|Title of Invention||
CODED MODULATION METHOD, WHICH TAKES TAILBITS AND THEIR CODING INTO ACCOUNT
|Abstract||A method for coded modulation of digital data, the digital data having useful bits carrying out the coded modulation at a number of levels dividing up the useful bits from the coded modulation into parallel signal streams subject each signal stream to channel coding by a respective coder channel coding the useful bits at, at least one fixed code rate supplementing the useful bits by tail bits channel coding the tail bits in the respective coder at a variable code rate in order to achieve a predefined number of bits for the respective coder so that all the parallel coded signal streams contain the same number of bits assigning the channel coded data to signal space points in order to produce modulation symbols.|
Method for coded modulation
The invention is based on a method for coded modulation of the generic type of the independent patent claim.
It is already known to employ coded modulation, in which channel coding and modulation are optimized together. The designation multilevel coding is also known as an equivalent to the term coded modulation.
Advantages of the invention
By contrast, the method according to the invention for coded modulation, having the features of the independent patent claim, has the advantage that a variable code rate is used for the tail bits, as they are known, and in each case set such that the existing transmission capacity can be utilized fully. The trans¬mission frame used prescribes the maximum number of modulation symbols. This leads to an optimal increase in capacity as compared with known methods, The tail bits are added on, either in the coders or in the upstream bit multiplexer. Furthermore, the method according to the invention can be used flexibly, since different transmission modes with different numbers of modulation symbols can be used. Furthermore, no additional signalling from the transmitter to the receiver is needed. Finally, the implementation of the method according to the invention is also simple, since no additional computing power is needed.
By means of the measures and developments listed in the dependent claims, advantageous improvements of the method for coded modulation specified in the independent patent claim are possible.
It is particularly advantageous that the variable code rate is achieved by variable puncturing. Puncturing means that, in order to achieve a higher code rate, some bits are not co-transmitted.
It is advantageous that the variable puncturing schemes is [sic] either stored in a table in the transmitter and in the receivers or is calculated via a known computing rule, the computing rule being known to the transmitter and the receivers.
Furthermore, it is advantageous that the fixed code rate for the useful bits and a predefined transmission rate are signalled to the receiver, so that the receiver can use them to extract the variable code rates for the tail bits, in order to synchronize itself with the transmitted data.
Furthermore, it is advantageous that convolution coding is used, which is a widespread technique for channel coding. The number of coded bits in this case is calculated from the number of modulation symbols multiplied by the coding level of the modulation m and divided by the number n of stages. In this way, the existing transmission capacity is utilized optimally.
Furthermore, it is advantageous that both a transmitter and as [sic] also a receiver have means for carrying out the method according to the invention.
Exemplary embodiments of the invention are shown in the drawing and will be explained in more detail in the following description. Figure 1 shows partitioning of a 4-ASK, Figure 2 shows a block diagram of the trans¬mitter according to the invention, Figure 3 shows a block diagram of the receiver according to the invention, and Figure 4 shows a flow diagram of the method according to the invention.
The digital transmission system Digital Radio Mondiale (DRM) for the transmission bands below 30 MHz is currently being developed. It has been decided that the channel coding used will be multilevel coding (MLC). In this case, the channel coding and the modulation are optimized together for which reason one also speaks of coded modulation, Channel coding adds redundancy to the data, by using which transmission errors can be detected and corrected if necessary,
In the case of a relatively high-level modulation method using a q-nary signal constellation, the signal alphabet has exactly q values. The basis for the MLC is formed by partitioning the signal alphabet into subsets, Each division step is allocated one component of the address vector of the signal space representation. In this case, each component is protected by a dedicated code. If a 2m-level signal constellation is considered, the result is a ' sub¬division into n levels, if m = n, corresponding to the address vector c (=co, C1, . . - , Cn-1) - The coding level m of the modulation is, for example, not necessarily equal to the number of levels if a 64-QAM (Quadrature amplitude modulation) is used.
Figure 1 shows the partitioning of a 4-ASK (Amplitude Shift Keying) . In the case of 4-ASK, four states are therefore coded. The coding of the data stream is carried out using n parallel coders, the code Co having the lowest code rate RQ, that is to say adding the most redundancy, and protecting that position of the address vector which is most susceptible to faults. In Figure 1, four states are marked by filled circles on the uppermost state strand. One then passes to the individual codable states in a 4-ASK via the two central state strands. The first level becomes either Co = 0 or 1. In a corresponding way, the four filled circles are distributed to two strands of numbers which have mutually complementary filled and empty circles. In the case of the lower four state strands, the individual states in a 4-ASK, namely 00, 01, 10 and 11, are then coded. In this case, the state 00 on the extreme left has a filled circle, - which is then followed by three empty circles. The state 01 has the filled circle in the third place, starting from the left. The state 10 has the filled circle in the second place, starting from the left, and the state 11 has the filled circle on the extreme right. The remaining positions are symbolized by empty circles for a 0.
Figure 2 shows a block diagram of the transmitter according to the invention. Contained in the data memory 1 are data which are to be transmitted by the transmitter according to the invention. However, other data sources can also be used here. These data are transferred from the data memory 1 to a source coder 2, which performs source coding in order to reduce the amount of data to be transmitted* The data source-coded in this way, with the useful bits, are then transmitted to a bit multiplexer 3, which distributes the data stream to n parallel lines. Connected to each of these
n lines, which are numbered consecutively from 0 to n-1, is a respective coder, which channel-codes one of the data streams (QO - - . qn-1) For example, here a coder 5 is shown in the line 0 and a coder 4 in the line n-1. At the output of the respective coder, the signals co and Cn-1 are present. The coders 4 and 5 carry out the channel coding by means of convolution coding, and therefore add redundancy to the useful bits again, tail bits, as they are known, also being connected to the useful bits, in order to transfer the coders 4 and 5 as convolution coders into a defined final state in each case. Such a coder 4 and 5 has shift registers, which are wired in accordance with the coding. The tail bits, here they are logic zeros, ensure that the coders 4 and 5 are in a defined state at the end of the coding, and also, in the receiver, the decoder is in a defined state at the end of the decoding, this state being detected by the fact that all the bits in the coder 4 and- 5 and the decoder are logic zeros.
In addition, these tail bits are acted on at a code rate but, according to the invention, this is variable here. This variable code rate is set in such a way that the transmission capacity which is available and is defined by the transmission frame is fully utilized,
Here, variable means that the code rate of the tail bits can be different for each level. The variable number of tail bits means that the number of coded bits, that is to say the coded useful bits plus the coded tail bits, is the same for each level in the coding scheme. Furthermore, the number of coded bits then corresponds to the number of modulation symbols multiplied by the coding level of the modulation m and divided by the number of levels n. In the case of a 4-ASK or 8-ASK, m = n, so that the number of coded bits corresponds to the number of modulation symbols. In the
case of a 64-QAM, however, m = 6 and n = 3 are possible, so that the number of coded bits corresponds to twice the number of modulation symbols.
The data channel-coded in this way are then assigned to signal space points in block 6, in order then to produce the respective modulation symbols,
The component codes used in the individual coders 4 and 5 are convolution codes with puncturing. Therefore, the code rates can be co-ordinated with one another, in order to achieve the best possible transmission. performance. The punctuated codes have a period which corresponds to the denominator of the code rate. For example, the code 4/5 has exactly five output bits in the case of four input bits. The period of the output bits is therefore 5, since no smaller number of output bits is possible, in order to maintain the code rate. In the case of MLC, a different code rate is used for each level. In order to ensure that the number of bits at the outputs of all the coders 4 and 5 is the same, this must be appropriately variable. However, this applies only to the tail bits, since the code rate for the useful bits remains the same. By means of changing the code rate of the tail bits, it is possible to influence the number of codes bits, In this case, the starting point is a maximum code rate for the tail bits, which is reduced, that is to say additional redundancy is added, in order to achieve adaptation. In this case, the reduction in the code rate of the tail bits must be carried out for each level to such an extent that, at each level, output bits are produced and these correspond to the number of modulation symbols multiplied by m and divided by n. The number of tail bits can be defined, via the puncturing scheme, to any value in a certain range. Alternatively, a minimum
code .rate may be assumed, and this may be increased by adapting the puncturing.
The following example is intended to clarify the
benefit. When convolution codes with a memory length of
6 are used, 6 tail bits are needed in order to transfer
the coder 4 or 5 into a defined final state. This
defined final state is intended to be reached for each
transmission frame, in order to prevent the propagation
of an error during the decoding in the receiver. The
tail bits can be added on either in the coders 4 and 5
or in the bit multiplexer 3. The basis used is an 8-ASK
as modulation with m = 3 and an MLC with n = 3 levels.
If the tail bits are provided at a base code rate of
2/3, exactly nine coded bits are produced at the coder
input, corresponding to the six zeros (tail bits) , If
there are 2 00 modulation symbols in a transmission
frame, the result is ideally 200 coded bits per level.
If the minimum number of nine tail bits is subtracted
from this, then the result is 191 coded useful bits as
the maximum possible number per level. If, for each
level, the period is considered, the result is that,
for stage 0 at a code rate of 1/3, exactly 189 coded
bits are produced (corresponding to 63 useful bits at
the code rate) , and eleven tail bits are consequently
needed. For level 1 at a code rate of 2/3, there are
exactly 189 coded bits (corresponding to 12 6 useful
bits at the present code rate), so that likewise eleven
tail bits are needed. At level 2, there are 190 coded
bits with a code rate of 4/5 (corresponding to 152
useful bits), and therefore ten tail bits are
necessary. The code rates of the tail bits of ' the
levels are changed from the base code rate of 2/3 to
6/11 for level 0 and level 1 and, respectively, 6/10
for level 2. This achieves the situation where all the
modulation symbols are filled with coded bits. The
result of this calculation is that 341 useful bits are
transmitted which, as compared with conventional methods, corresponds to an increase in capacity. The result in this case is 568 coded bits, which correspond to the 341 useful bits. In the case of conventional methods with the code rates 4/5 and 2/3, it would have been necessary to select a number of coded bits and therefore of useful bits which can be divided by 3 and 5, Using the variable code rate, it is now possible to achieve an optimum value for the coded bits.
In Figure 2, the modulation symbols coded in this way are then transferred from the function block 6 to an OFDM modulator 7, which distributes the individual modulation symbols to frequency carriers which lie close beside one another and are mutually orthogonal, The OFDM signals produced in this way are then mixed in an analogue high-frequency part 8, are amplified and then emitted by an antenna 9.
Figure 3 shows a block diagram of the receiver according to the invention. An antenna 10 for receiving the OFDM signals is connected to an input of a high-¬frequency receiving part 11. The high-frequency receiving part 11 converts the received signals to an intermediate frequency, amplifies and filters them. Furthermore, the high-frequency receiving part 11 transfers these signals to a digital part 12, which digitizes the received signals and carries out OFDM demodulation- The modulation symbols obtained in this way are then demodulated by a processor 13 and transferred into a data stream, which is converted by the processor 13 into analogue signals, which are then amplified by the audio amplifier 14 in order ultimately to be reproduced by the loudspeaker 15. Alternatively, it is also possible here to receive multi-media data, which are then reproduced optically. In addition, the fixed code rate for the channel coding of the useful
bits and the transmission rate are signalled to the receivers by the transmitter. It is therefore possible for the receivers to determine the respective variable code rate for the tail bits.
Figure 4 illustrates the method according to the invention as a flow diagram. In method step 16, the data from the data memory 1 are prepared and subjected to source coding by the source coder 2, In method step 17, the data stream produced in this way is divided up into parrallel data streams by the bit multiplexer 3, In method step 18, the individual coders carry out the channel coding, the useful bits being coded at a fixed code rate and the tail bits being coded at a variable code rate, which depends on the number of output bits from the individual coders 4 and 5 agreeing with the number of modulation symbols. This is achieved by the puncturing scheme, as it is known, for the tail bits at each level. In method step 20, in function block 6 the channel-coded data produced in this way is assigned to signal space points in order to produce the modulation symbols. In method step 21, the modulation symbols are subjected to OFDM modulation and, in method step 22, the OFDM signals are amplified and transmitted- In addition, in this case the fixed code rate for the useful bits and the transmission rate are also trans¬mitted from the transmitter to the receivers as a signal, so that the receivers are capable of calculating the variable code rates by using the received data.
1. Method for coded modulation of digital data, the
digital data having useful bits, the coded
modulation being carried out at a number of
levels, characterized in that the useful bits from
the coded modulation are divided up into parallel
signal streams, in that each signal stream is
subjected to channel coding by a respective coder
(4, 5) , the useful bits being channel-coded at at
least one fixed code rate, in that the useful bits
are supplemented by tail bits, in that the tail
bits are channel-coded in the respective coder (4,
5) at a variable code ^rate in order to achieve a
predefined number of bits for the respective coder
(4, 5), so that all the parallel-coded signal streams contain the same number of bits, and that the channel-coded data are assigned to signal space points in order to produce modulation symbols.
2. Method according to Claim 1, characterized in that the variable code rate of the tail bits is achieved by means of variable puncturing,
3. Method according to Claim 2, characterized in that all possible puncturing schemes are made known to a transmitter and receivers,
4. Method according to Claim 2, characterized in that all possible puncturing schemes for the tail bits
are calculated, the calculation method being known to a transmitter and receivers.
5. Method according to Claim 3 or 4, characterized in that the fixed code rate and a transmission rate are signalled to the receivers.
6 * Method according to one of the preceding claims, characterized in that the channel coding used is convolution coding,
7. Method according to one of the preceding claims, characterized in that the number of coded bits is produced by the number of modulation symbols multiplied by the level m of the modulation and divided by the number of levels.
8. Transmitter for carrying out the method according to one of Claims 1 to 7, characterized in that the transmitter has a multiplexer (3) for dividing the useful bits up into the parallel signal streams of a number of coders (4, 5) and means (6) for assigning the coded data to the signal space points, the number of coders (4, 5) corresponding to the number of levels in the modulation.
9. Receiver for carrying out the method according to Claim 3, 4, 5, 6 or 7, characterized in that the receiver has means for evaluating the signal (12, 13) for the demodulation of the modulation symbols and for the channel coding.
10. Method for coded modulation of digital data, substantially as hereinabove described and illustrated with reference to the accompanying drawings.
|Indian Patent Application Number||IN/PCT/2002/1054/CHE|
|PG Journal Number||07/2008|
|Date of Filing||09-Jul-2002|
|Name of Patentee||ROBERT BOSCH GMBH|
|Applicant Address||Postfach 30 02 20 70442 Stuttgart|
|PCT International Classification Number||H04L 25/49|
|PCT International Application Number||PCT/DE2001/004124|
|PCT International Filing date||2001-10-31|