DE102010043610A1 - Method of decoding data received in radio receiver as frame, involves carrying out turbo-demodulation and convolutional decoding processes repeatedly, and performing convolutional decoding process based on soft output Viterbi algorithm - Google Patents

Abstract

The method involves carrying out de-interleaving processes before performing turbo-demodulation process and convolutional decoding process. The turbo-demodulation and convolutional decoding processes are carried out repeatedly. The convolutional decoding process is carried out based on soft output Viterbi algorithm (SOVA). Independent claims are included for the following: (1) device for decoding data received as frame; and (2) computer program product comprising machine-readable medium storing instructions for decoding data received as frame.

Description

The invention relates to a method and a device for decoding data received in frames. In digital broadcasting systems, such. For example, according to the DAB (Digital Audio Broadcasting) standard, a multi-carrier method wherein each carrier is modulated with a differential 4-phase shift keying (4-DPSK) is used. Phase shift keying (PSK) is a phase modulation for digital signals. In this method, the signal has a constant frequency and a constant amplitude. The phase angle of the carrier signal changes abruptly in the rhythm of the digital modulation signal. For a DAB transmission, several audio streams together with also possible pure data services are combined to form a so-called ensemble with a high data rate. The resulting multiplex is modulated using OFDM (Orthogonal Frequency Division Multiplex). OFDM is a modulation method which uses a plurality of orthogonal carrier signals for digital data transmission. Each carrier is modulated separately. Furthermore, according to the DAB standard ( ETSI 300.401 ) for spreading bundle springs a frequency interleaving and a time interleaving used. For error correction is a puncturable convolutional code (Convolutional Code). Furthermore, according to the DAB standard in the receiver, the demodulation is performed by a so-called soft-decision demodulation and the convolutional decoding using a corresponding Viterbi decoder. The use of a convolutional code offers the possibility of forward error correction. In this case, a higher level of protection against transmission errors is achieved by additionally introduced redundancy. The information content of the individual user data points is distributed over several positions of the code word by the mathematical method of convolution. The convolutional decoding in the receiver using a Viterbi decoder is based on processing individual positions of a codeword with specific probabilities. The Viterbi algorithm is a dynamic programming algorithm for determining the most probable sequence of hidden states. The Viterbi decoder can process both binary and continuous input sequences.

State of the art

By " C. Berrou, A. Gelavieux, P. Thitimajshima: Near Shannon limit error correcting coding and decoding Proceedings of IEEE International Conference on Communications, 1993; ICC 93. Geneva "Is the principle known for turbo decoding parallel concatenated convolutional codes.

By " Dr. Ing. Thomas May: Differential Modulation and Channel Coding in Broadband OFDM Radio Transmission Systems Dissertation; TU Hamburg-Harburg "Is known that the combination of differential modulation and simple convolutional code formally corresponds to a concatenated convolutional code and thus the principle of turbo decoding is applicable.

In the EP 1 284 548 A2 For example, a method and apparatus for assigning pointers to data received in frames will be described. In this case, it is possible to achieve dynamic allocation of the data to a channel in a memory.

Disclosure of the invention

The invention has for its object a method and an apparatus for decoding data received in frames, such. B. data which are received via digital broadcast signals to improve so that at the same signal to noise ratio, a lower error rate can be achieved or that at lower signal to noise ratio, the maximum error rate is exceeded.

This object is achieved with a method and a device for decoding received in frame data according to the features of claims 1 and 8.

Hereinafter, at least a turbo-demodulation, a de-interleaving, and a convolutional decoding are performed in the above-mentioned method of decoding data received in frames. Turbo-demodulation is to be understood here as a method for decoding, wherein a plurality of decoders are connected in parallel or in series with one another. For this purpose, statistical information is exchanged between the decoders, the decoding process being carried out iteratively. By de-interleaving is meant a so-called descrambling. The data which was scrambled on the transmitter side for the purpose of reducing burst errors is put back in the correct order. Additional steps for decoding the differentially modulated and convolutional encoded data are also contemplated in the prior art. This includes, for example, that by a corresponding Fourier transform, z. As FFT (Fast Fourier Transform) or DFT (Discrete Fourier Transform), each symbol is transformed into the corresponding frequency range. Such a Fourier transformation is carried out, for example, as one of the first essential method steps in the decoding of the data. According to the invention, the turbo demodulation, the de-interleaving and the convolutional decoding are temporally rearranged such that the de-interleaving is performed before the turbo-demodulation. In this case, the deinterleaving may consist of a combination of frequency deinterleaving and time deinterleaving. Furthermore, the decoding method according to the invention is characterized in that the turbo-demodulation and the convolutional decoding are performed in an iterative processing loop. The convolutional decoding is based at least partially on a so-called SOVA (Soff-Output-Viterbi algorithm). The SOVA generates in addition to the decoding of the individual bits also information regarding their reliability. The SOVA is a modified Viterbi algorithm. In this case, the error probability is calculated for the SOVA for each information bit and the so-called soft outputs output the probability of the decisive bit for the downstream soft-decision process.

Due to this special time arrangement of the method steps, the use or the use of the turbo-decoding principle for the differentially modulated and convolution-coded data is also in systems such. As digital radio reception systems, practically usable without the additional high storage or computing power required. Thus, a lower error rate can be achieved with the same signal to noise ratio. Furthermore, with a lower signal-to-noise ratio of, for example, 2.5 dB, the maximum error rate can be undershot. This results in an AWGN (Additive White Gaussian Noise) channel gain of up to 2.5 dB. By AWGN is meant additive white Gaussian noise (AWGR), which denotes white noise whose signal amplitudes are Gaussian distributed and whose spectral noise power density is constant.

In the reconstruction of the multiplex signal, it is preferable to perform the assignment of data, in particular the assignment of the real part and the imaginary part of a bit, with the aid of pointers and a corresponding interleaver memory. With the help of such a pointer concept related data such. For example, real-part bits and imaginary-part bits, which may be placed at different locations and even in different codewords, are again reliably assigned to each other. This has the advantage that the data can remain in place in the memory and the assignment is carried out solely by pointers which are stored in a table in the memory.

Preferably, the inventive method in a receiving chain for a digital radio receiver according to a digital broadcasting system such. DAB, DAB + and DMB. The method may further be used for transmission methods which use differential modulation such as DAB, DAB + and DMB. The method can be used for all services of the proposed transmission method according to their standards.

Preferably, so-called soft outputs are generated by the soft output Viterbi algorithm used only for non-punctured locations in the convolutional code. Since the demodulator can not use the punctured locations in the soft outputs generated by the iterative processing loop from the convolutional decoder, the corresponding expense can be reduced by this preferred measure. Soft outputs indicate a probability of whether a bit is a "0" or a "1". In contrast, hard outputs are the decided bits. Each hard output is thus assigned a soft output. The punctured digits are bit positions of a codeword, which have been omitted by a so-called puncturing transmitter side. The sender-side puncturing allows codeword lengths to be designed exactly to a specific frame length for subsequent data transmissions. Therefore, a de-puncturing is provided on the receiver side.

Furthermore, it is preferred that in the soft outputs of the demodulator the puncturing sites are replaced by so-called zeros. As a result, the soft outputs of the demodulator are not depunctured at all, but a "0" is inserted at each punctuation point by the convolutional decoder or the soft output Viterbi algorithm, without accessing the memory of the soft output values. Thereby, the memory for the punctured positions can be saved until the final decoding of the convolutional code. Preferably, only after the last iteration the de-puncturing is done at the output of the hard outputs.

Furthermore, it is preferable to use the values of the soft outputs also as a termination criterion of the iterative processing loop. In this case, as soon as the value of a soft output exceeds a predetermined limit value, the decoding result is regarded as sufficiently reliable and thus the iterative processing is aborted. Preferably, after an initially defined maximum number of iterations, if after this maximum number of iterations very small amounts of the soft outputs are still present, this is used as an indicator for a so-called erasure decoding of an outer code or an error concealment. These measures are necessary to avoid possible damage to the reconstructed signal due to imperfections in particularly sensitive information. Since it is possible in the case of a differential modulation in the time direction to demodulate the OFDM subcarriers independently of one another, the cost of the iterative processing can thus be further reduced by iteratively reinterpreting only the subcarriers demodulating, where the convolutional decoder provides soft outputs with very small amounts, since this implies a rating as being particularly uncertain. As a result, a new demodulation of the already decided decided bits can be saved.

The device according to the invention for decoding data received in frames has a demodulator and a convolutional decoder. The demodulator serves for the demodulation of the differentially modulated data and according to the invention is designed as a turbo-demodulator. Furthermore, the decoding device has at least one de-interleaver, wherein the de-interleaver is arranged in front of the turbo-demodulator. Among other things, it is provided that the de-interleaver is composed of two separate de-interleavers, a time and a frequency de-interleaver. According to the invention, the convolutional decoder is designed as a SOVA decoder and, together with the demodulator, forms an iterative processing loop for the data received for demodulation.

Advantageously, the device has an interleaver memory for reconfiguring the multiplexed signal. By such an interleaver memory, it is possible a so-called pointer concept for the assignment of z. Real-bit and imaginary-part bits, which may be placed at different locations and even at different codewords.

The invention will now be described by way of example with reference to the accompanying drawings.

Show it:

1 A block diagram of one of the prior art ( ETSI 300.401 ) known receiving chain of a DAB system;

2 A block diagram showing one of the prior art (Dr. Ing.

Thomas May: Differential modulation and channel coding in broadband OFDM radio transmission systems ") known principle of turbo-decoding for DPSK and convolutional code;

3 A block diagram of the method according to the invention and of the device according to the invention for the decoding of data received in frames; and

4 A further block diagram of the method according to the invention and of the device according to the invention for decoding data received in frames.

1 shows a typical receive chain of a DAB system as z. B. from the Standard ETSI 300.401 is known for digital broadcasting systems. The receive chain provides for the data received in frames of an OFDM transmission, the data being differentially modulated and convolutionally encoded, by means of a Fourier transform 10 to transform such that each symbol is transformed back into the corresponding frequency domain. In a subsequent differential demodulation 11 the complex signal states are recovered. Since the data is subjected to frequency and time interleaving according to the DAB standard in the sender for spreading burst errors, the demodulated data in the receiver becomes subsequent to the demodulation 11 a de-interleaver, which consists of a frequency 12 and a time-de-interleaver 13 exists, subject. Finally, the typical DAB receive chain sees a depuncture 14 and a soft-decision demodulation by a corresponding Viterbi decoder 15 in front.

2 shows the principle of turbo decoding for differential PSK (DPSK) modulated and convolutionally encoded signals. In the "Dr. Ing. Thomas May: Differential Modulation and Channel Coding in Broadband OFDM Radio Transmission Systems "is derived, for example, that the combination of differential modulation and a simple convolutional code formally corresponds to a concatenated convolutional code and thus the turbo-decoding principle is applicable. This will be a so-called SOVA demodulator 17 , Folding decoder 20 , De-interleaver, or the FrequenzDe interleaver 12 and Time De-interleaver 13 , together with an interleaver 19 interconnected in an iterative processing loop. However, a practical implementation of this concept for a DAB system is very expensive. In a through 2 As shown, the interleaver would be in the DAB transfer mode 1 for example, over 288 OFDM symbols are running. In combination with the FrequencyDe interleaver, this leads to a high storage and computing power requirement.

To avoid an increased storage or computing power requirement when using the in 2 shown turbo decoding is the de-interleaving, consisting of frequency de-interleaver 12 and time-de-interleaver 13 , before demodulation 17 realized. This is possible because the frequency interleaver rule refers to the position of the carriers and is not changed by the demodulation. The time interleaver rule repeats every 16 bits. The bits modulated on the real and imaginary parts of a subcarrier are separated by a mapping. They are placed at locations in the codeword that differ by the number of carriers. Because in the DAB system in all modes the Carrier number is a multiple of 16, the bits of the real part and the imaginary part have the same time interleaver rule. This also makes it possible to use the TimeDe interleaver 13 to place in front of the demodulation. Such an arrangement in the demodulator 17 the Fourier transform 10 , the frequency-de-interleaver 12 and the TimeDe interleaver 13 follows is in 3 shown. The iterative processing loop for turbo demodulation becomes as in 3 also shown through the functional blocks SOVA demodulator 17 and a SOVA convolutional decoder 20 along with the required depuncturing 14 and a puncturing puncture unit 21 the soft outputs 22 realized. To the repeated puncturing 21 and depuncturing 14 to simplify in the iterative loop, the demodulated soft outputs 22 only once at the depuncturing 14 added in the form that zeros are used at the puncturing sites. In later demodulation steps, the soft outputs will be 22 overwritten directly at the right places.

4 shows a block diagram based on the arrangement in 3 where the depunctation 14 after the SOVA convolutional decoder 20 is provided. Thus, within the iteration loop, the depuncturing 14 and punctuation 21 be saved. As a result, the required computing power requirement can be further reduced. In this case, memory for the punctured digits can also be saved until the final decoding of the convolutional code. The soft outputs 22 of the demodulator 17 are not depuncted at all in this case. The SOVA convolutional decoder 20 always sets a zero for the punctuation points, without accessing the memory of the soft output values. Only after the last iteration will the depunctation 14 at the output of the hard outputs 23 performed.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 1284548 A2 [0004]

Cited non-patent literature

• ETSI 300.401 [0001]
• C. Berrou, A. Gelavieux, P. Thitimajshima: Near Shannon limit error correcting coding and decoding Proceedings of IEEE International Conference on Communications, 1993; ICC 93. Geneva [0002]
• Dr. Ing. Thomas May: Differential Modulation and Channel Coding in Broadband OFDM Radio Transmission Systems Dissertation; TU Hamburg-Harburg [0003]
• ETSI 300.401 [0018]
• Standard ETSI 300.401 [0023]

Claims (13)

Method for decoding data received in frames, the data received in frames being differentially modulated and convolutionally coded, characterized in that at least one turbo-demodulation ( 17 ) and a deinterleaving ( 12 . 13 ) and a convolutional decoding ( 20 ), whereby de-interleaving ( 12 . 13 ) before turbo-demodulation ( 17 ), wherein the turbo-demodulation ( 17 ) and the convolutional decoding ( 20 ) are performed by an iterative processing, wherein the convolutional decoding ( 20 ) is performed based at least in part on a soft output Viterbi algorithm. A method according to claim 1, characterized in that a dynamic assignment of data is provided by pointers, wherein the assignment is used in particular the assignment of real part and imaginary part of a bit. A method according to claim 1 or 2, characterized in that the received data in frame via digital broadcast signals are received. Method according to one of the preceding claims, characterized in that the Soft-Outbut-Viterbi algorithm soft outputs ( 22 ) are generated only for non-punctured locations in the convolutional code. Method according to one of the preceding claims, characterized in that in the soft outputs ( 22 ) of the demodulator ( 17 ) Puncturing points are replaced by zeros. Method according to one of the preceding claims, characterized in that a de-puncturing ( 14 ) only when outputting the hard outputs ( 23 ) is made. Method according to one of the preceding claims, characterized in that the values of the soft outputs ( 22 ) also serve as a termination criterion for iterative processing. Device for decoding data received in frames, the device comprising a demodulator ( 17 ) for differentially modulated data and a convolutional decoder ( 20 ), characterized in that the demodulator ( 17 ) as a turbo-demodulator ( 17 ), wherein at least one de-interleaver ( 12 . 13 ) in front of the turbo-demodulator ( 17 ), the convolutional decoder being a SOVA decoder ( 20 ) is formed. Apparatus according to claim 8, characterized in that the device has a memory, wherein the memory is designed as an interleaver memory. Computer program product comprising program parts for carrying out a method according to at least one of the preceding claims 1 to 7. Machine-readable, in particular computer-readable, data structure produced by a method according to at least one of claims 1 to 7 and / or by at least one computer program product according to claim 10. Machine-readable, in particular computer-readable, data carrier on which at least one computer program product according to claim 10 is recorded and / or stored and / or on which at least one data structure according to claim 11 is kept ready for retrieval. Use of a device according to claims 8 to 9 in a radio receiver for decoding via digital broadcast signals in frames of received data.

Images