DE10120156A1 - Turbotrellis coded modulation - Google Patents

Abstract

In a method of transmitting data over a data transmission channel, at least some of the bits of an incoming bit stream are passed through a turbo encoder to produce turbo-coded output bits, and words corresponding to symbol points in a constellation in a trellis code modulation scheme are generated using at least the bits that passed through the turbo encoder, possibly in conjunction with other bits that are not passed through the turbo encoder. Typically, the turbo coded bits are the least significant bits.

Description

Field of the Invention

This invention relates to the field of data transmission and, more particularly, to a Modulation scheme for transmitting data over a data transmission channel, for example in a discrete multi-tone modulation system.

Background of the Invention

To increase the efficiency of data transmission over a data transmission channel, the data is transmitted as symbols, each representing a number of bits. In For example, in a QAM (Quadrature Amplitude Modulation) system, the symbols represented by the amplitude and phase of the signals. Sixteen unique symbols, i. H. Combinations of amplitude and phase, for example, represent four bits at a time Symbols form a constellation of points in a phase-amplitude diagram. As the number of symbols increases, so does the possibility of transmission errors to. Forward error coding schemes are used to enable the receiver Detects errors and restores the correct transmitted symbol.

A preferred form of coding in data transmissions is convolutional coding is a bit-level coding scheme that depends on the preceding bit sequence. In the trellis coded modulation, the number of symbols is increased to provide redundancy to provide. Only certain transitions are allowed. In the event of an error, the Receiver the most likely correct transition with knowledge of all possible recognize permissible transitions.

In contrast to block codes that send data in predetermined blocks, come Convolutional codes do not cope well with tuft errors. Partly in response to this problem Turbo codes were developed. A turbo code essentially consists of two or a plurality of convolutional codes separated by an interleaver that the Input sequence of the first encoder processed. For example, see "Application of Turbo  Codes for Discrete Multi-Tone Modulation ", Hamid R. Sadjapour, AT Shannon Labs., 1996.

The turbo code is attracting more and more interest due to its greater coding gain. In a DSL (digital subscriber multiplex) system, a turbo code was used to replace a trellis code for better bit error rate (BER) performance. However, as the constellation size increases, the coding gain of the Remove turbo codes. This is because the redundant bits are the constellation size make it bigger.

An object of this invention is to increase the data transfer rate, for example in a DMT system.

Summary of the invention

According to the present invention, a method for transmitting data over a Data transmission channel provided, comprising receiving an incoming bit stream, Passing at least some of the bits through a turbo encoder to turbo coded ones Generate output bits and generate words corresponding to symbol points in one Constellation in a trellis code modulation scheme using at least the bits passed through the turbo encoder.

In this invention, the turbo encoder is preferably used to do just that to code the least significant bit (LSB) in the constellation, since the LSB for errors on is most sensitive. The data rate that can be achieved by this means is only a few dB of that Shannon capacity removed. The invention combines the powerful turbo code with a trellis-coded modulation scheme in order to optimize the data rate in a DMT Increase (discrete multi-tone) system.

In a DMT system, several subchannels are used to transmit data, each with different carriers and different QAM constellations that contain different numbers of bits per constellation point. The number of bits each constellation point is usually an integer and a subchannel is not can be used if it cannot support a data bit. According to the invention, a  Spread spectrum algorithm can be combined with a turbotrellis coded modulation, so that channels that transmit less than one bit of information can also be used. As a result, the overall channel capacity can be greatly increased.

Brief description of the drawings

The invention will now be described only by way of example with reference to the associated Drawings described in more detail, in which the following applies:

Fig. 1 shows an encoder according to the principles of the invention, for x and y <1, and y is the number of bits in each constellation point (symbol) wherein x;

Fig. 2 shows the encoder structure for x = 1 and y <1, the turbo coding rate being 2/3;

Fig. 3 shows an encoder structure for the case y = 1 and x <1;

Fig. 4 shows the encoder for x = y = 1, where the coding rate is ½;

Fig. 5 is a block diagram of the decoder;

Fig. 6 shows a constellation showing how the last three bits are determined; and

Fig. 7 is a constellation showing the determination of the most significant bits.

DETAILED DESCRIPTION OF THE PREFERRED EXAMPLES

The invention is used in connection with a DMT (discrete multi-tone system) described, which could typically contain 1000 subchannels, each for Transfer a different number of symbols that have a different number of bits are capable of, d. H. the number of constellation points for each subchannel can vary, and thus the number of bits per constellation point can vary.

Encoder

As shown in Fig. 1, part of an incoming bit stream is supplied to an encoder data block 10 , which is an addressable memory. Assuming 10 bits per symbol, including one check bit per two subchannels, the 1000 channels can transmit 9500 bits at a time. Thus, typically 9500 bits of an incoming bit stream are fed into the encoder. A fraction of these, typically 1500, are fed to encoder data block 10 .

The encoder can preferably be implemented as a parallel encoder, as in our Concurrent GB Application No. 0010330.9, filed April 18, 2000, the content of which is incorporated herein by reference.

In the example shown, three bits u ₁ , u ₂ , u _{3 are} successively output by the encoder data block 10 and three bits u ' ₁ , u' ₂ , u ' ₃ are output as interleaved data. The data bits u ₂ and u ₃ form components v ₀ , v _{1 of} the first output word v and the bit u ₁ forms the bit w _{1 of} the second output word w. Bit w ₀ is formed by turbo coding the groups of bits u ₁ , u ₂ , u ₃ and u ' ₁ , u' ₂ , u ' ₃ with recursive systematic convolutional encoders 12 , 14 after passing through respective shift registers 16 , 18 .

The constellation encoder structure used is similar to that used in an ADSL system. The binary word u = (u _{z '} , u _z'-1 , _... , U ₁ ) determines two binary words v = (v _z'-y , _... , V ₀ ) and w = (w _y-1 , ..., w ₀ ) (where z '= x + y - 1), which are used to find two constellation points (each containing x and y bits, respectively) in the encoder look-up table.

Fig. 1 shows the encoder structure for x <1 and y <1, the turbo encoder used being a systematic encoder with a coding rate of 3/4 which is interrupted at a rate of 1/2. The turbo encoder 20 consists of the two recursive systematic convolutional encoders 12 , 14 (RSC1 and RSC2). The coder RSC 1 assumes continuous data from Codiererdatenblock 10 and the coder RSC 2 takes nested data from the same data block 10th

The length of the data block depends on the number of data transmitted in each signal frame, 9500 bits in the example given above. Usually an integer number of data blocks are transmitted in each signal frame. The FIGS. 2 to 4 show the encoder structure for other values of x and y.

Fig. 2 shows the encoder structure for x = 1 and y <1, the turbo coding rate being 2/3. For the case y = 1 and x <1, the encoder structure shown in FIG. 3 is similar to that shown in FIG. 2.

FIG. 3 shows the encoder structure for the case x = y = 1, the coding rate being 1/2.

For y <1 (or x <1), an encoder structure similar to that shown in Figs. 1 to 4 can be used depending on the value of x (or y). The only difference is that a bit is transmitted using K subchannels, where y = 1 / K, using a spreading code.

If the spreading code used [b ₁ , b ₂ ,. . ., b _k ], can be 0 as [b ₁ , b ₂ ,. . ., b _k ] are transmitted, where (k = 1, 2,..., K), and 1 is called b₁, b₂,. . ., transfer. The constellation for each subchannel in the group of K subchannels uses a constellation with one bit per channel and the kth channel transmits the bit _bk . A total of K subchannels are required to transmit one data bit. The advantage of such an arrangement is that self-crosstalk can be greatly reduced if different spreading codes are used for different modems in the same binding group. Suitable spreading codes are described in IEEE Communications Letters, Volume 4, No. 3, pp. 80-82, March 2000, RV Sonalkar and RR Shively.

Decoder

The decoding procedure for turbotrellis coded modulation consists of the following steps:

• 1. Soft decoding of the least significant bit (LSB);
• 2. Hard decoding of the most significant bits (MSB);
• 3. Decode the LSB using a turbo decoder algorithm; and
• 4. Determine all data bits.

If an N-bit constellation is used for data transmission in a given subchannel, the constellation location can be represented by two-dimensional vectors: X _b = [b _{x M} , b _{x (M-1)} ,. . . b _x1 , 1] and Y _b = [b _yM , b _{y (M-1) ,.} . . b _y1 , 1], where M = N / 2 for an even number N and M = (N + 1) / 2 for an odd number N. The decoder is the same for both X _b and Y _b .

The received data should be (X, Y). If (-2 ^M + 1 + 2k) <X <(- 2 ^M + 1 + 2 (k + 1)), where k = 0, 1,. . ., 2 ^M -1, and X ₁ = (-2 ^M + 1 + 2k) and X ₂ = (-2 ^M + 1 + 2 (k + 1)), the decoder result depends on whether the last XX ₁ or assumes X ₂ , from the LSB. For N <1, the soft bit (log probability without constant) for the LSB in X is determined as

where σ ^{2 is} the noise power. The soft bit for the LSB in Y can similarly be obtained by swapping X for Y in the equation above.

If N = 1, the soft bit is

If N <1 and the spreading code [b ₁ , b ₂ ,. . ., b _k ], the soft bit can be calculated as

The soft bit output is sent to the turbo decoder circuit shown in FIG. 5. The turbo decoder consists of two LOG-MAP decoders 30 , 32 . Each contains a forward (α) iteration, a backward (β) iteration, and performs the final soft bit output calculation. The only difference is that the output signal contains not only the data bit but also the error check bit as its last iteration.

The reason the error check bit output is required is that the LSB is required to determine X (or Y) from two possible constellation points X ₁ and X ₂ (or Y ₁ and Y ₂ ) while some of these LSBs Error check bits are. A detailed example of a turbo decoder is in the Sadjapour article cited above and also in C. Berrou and A., "Near Optimal Error Correcting Coding and Decoding Turbo Codes", IEEE Trans. On Communications, Volume 44, No. 10, Oct. 1996 to find.

The soft output error check bits at time k are calculated as

P _ck1 = prob (b _ck = 1) = MAX _{(s, s')} [γ _ck1 (R _k , s, s') α _k-1 (s') β _k (s)]
P _ck0 = prob (b _ck = 0) = MAX _{(s, s')} [γ _ck0 (R _k , s, s') α _k-1 (s') β _k (s)]

where s is the state of the turbo encoder at time k and s' is the state at time k-1. R _k represents the received data. Β _k (s) is the probability in state s (time k) for the backward iteration and α _k-1 (s ') represents the probability in state s' (time k-1) for Forward _iteration . Γ _ck0 (R _k , s, s') and γ _ck1 (R _k , s, s') are the probability of the transition from state s' to s, the received data being R _k and the error _{check bit} 0 or 1.

After running through the turbo decoder, the LSBs are determined, and if N <1, the MSBs are still to be determined from two possible constellation points. X is taken as an example that has two possible values X ₁ or X ₂ (which are two adjacent points in the constellation). For two adjacent constellation points, the LSB for X ₁ and X _{2 must be} different. Hence X _b = [b _xM , b _{x (M-1) ,.} . . b _x1 , 1] can be determined from X ₁ and X ₂ by examining its LSB. Likewise, Y _b = [b _yM , b _{y (M-1) ,.} . . b _y1 , 1] can be determined. After X _b and Y _{b are} determined, the last received data bits are obtained for the following three cases:

If N is even, the last bits are [b _N , b _N-1,. . ., b ₁ ] = [b _xM , b _yM , b _{x (M-1)} , b _{y (M-1) ,.} . . b _x1 , b _y1 ].

If N = 3, the last three bits are determined by the constellation shown in FIG. 6, which is further tabulated in Table 1.

If N is an odd number and N <3, the low bit (N-5) can be determined in the same way as for the even N case, ie [b _N-5 , b _N-6,. . ., b ₁ ] = [b _{x (M-3)} , b _{y (M-3)} , b _{x (M-4)} , b _{y (M-4)} ,. . ., b _x1 , b _y1 ]. The 5 MSBs are determined according to the constellation in FIG. 7, which is also shown in table 2.

It can be seen that the use of a turbo code as described in combination enables better performance with a trellis code than with current Trellis codes used is possible. If a spread spectrum algorithm with a turbotrellis coded modulation is combined, it is possible to use channels that transmitted less than one bit of information, resulting in a significant increase in Channel capacity leads.

The blocks described can be implemented in a digital signal processor using Standard DSP techniques known to those skilled in the art can be implemented.

Claims (18)

1. Method for transmitting data over a data transmission channel, comprising receiving an incoming bit stream, routing at least some of the bits through a turbo encoder to generate turbo encoded output bits generate, and generate words corresponding to symbol points in one Constellation in a trellis code modulation scheme using at least the bits passed through the turbo encoder. 2. The method of claim 1, wherein the bits passed through the turbo encoder are the least significant bits (LSBs) of the incoming data and the most significant bits are forwarded directly to the output word. 3. The method of claim 1, wherein a part of the incoming bit stream one Memory is supplied and nested groups of bits from the memory recursive systematic convolutional encoders are fed to the generate turbo coded output bits. 4. The method of claim 1, wherein the symbols are in a plurality of Subchannels are transmitted in a discrete multi-tone (DMT) system. 5. The method of claim 1, wherein for at least some of the subchannels Constellation symbol points contain less than one bit and for such Subchannels an entire bit over K subchannels using a Spreading codes is transmitted. 6. The method according to any one of claims 1 to 5, wherein the words error check bits contain.   7. The method of claim 1, wherein the words are used to derive the Obtain constellation points from an encoder lookup table. 8. Encoder for encoding an incoming data stream for transmission over a data transmission channel, with an encoder data block for storage part of the incoming data stream, a first recursive systematic Convolutional encoder that receives continuous data from the encoder data block, and a second recursive systematic convolutional encoder for receiving of interleaved data from the encoder data block, the Convolutional encoder at least the least significant bit of an output data word output that a symbol point in a constellation in a trellis code Modulation scheme forms. 9. The encoder of claim 8, further comprising a shift register associated with connected to the input of each recursive systematic convolutional encoder where each cell of the shift register is a respective bit of that Encoder data block received. 10. The encoder of claim 8, wherein for x = y = 1, where x and y are the number of bits are in two constellations, the inputs of the recursive systematic Convolutional encoder are directly connected to the encoder data block. 11. Method for decoding a signal modulated with a turbotrellis code with:

• a) soft decoding of the received signal for the least significant bits;
• b) hard decoding the input signal for the most significant bits;
• c) decoding the least significant bit using a turbo decoder algorithm; and
• d) determining all data bits.

12. The method of claim 11, wherein the received signal is processed to determine the soft bit (log probability without constant) for the LSB as
where σ ^{2 is} the noise power. 13. The method of claim 12, wherein the soft bit to a turbo decoder is forwarded to generate a data bit and an error check bit. 14. The method of claim 13, wherein after determining the LSBs, the MSBs determined from possible constellation points by examining the LSBs will. 15. The method of claim 11, wherein an LSM [LSB] and MSB decoder for different constellations is combined. 16. Decoder device for decoding a received, with a Turbotrelliscode modulated signal, with a pair of decoders used for Receiving soft input bits are connected, one Interleaver, a deinterleaver, and wherein the Decoder of the soft input bits one data bit and at least one check bit deduce.   17. The decoder apparatus of claim 16, wherein each decoder has a forward and backward iteration and a calculation of the last soft bit output carries out. 18. The decoder device according to claim 16, wherein the decoders LOG-MAP- Are decoders.