described above, not every bit sequence in a frame should use the approach as shown in the present invention. This also applies to bit manipulation implementations.

35 shows a transmitter 3500 that can be used to transmit data in accordance with the method described above.

At the transmitter 3500, the bit stream to be transmitted is encoded at the encoder 3501. The encoded bit stream is interleaved in the random bit interleaver 3502. In the S / P unit 3503, the bit groups are combined into bit sequences (plural bits) which are later represented by one transmitted symbol. The number of bits to be combined depends on the number of modulation states available. For example, ld 16 = 4 bits are combined into one sequence for 16-QAM, and ld 64 = 6 bits are combined into one symbol for 64-QAM. At the iterator 3504, the symbol is repeated for retransmission. The proportion of symbols to be repeated and the repetition factor depend on the particular version of the method. This is controlled by the iteration determiner 3505. Invert bit determining unit 3506, which also includes a memory 3507 for storing a table containing bit inversion information, determines a particular bit of the repeated bit sequence to be inverted in optional bit inverter 3508, which is As described above, it depends on the modulation scheme. The bits are determined for inversion based on information received from the parallel entity, by executing each algorithm, or by reading the stored information from memory. The inverting bit determination unit 3506 executes a sub-step of the method of determining a subset of bits for inversion and a sub-step of the method of determining a subset of bits to be replaced with PAO bits as described above. It may further include (3509-3512). The bit inversion unit 3508 may further include a bit replacement unit for replacing the PAO bit with a selected fixed value. The transmitter 3500 may further include a control data transmitter 3513 that transmits information about inverted bits and bit sequence repetitions via the same transport channel or other transport channel.

Mapper 3514 maps symbols each representing one bit sequence to a modulation state using a mapping that does not change between transmission of a symbol having at least some of the inverted bits and retransmission of the same symbol, as described above. .

After mapping, pilot data is added and the frames are combined in pilot / data frame generation unit 3515 before the information is modulated onto a carrier in modulator 3516. The modulated signal is delivered to the receiving entity via channel 3517.

According to certain implementations, the transmitter 3500 may further include a unit such as an IF stage, mixer, power amplifier, or antenna. In terms of signal flow, such additional units may be seen included in channel 3517 as they may add noise to the signal or cause phase shifts or attenuation in the signal.

Units 3501 through 3516 may be implemented within dedicated hardware or digital signal processors. In this case, the processor executes the method described herein by executing instructions read from a computer readable storage medium such as read only storage, electrically erasable read only storage or flash memory. Such instructions may be further stored on another computer readable medium, such as a magnetic disk, optical disk, or magnetic tape, and downloaded to the device prior to use. It is also possible to implement mixed hardware and software.

Obviously the technique described above reduces the data transmission capability (capacity) of the transport channel. Therefore, the receiver must know how to process the received original and duplicate data. This knowledge can be obtained, for example, by sending a signal from a transmitter to a receiver. Preferably, some predetermined pattern is defined for the communication system, which defines the method and location for which part of the data the method described above is applied in what form. It is sufficient to signal a simple parameter indicating or representing one of these predefined patterns, from which the receiver can reproduce the particular method and form employed by the transmitter.

The method outlined above will mean, for example, that one or more transmitted bits are replaced or invalidated. In other words, the original value of such a bit is lost and passed to the receiver. The method described in the preceding paragraph will allow the receiver to have knowledge of which bits are affected by that method, so the results can be adapted to this situation. The receiver should set the information about those affected bits to a value meaning "unknown." For example, if the receiver (demodulator) uses LLR information as output, the LLR value indicating "unknown" is zero. If we use bit probabilities, each has a probability value of 0.5. If a hard decision, i.e. just 0 or 1 is used, the receiver may randomly generate a bit value since there is no information to base on determining the value of the replaced or invalidated bit. Preferably, the bits replaced or invalidated at the transmitter are part of the bit sequence after adding the FEC encoding, ie redundancy. In such cases, simply replacing or invalidating the bits eliminates some of the redundancy but does not automatically introduce loss of bit information or bit errors. The remaining transmitted redundancy may still compensate for such redundancy so that no bit or block error occurs after FEC encoding.

45 shows an exemplary structure of a receiver that can be used to receive data sent by the transmitter 1300 or 3500. The channel estimate values are provided to the LLR calculation unit 4507 to be taken into account when calculating the LLR value. Unit 4508 is an appropriate value (0 for LLR or 0,5 for linear probability) that is replaced by a fixed value or invalidated at transmitter 1300 or 3500 before all LLR values are repeatedly combined in unit 4509. Insert To determine which bit the LLR value should be inserted in, the control data receiver 4510 may receive respective information from the transmitter. It is also possible to directly identify the bits in which the received data has been replaced or, for example, a predefined manner stored in the table 4511, from which such information may be derived. Unit 4512 thus uses this information in control unit 4508. Optionally, unit 4507 may be controlled to omit the calculation of meaningless LLR values to reduce computational requirements.

The transmitter 1300 or 3500 and / or receiver 4500 may be part of the base station 3600 as shown in FIG. 36. Such a base station may further include a corresponding receiver 3604, a core network interface 3603, and data processing units 3601 and 3602, which are also configured as shown in FIG. 45.

The counterpart of base station 3600 would be a mobile station 3700 as shown in FIG. In addition to the receiver 3710 and the transmitter 1300 or 3500 (optionally configured as shown in FIG. 45), the mobile station includes an antenna 3701, an antenna switch 3702, a data processing unit 3703 and a controller 3704. It may further include.

The mobile station 3700 may be a module or mobile phone to be integrated into a portable computer, PDA, vehicle, bending machine, or the like. The mobile phone can further include a mixed signal unit 3705 and a user interface that includes a keyboard 3706, a display 3707, a speaker 3708, and a microphone 3709.

The method and transmitter according to the embodiment described above may completely eliminate multiplicity in the result of combining the retransmitted symbols. This can advantageously improve the reliability of channel estimation in a digital communication system. Better channel estimation has the advantage of reduced error rate and may provide connectivity with wireless communication systems in areas of poor coverage, fast fading conditions and other adverse environmental conditions.

The general and detailed description has shown how data symbols can be used, for example for channel estimation purposes. Assuming that the phase diversity is reduced to 1 by pinning a particular bit, and this bit is denoted as "pilot bit" in the figure, this process is shown again in a simple manner in Figures 46 (a) and 46 (b). The pilot bits are multiplexed with the data bits to generate the original sequence and ultimately used to generate the original symbol and at least one duplicate symbol. In the following, what kind of data is actually transmitted preferably on such a symbol. This is outlined as being most applicable to mobile radio system schemes but can be applied to fixed lines or other types of communication systems with modifications requiring the same consideration.

To simplify the description, the following terms are defined below.

Original symbol: A symbol generated from the original bit sequence as shown in FIG.

* Duplicate symbol: at least one symbol generated from the original symbol or from at least one duplicate sequence for the original bit sequence as shown in FIG.

Pseudo-pilot symbols: a combination of the original symbol and the corresponding duplicate symbol

* Pilot symbol: a single symbol that can be used as a reference symbol for channel estimation

Simple data symbol: A single symbol that carries data bits to one or more receivers

Simple control symbol: A single symbol that conveys the information required or helpful for successful system operation.

In general, simple data symbols can carry any kind of data. This may include data attached to a user or service application, such as voice data, video data, software data, etc., as well as data carrying control data or signals.

Simple control symbols on the physical layer are generally used for the purpose of transmitting signals. Much information needs to be sent between the network and the terminal for the purpose of carrying the signal. This information includes not only signaling messages generated on the physical layer, but also the required physical layer, which is necessary for system operation but not necessarily distinct for higher layer functionality. This kind of information is usually transmitted as simple control symbols.

The next channel is described in connection with the UMTS network. Different networks may use different names, but regardless of the name, there will be some data that performs the same or similar functions as described here. Therefore, it should be understood that what is described is not limited to the UMTS system or to a channel of a given name.

A sync channel is required for cell searching. These channels determine frame and slot synchronization, as well as information about the group to which the cell belongs.

A broadcast channel is used to transmit information about a predetermined cell or network specific information. The most typical data needed for any network is a kind of transmit diversity method used with a valid random access code in a cell and an access slot or other channel for that cell. Since the terminal cannot be written to the cell without the possibility of broadcast channel decoding, such a channel is required for transmission with relatively high reliability to reach all users within the planned coverage area.

For example, after the base station receives the random access message, the forward access channel carries control information to a terminal known to be located in a given cell. The access channel can also be used to transport packet data to the terminal.

That is, the paging channel carries data related to the paging procedure when the network wants to initiate communication with the terminal. A simple example is a speech call to a terminal, and the network sends a paging message to the terminal in the cells belonging to the location area where the terminal is expected to be embedded.

It is intended that a random access channel will be used to carry control information from the terminal to the network. Such channels are typically used for signaling purposes to register a terminal after power is applied to the network or to perform a location update after a move from one location to another or to initiate a call. For proper system operation, a random access channel must be heard from the entire required cell coverage area, which requires relatively high reliability of transmission data.

A collection indication channel is used to indicate reception of a random access channel signature sequence from a base station. Therefore it needs to be heard by all terminals in the cell, which requires a relatively high reliability of the transmitted data. Typically these channels are not noticeable at higher layers.

The paging indication channel works in conjunction with the paging channel to provide efficient sleep mode operation to the terminal. This channel must therefore be heard by all terminals in the cell, which requires a relatively high reliability of the transmitted data.

The shared control channel provides the physical layer control information necessary to enable the reception / demodulation / decoding of data on the shared data channel and to implement possible physical layer combinations of data delivered on the shared data channel in the case of retransmission or bad data packets. To carry.

The dedicated physical control channel may also carry essential control information including feedback signals, such as ARQ recognition (positive ACK and negative NAK), as well as link quality information (such as channel quality indicator CQI).

The shared control channel may include information that specifies one or more of the following.

Information about one or more of the frequency (sub-) carriers, time instants, and spreading codes used for data transmission;

Modulation schemes used for data transmission, eg BPSK, QPSK, 8-PSK, 16-QAM, 64-QAM, etc.

Multiple redundancy versions, ie redundancy versions of data blocks in the case of ARQ with so-called "incremental redundancy"

· Number of ARQ courses if multiple ARQ courses can exist in parallel

A first transmit / retransmit indicator that indicates whether the receiver combines the actual received data with previously received data or whether the buffer should be flushed and filled only with new data

Channel Coding (FEC) Types and Ratios

To reduce interference, it may be advantageous to reduce the correlation between different signals in the communication system. This process is sometimes called "orthogonalization" when reducing correlation to zero. Orthogonalization may be achieved, for example, by spreading orthogonal sequences or multiplexing into orthogonal sequences, where the orthogonal sequence is an OVSF sequence resulting from, for example, the Walsh-Hadamard matrix. The possibility to reduce correlation is to scramble or multiplex non-orthogonal sequences, such as pseudo-noise sequences, for example Gold sequences.

Orthogonalization or correlation reduction techniques can also be applied to the present invention. Symbol-based orthogonalization or correlation reduction techniques can be achieved by jointly applying pseudo-pilot symbols, or by applying these techniques individually to each of the original and duplicate pseudo-pilot symbols. This is illustrated in Figures 57-58 for multiplexing with spreading codes.

Alternatively in the case of bit-based orthogonalization or correlation reduction techniques, they apply equally to the original and duplicate sequences or individually to each of the original and duplicate bit sequences.

Of course, pseudo-pilot components can also be spread through bandwidth spreading. Again this spreading can be based on the components individually or jointly on the basis of pseudo-pilot symbols. 59-60 show examples of bandwidth spreading with spreading codes.

In addition, it would be beneficial for the system to modify the pseudo-pilot symbol prior to transmission, for example by multiplexing to a constant phase condition. For carrier tracking reasons, it would be desirable that neither the real part nor the virtual part of the pseudo-pilot symbol is zero. However, if the pseudo-pilot is designed such that the pseudo-pilot symbol lies on one of the orthogonal axes, the pseudo-pilot symbol may shift in phase. Clearly, the phase shift of the pseudo-pilot symbol is equivalent to the phase shift of the original and corresponding duplicate symbols. Although FIG. 61 shows the rules for a constant phase shift applied to all pseudo-pilot symbols, those skilled in the art will appreciate that the transition may vary from symbol to symbol.

Figure 47 shows a simple case where the ratio between shared control symbols to pilot symbols is one, i.e., the number of such symbols per frame is the same. It is therefore easy to combine each one pilot with each one control symbol for one pseudo-pilot symbol. However, it is also possible in the system that the ratio is not equal to one. One solution is to construct as many pseudo-pilot symbols as there are both pilot symbols and control symbols. For example, if there are n pilot symbols and m control symbols, min (n, m) pseudo-pilot symbols may be generated, and additionally, nm pilot symbols or mn control symbols may be simple according to conventional methods. Transmitted as a symbol.

If a transmission using a pseudo-pilot requires a modulation scheme that carries at least two bits per symbol, data belonging to the same transport channel (e.g., shared control channel) may be completely mapped onto the pseudo-pilot symbol. Unexpected situations may arise. In general, excess data may be transmitted using a modulation scheme independent of the pseudo-pilot modulation scheme as shown in FIG. 65. However, from a uniform design point of view, it would be desirable to transmit a transport channel using a single modulation scheme. In such a case, it would be desirable to repeat some of the non-pseudo-pilot symbols or reduce the number of symbols to fill the optionally available bandwidth as shown in FIG. 66.

As a reason for timing, it would be desirable to transmit control or signaling data within the first time slot of the frame. Particularly for other channels or shared data channels that transmit user data at least partially in a time-multiplexed manner, in terms of timing, the time required to process the control signals to the receiver and take the necessary actions to properly receive data information. To allow it, it would be desirable to transmit the control channel to which the control data belongs prior to the corresponding data channel. This is particularly applicable to shared control and data channels. An example of a conventional solution for an OFDM system is shown in FIG. 47. An OFDM frame consists of several time slots, in this case seven "OFDM symbols" and several carrier frequencies, here eight "subcarriers." The pilot and shared control symbols are frequency-multiplexed within the first OFDM symbol, and the two symbols are time-multiplexed together with the shared control symbol.

A corresponding solution according to the invention is shown in FIG. 48. This figure shows time-multiplexing of pseudo-pilot symbols and shared data symbols. The pseudo-pilot symbol includes multiplexing of pilot bits and shared control bits according to FIG. 46. In this case, since the pseudo-pilot symbols, i.e., the original and duplicate symbols carry shared control information, one pseudo-pilot symbol can be used for channel estimation and each of the configuration (original and duplicate) symbols carries shared control information. . As pilot and control information is ultimately multiplexed onto modulation symbols, it can be interpreted as "modulation division multiplexing (MDM)" or "modulation multiplexing" of pilot and control information on the same symbol. Since the multiplexing of the original and replica symbols is in accordance with FIG. 48 in the frequency-domain, it is called " frequency replication multiplexing (FCM). In summary, the first OFDM symbol has an FCM-MDM structure.

However, multiplexing of the original and duplicate symbols may be realized as shown in FIG. 49 in the time domain. Here we have a TCM-MDM structure, "Time Replication Multiplexing-Modulation Division Multiplexing", where the shared data portion is time multiplexed before the pseudo-pilot.

50 and 51 show a similar approach where the pseudo-pilot and shared data symbols are frequency-multiplexed.

Of course, neither multiplexing between the pseudo-pilot and shared data nor original / replicating multiplexing should be the same in one OFDM frame. Examples are shown in Figs. 52-56, in which various stages of change are realized with respect to pseudo-pilot / shared data multiplexing and original / replicating multiplexing.

It should be apparent to those skilled in the art that the order of the original and duplicate symbols in FIGS. 48-56 is not critical. For example, in FIG. 49, the first OFDM symbol may always transmit a duplicate symbol, while the second OFDM symbol may always transmit the original symbol. Mixed forms are of course also possible.

Complex combinations of original and duplicate symbols, e.g. addition of complex values, are a separate matter and channel estimation by reducing the level after combination / component state / precombination compared to the number of parameters / component states / number of parameters It is also possible to combine the components or other parameters of these symbols to improve the reliability of the. The parameter or component of such a symbol may be, for example, an actual part, an imaginary part, power, magnitude, phase, or amount or duration derived from one or more of these.

In another embodiment of the present invention, a target that improves channel estimation capability may be determined by reducing the number of possible magnitude levels, thereby reducing the number of different combination values obtainable for all data word values, thereby reducing each data word value. Is achieved by adding magnitude values associated with the data word values to a value less than the number of magnitude levels in the first mapping in accordance with the first and at least one additional mapping.

 In yet another embodiment of the present invention, a target that improves channel estimation capability is characterized by reducing the number of possible power levels, thereby reducing the number of different combination values obtainable for all data word values, thereby reducing each data word value. Is achieved by adding power values associated with the data word values to less than the number of different power levels in the first mapping in accordance with the first and at least one additional mapping.

In another embodiment of the present invention, a target that improves channel estimation capability reduces the number of possible phase levels, thereby reducing the number of different combination values obtainable for all data word values, thereby reducing each data word value. Is achieved by adding phase values associated with the data word values to less than the number of different phase levels in the first mapping in accordance with the first and at least one additional mapping.

For each of the aforementioned level reductions or any combination of the aforementioned level reductions, a duplicate symbol or sequence can be easily generated by applying the outlined principle of making the necessary changes in the case of a coherent combination. General principles are shown in the flowcharts of FIGS. 67-69. Of course, if a combination of level reductions is needed, the step of determining the replication arrangement can only be achieved by considering the combined requirements. If both power and phase levels should be reduced, determining the replication arrangement should be modified so that "the corresponding power and phase in each corresponding replication backstroke is achieved for each symbol." It should also be clear that the order of these steps may change. For example, if both power and phase levels should be reduced to 1 and 2, respectively, FIGS. 62-64 show exemplary solutions for QPSK, 8-PSK and 16-QAM, respectively. It should be noted that the terms "original array", "duplicate array" here and in the following section are used to describe behavior on a symbol level and are not limited to only one of the approaches for pseudo-pilot generation in accordance with FIG. 46.

Assuming that the average power should be 1, applying the flowcharts of FIGS. 67 and 69 to the original QPSK of FIG. 62, the following power and phase levels are determined.

Obviously it is not unusual to achieve a single power level after the combination. It is then defined that the bit sequence should have a target phase level after the combination of zeros. This results in the following for the duplicate array in the last step:

This is shown as a replication arrangement in FIG. 62. In this example, it will be appreciated that a result with the same effect can be achieved by inverting the second bit from the original bit sequence to obtain a duplicate sequence and using the original arrangement to obtain a duplicate symbol from the duplicate sequence. Those skilled in the art will appreciate that the bit operation approach is generally a possible alternative to the modified arrangement.

Assuming that the average power should be 1, applying the flowcharts of FIGS. 67 and 69 to the original 8-PSK of FIG. 63, the following power and phase levels are determined.

Again, it is not unusual to achieve a single power level after the combination. It is then defined that the symbol must have a target phase level after the combination of zeros. This results in the following for the duplicate array in the last step:

This is shown in the replication arrangement in FIG. If the number of symbols is transformed into a bit sequence, those skilled in the art will be able to easily apply the bit operation to achieve the same result.

Assuming that the average power should be 1, applying the flowcharts of FIGS. 67 and 69 to the original 16-QAM of FIG. 64, the following power and phase levels are determined.

The only target power level after the combination is set to 2.0. It is then defined that the symbol must have a target phase level after the combination of zeros. This results in the following for the duplicate array in the last step:

This is shown in FIG. 64 as a replication arrangement. Once the number of symbols has been transformed into a bit sequence, those skilled in the art can easily adapt the bit operation to achieve the same result.

According to the investigation of FIGS. 62 to 64, if combinations are made separately for power / magnitude and phase, then after combination of original and replica these are sufficient to reduce the number of power or magnitude and the number of phases to one. You will find out.

및

로 각각 기술되며, 마찬가지로 크기 레벨은

및

, 위상 레벨은

및

로 기술된다고 가정한다. 아래 수식에서와 같이 채널 계수 h가 크기 이득 k와 위상 천이 δ로 분해될 수 있다고 가정한다.In such a case, the actual estimation of the channel coefficient h may preferably employ the following strategy. The power level of the symbols from the original and duplicate arrays

And

Are described as

And

, Phase level is

And

Assume that It is assumed that the channel coefficient h can be decomposed into magnitude gain k and phase shift δ as in the following equation.

The following characteristics are obtained for received power, magnitude and phase level (ignoring other channel effects).

By adding the received values, we get

Therefore, the channel size gain k and the phase shift δ can be estimated as follows.

It will be appreciated that these equations are given for a simple case where one original symbol and one duplicate symbol are sufficient. If multiple replication arrays are used, the denominator in the channel size gain expression should describe the sum of all these replication arrays instead of a single array; You must add 1 to the number.

To check the power, magnitude, and phase levels in more detail than Table 1, Table 2 lists the actual levels assuming each array is normalized to the average power per symbol.

This will then be illustrated for the 16-QAM shown in FIG. 64. For any 16 symbols through Table 2 and FIG. 64 the sum in this case is always

(또는 각도 해석에 의존하여 등가적으로

)이 되는 것을 안다.

(Or equivalently depending on the angle interpretation

I know).

Using the values for this 16-QAM example, the following equation is obtained.

For the example of QPSK and 8-PSK of FIGS. 62 and 63, the sum of the power level and the magnitude level

It becomes known.

Therefore, any combination of magnitude or power level may be used to arrive at an estimate for channel size gain k . This is possible for any pure PSK scheme. It may be concluded from Table 2 that the size level is preferred for the pure ASK scheme since the array can easily be formed to be sufficient to reduce the combination to a single size level. Since the mix of ASK and PSK should focus on the choices (or limitations) of the respective configurations, the size level is the case even if ASK prefers size level combinations over power level combinations because there is only one single copy required. Combinations are preferred.

This is further illustrated for 4-ASK / 4-PSK in FIG. For generalization, each array point is labeled with a bit sequence (number) as well as a symbol label (alphabet). Applying the flowcharts of Figs. 68 and 69 to the original 4-ASK / 4-PSK and assuming that the average power is 1, the following magnitude and phase level are determined.

The only target size level after the combination is set to 8 / sqrt (21). The target phase level is then defined after the combination of zeros. This results in the following for the duplicate array in the last step:

This is shown as a replica arrangement in FIG. In this example, it will be appreciated that a result with the same effect can be achieved by inverting the first and third bits from the original bit sequence to obtain a duplicate sequence and using the original arrangement to obtain a duplicate symbol from the duplicate sequence.

The effect to reach the copy array is by modifying the original bit sequence with the copy sequence or modifying the bit sequence's mapping rule to the modulation state before mapping the copy sequence to the modulation state according to the mapping rules that are also used for the original bit sequence mapping. Is achieved.

Assuming that the copy array does not have to have the same layout as the original array within the complex plane, in the case of power or magnitude or phase combinations, a single copy array is always formed to be sufficient to achieve the purpose of power / size / phase level reduction. Such different layouts can be seen, for example, by comparing the right and left arrangements in FIG. 6.

It is noted that the described possibility of using pseudo-pilot for data transmission, in particular data type, for example control data, signaling data, broadcast data, etc., is applicable regardless of how to achieve a reduction in the level of versatility for pseudo-pilot. You will know Therefore, for example, when generating a pseudo-pilot using a power- and phase-combination method, it is also preferable to transmit a shared control channel using a pseudo-pilot as in FIGS. 48 to 56. Those skilled in the art will realize that there is no fundamental difference in what kind of data can be transmitted in a pseudo-pilot generated using a complex-combination method compared to using one or more power- / size- / phase-combination methods. will be.

Although the invention has been described in connection with embodiments constructed in accordance with the description herein, various modifications, variations and modifications of the invention within the scope of the appended claims and in light of the foregoing have been made without departing from the spirit and intended scope of the invention. It will be apparent to one skilled in the art that improvements may be made. In addition, the areas believed to be familiar to those skilled in the art are not described herein in order not to unnecessarily obscure the invention described herein. It is, therefore, to be understood that the invention is not only limited by the specific embodiments, but also by the scope of the appended claims.

Claims (62)

In the method for transmitting data in a digital communication system, a) selecting 1208 a subset of all available modulation states in the predetermined modulation scheme to be used for transmission, and b) transmitting a first symbol representing a first plurality of bits, the symbol having a first modulation state 1205 having a first modulation state included in the subset; c) transmitting additional symbols representing the first plurality of bits, each of the additional symbols having at least one additional transmission step 1206 having additional modulation states included in the subset Including but not limited to: For each bit value combination, the addition of the complex value associated with the first and the additional modulation state yields the same phase of the complex result for all combinations of bit values in the plurality of bits. How to send data. The method of claim 1, For each bit value combination, the addition of the complex value associated with the first and the additional modulation state yields the same result for all combinations of bit values in the plurality of bits. The method according to claim 1 or 2, Wherein the first modulation state is obtained by first mapping a bit value combination to a modulation state, The at least one further transmission step comprises exactly one further transmission step, Said one additional modulation state is obtained by second mapping a bit value combination to a modulation state, The second mapping, which maps the bit value combination to the modulation state, Obtained by first mapping a data word value to a modulation state, i. Dividing the complex plane representing the first mapping of the bit value combination into the modulated state into at least two non-overlapping adjacent sub-planes (404, 405), Ii. Determining an axis of symmetry 412 for at least a portion of the subplane for modulation states contained within each portion of the subplane; Iii. Allocating a complex value 409 to at least a portion of the bit value combination in the second mapping 402, wherein the complex value 409 is the complex value 408 according to the first assignment 401. With respect to the point 407 on the axis of symmetry of this sub-plane 405, in the complex plane mirrored essentially from the position 408 of the complex value assigned to the bit value combination according to the first assignment 401 And having a location, The subset of modulation states to be used for transmission includes all modulation states located in one of the at least two sub-planes How to send data. The method of claim 3, wherein The point 407 on the axis of symmetry 412 that serves as the center for mirroring in step VII is all complex numbers assigned to at least a portion of the modulation state in the first assignment 401 and located within the subplane 405. A method of sending data that is the average of the values. The method according to claim 1 or 2, Wherein the first modulation state is obtained by first mapping a bit value combination to a modulation state, The at least one further transmission step comprises m-1 additional transmission steps, The m-1 additional modulation states are obtained by m-1 additional mapping of a bit value combination to a modulation state, The m-1 mapping that maps a combination of bit values into a modulation state, Obtained from the first mapping of the bit value combination to the modulation state, i. Dividing the complex plane representing the first mapping of a bit value combination into a modulation state into at least two non-overlapping adjacent subplanes 706, 707, wherein the number of modulation states in at least a portion of the subplane is m; , Ii. Assigning at least a portion of said data word value different modulation states in the same subplane, one modulation state for each mapping, The subset of modulation states to be used for transmission includes all modulation states located in one of the at least two sub-planes How to send data. The method according to claim 1 or 2, In a method applied to a digital communication system using phase shift key modulation, Wherein the first modulation state is obtained by first mapping a bit value combination to a modulation state, The at least one further transmission step comprises exactly one further transmission step, Said one additional modulation state is obtained by second mapping a bit value combination to a modulation state, The second mapping, which maps the bit value combination to the modulation state, Obtained by first mapping a data word value to a modulation state, i. Split a complex plane representing a first mapping of a bit value combination to a modulation state into at least two non-overlapping adjacent subplanes 804, 805, wherein all modulation states in which at least a portion of the subplane is contained within the subplane. Having an axis of symmetry 806 in relation to the position of Ii. Determining an axis of symmetry 806 for at least a portion of the subplane associated with a modulation state contained within each portion of the subplane; Iii. Assigning a complex value to at least a portion of the data word value in the second mapping (802), wherein the complex value is in the sub-plane where the modulation state 807 according to the first mapping (801) is located. With respect to the axis of symmetry 806, having a position 808 in the complex plane approximately mirrored from the position 807 of the modulation state assigned to the bit value combination according to the first mapping 801, The subset of modulation states to be used for transmission includes all modulation states located in one of the at least two sub-planes How to send data. The method according to any one of claims 3 to 6, And the complex plane divides the complex plane into the subplanes in relation to an axis of symmetry (203, 303, 403, 503, 803) taking into account the positions of complex values of all modulation states included in the first mapping. The method according to any one of claims 3 to 7, And dividing the complex plane into the subplane such that no complex value of any modulation state is located on the boundary line between the subplanes (203, 303, 403, 503, 803). The method according to any one of claims 3 to 8, And the non-overlapping adjacent sub planes are half planes of the complex plane. The method according to any one of claims 1 to 9, And said transmitting step is then achieved on the same transmission channel. The method according to any one of claims 1 to 9, The digital communication system includes at least one of time division, frequency division, code division or OFDM component, and wherein the transmitting step is performed in an adjacent instance with respect to at least one of the components. The method according to any one of claims 1 to 11, And said transmitting step is applied to each transmitted symbol. The method according to any one of claims 1 to 11, And said second and further transmission step are applied to a predetermined number of data symbols per transmission frame. The method according to any one of claims 3 to 13, And the number of subplanes (404, 405) is two. The method according to any one of claims 1 to 14, The first mapping and the at least one additional mapping are obtained from a common super-mapping by pre-assigning a primary control word to each data word, the super-mapping of which is a chained value of the control word and the data word. Mapping to a modulation state, each transmission being associated with a specific value of a control word. The method according to claim 1 or 2, Step a) includes replacing at least one of the first plurality of bits with a fixed value to obtain a second plurality of bits, Step b) includes mapping the second plurality of bits to the first symbol having the first modulation state in accordance with a predefined mapping of a bit sequence to a modulation state; Step c) inverts the bits of the at least one subset of the second plurality of bits to obtain at least one additional plurality of bits and retains bits that are not included in the unchanged subset; Mapping the at least one additional plurality of bits to the at least one additional symbol having the at least one additional modulation state in accordance with the predefined mapping that maps a sequence to a modulation state. How to send data. The method of claim 16, In step a) replace one of the first plurality of bits with a fixed value to obtain a second plurality of bits, and consequently, in step b) all modulation states that can be generated from the second plurality of bits are associated with a complex number And a data transmission method placed in a half plane of one of said complex planes by said value. The method according to claim 16 or 17, Step c) inverts all bits included in the second plurality of bits, and wherein the subset is a subset of the combination set from one of the second plurality of bits to all but one of the bits. Including, Step a) includes replacing one of the first plurality of bits corresponding to one bit from the second plurality of bits not included in the combination set with a fixed value. How to send data. The method of claim 18, And the predefined mapping is a gray mapping mapping that defines a modulation state of phase shift key modulation. The method of claim 18, The predefined mapping is a gray mapping defining a modulation state of a mixed modulation comprising magnitude shift key modulation and phase shift key modulation, and wherein the first plurality of bits are associated with the plurality of bits in the gray mapping. A set of magnitude shift keys defining an absolute value of a complex value of a phase shift key set defining a phase value of a complex value of a modulation state associated with the plurality of bits in the gray mapping, Step c), i. At least one inverting sub-step, executed on the magnitude shift key set, or Ii. At least one inverted sub-set executed on said phase shift key set as defined in claim 15, The sub-set i. Inverts all bits contained in one subset of the magnitude shift key set, and the same value for half of all the plurality of bits mapped to the modulated state with the lowest transmit power among all existing modulation states. Consisting of bits 2210 having, or, Inverts all the bits contained in one subset of the magnitude shift key set, and sets the same value for half of all the plurality of bits mapped to the modulation state with the highest transmit power among all existing modulation states. Comprising a bit 2210 having How to send data. The method of claim 18, The predefined mapping is a gray mapping that defines a modulation comprising a first component and a second component, the second component being essentially orthogonal to the first component, and wherein the first plurality of bits The second configuration according to the first bit set associated with the first component and the third gray mapping that maps the bit sequence to the second modulation state set according to a second gray mapping that maps a bit sequence to a first set of modulation states. A second set of bits associated with the element, Step c), i. Inverting the bits included in the first set of bits (2702), and all of the bits mapped to the modulation state having the lowest transmit power among all existing modulation states in the first set of modulation states according to the second gray mapping; A sub-step with the same value for half of the plurality of bits, or Inverting the bits included in the first set of bits (2702) and all of the bits mapped to the modulation state with the highest transmit power among all existing modulation states in the first set of modulation states according to the second gray mapping A sub-step having the same value for half of the plurality of bits, Ii. Inverting a bit included in the second set of bits (1203), the same sign of the second component of the complex value of the second set of modulation states associated with the plurality of bits in the third gray mapping A sub-step having the same value for all the plurality of bits mapped to the second set of modulation states with How to send data. The method of claim 21, And the modulation is a square quadrature amplitude modulation. The method according to any one of claims 16 to 22, Transmitting information associated with an identity of the at least one bit of the first plurality of bits, replaced with a fixed value. A computer readable storage medium storing program instructions which, when executed in a processor of a transmitter of a digital communication system, cause the transmitter to perform the data transmission method according to any one of claims 1 to 23. 24. A transmitter (1300, 3500) for a digital communication system, implemented to carry out a data transmission method according to any of the preceding claims. A base station 3600 for a mobile communication system comprising a transmitter according to claim 25. A mobile station 3700 for a mobile communication system comprising a transmitter according to claim 25. In a method for receiving data in a digital communication system, a) first and second receiving steps for receiving first and second symbols representing a first plurality of bits, b) calculating a probability of computing a possible value from the received first and second symbols for at least the first plurality of bits; c) setting a possible value for at least one predetermined bit of said first plurality of bits to a value representing an unknown bit value Data receiving method comprising a. The method of claim 28, The possible value comprises a log of probability ratios, and the possible value representing an unknown bit value is zero. The method of claim 28, The possible value comprises a linear probability and the possible value representing an unknown bit value is 0.5. The method according to any one of claims 28 to 30, Receiving information regarding the identity of the at least one predetermined bit. A computer readable storage medium storing program instructions, when executed in a processor of a receiver of a digital communication system, that cause the receiver to perform the method of receiving data according to any of claims 28 to 31. 32. A receiver for a digital data communication system, adapted to perform the method of any of claims 28 to 31. A base station 3600 for a mobile communication system comprising a receiver according to claim 33. A mobile station 3700 for a mobile communication system comprising a receiver according to claim 33. In the method for transmitting data in a digital communication system, a) transmitting a first symbol representing a first plurality of bits, wherein the symbol has a first modulation state; b) transmitting additional symbols representing the first plurality of bits, each of the additional symbols having at least one additional transmission state having an additional modulation state, By combining at least one parameter of the at least one parameter of the additional symbols with at least one parameter of the first symbol, the number of different possible parameter states after the combination is smaller than the number of different parameter states before the combination How to send data. The method of claim 36, At least one of the parameters is power, and the combination is realized by combining the power of the at least one additional symbol with the first symbol. 38. The method of claim 36 or 37, At least one of said parameters is a magnitude, and said combination is realized by combining a phase of said at least one additional symbol with said first symbol. 39. The method of any of claims 36-38, At least one of the aircraft parameters is a phase, and the combination is realized by combining the phase of the at least one additional symbol and the first symbol. The method according to any one of claims 36 to 39, And the combination is addition. The method according to any one of claims 36 to 40, The number of different parameter states after the combination is one. In the method for transmitting data in a digital communication system, a) generating an original symbol by mapping bits of the original bit sequence using a modulation array; b) generating at least one duplicate symbol from at least one duplicate bit sequence generated from said original bit sequence or from said original symbol, Combining the original symbol with the at least one replica symbol to form a quasi-pilot symbol How to send data. The method of claim 42, Generating the original bit sequence by multiplexing at least one pilot bit and at least one control data bit. The method of claim 42 or 43, And constituent symbols of the pseudo-pilot symbols span at least two frequency carriers. The method of claim 42 or 43, And constituent symbols of the pseudo-pilot symbols span at least two time slots. The method according to any one of claims 42 to 45, And the pseudo-pilot symbol is transmitted in an OFDM system. The method of any one of claims 42-46, A data transmission method for generating the original symbol and the at least one duplicated symbol using a data transmission method as described in any one of claims 1 to 23 and 36 to 41. The method of any one of claims 42-47, Multiplying the pseudo-pilot symbol with at least one of the original symbols and at least one duplicate symbol having a defined sequence. In a method for receiving data in a digital communication system, a) receiving a first and at least one additional symbol, b) obtaining at least one combination of at least one parameter of the first symbol and at least one parameter of the at least one additional symbol; c) using said at least one combination to obtain an estimate of a communication channel parameter Data receiving method comprising a. The method of claim 49, Wherein said at least one combination comprises a power combination of said first symbol and at least one additional symbol. 51. The method of claim 49 or 50, Wherein said at least one combination comprises a size combination of said first symbol and at least one additional symbol. The method of any one of claims 49-51, Wherein said at least one combination comprises a phase combination of said first symbol and at least one additional symbol. The method of any one of claims 49-52, And said at least one combination comprises a complex value combination of said first symbol and at least one additional symbol. The method of any one of claims 49-53, Wherein said communication channel parameter comprises at least one of magnitude gain and phase shift and a complex channel coefficient. A transmitter for a digital communication system configured to perform the method of any of claims 36 to 48. A base station for a mobile communication system comprising a transmitter according to claim 55. A mobile station for a mobile communication system comprising a transmitter according to claim 55. A computer readable storage medium storing program instructions, when executed in a processor of a transmitter of a digital communication system, that cause the transmitter to perform the data transmission method according to any one of claims 36 to 48. A computer readable storage medium storing program instructions, when executed in a processor of a receiver of a digital communication system, that cause the receiver to perform the method of receiving data according to any of claims 49-54. A receiver for a digital data communication system configured to perform the method of any of claims 49-54. A base station for a mobile communication system comprising a receiver according to claim 60. A mobile station for mobile communication system comprising a receiver according to claim 60.

Images