Versatile serial concatenated convolutional codes
AssigneeSTMicroelectronics S.r.l.
InventorsFabio Scalise, Fabio Osnato, Stefano Valle, Massimiliano Siti, Sergio Benedetto
Life of this patent
Estimating the term…
Abstract
An input digital signal is encoded by subjecting it to a first convolutional coding step followed by an interleaving step and a second convolutional coding step. The serial concatenated convolutional coded signal thus obtained is then subjected to modulation by means of a two-dimensional modulation scheme such as M-PSK or M-QAM. The corresponding decoding process involves an iterative decoding algorithm based on cascaded logarithmic soft-input soft-output processing steps.
Claims82
A method of encoding an input digital signal, comprising:
- subjecting said input digital signal to a cascade of a first convolutional coding step, an interleaving step and a second convolutional coding step to generate a serial concatenated convolutional coded signal;
- and
- modulating said serial concatenated convolutional coded signal by a two-dimensional modulation scheme, wherein at least one of said first and second convolutional coding steps generates a respective convolutional coded signal to be transmitted with a required code rate and further including at least one rate-matching step implemented by puncturing said respective convolutional coded signal by determining what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate.
A method of encoding an input digital signal, comprising:
- subjecting said input digital signal to a cascade of a first convolutional coding step, an interleaving step and a second convolutional coding step to generate a serial concatenated convolutional coded signal;
- modulating said serial concatenated convolutional coded signal by a two-dimensional modulation scheme;
- generating, by said first convolutional coding step, an outer code having an outer code rate;
- generating, by means of said second convolutional coding step, an inner code having an inner code rate;
- selecting said outer code rate as ro=(rc)γ;
- and
- selecting said inner code rate as ri=(rc)1−γ;
- where rc is the product of said inner and outer rates and γ is an optimized rate splitting parameter.
A method of decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the method comprising:
- subjecting said coded signal to a cascade of a first decoding step, an interleaving step and a second decoding step;
- and
- subjecting said coded signal to a log-likelihood ratio computation of coded symbols associated with said coded signal, wherein said coded signal is subjected to row-column interleaving before being modulated by means of said two-dimensional modulation schemes, wherein decoding said coded signal involves the step of subjecting said coded signal to a complementary row-column interleaving function and wherein said log-likelihood ratios are written row wise and read column wise.
A method of decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the method comprising:
- subjecting said coded signal to a cascade of a first decoding step, an interleaving step and a second decoding step;
- and
- subjecting said coded signal to a log-likelihood ratio computation of coded symbols associated with said coded signal, wherein said log-likelihood ratio computation include the steps of:
- evaluating log-likelihood values of all modulation signal, said signals being designated si, according to the relationship:
- L i = exp ( - [LeftDoubleBracketingBar] r - s i [RightDoubleBracketingBar] 2 N 0 )
- where r is the received signal and N0 is a factor proportional to the noise variance, and
- subsequently computing the log-likelihood ratios of the individual symbols as follows:
- λ k = log ( ∑ i :
- b k ( s i ) = 1 L i ) - log ( ∑ i :
- b k ( s i ) = 0 L i )
- where bk(.) are the mapping functions that associate to the i-th modulation signal the value of the k-th symbol of its label.
A method of decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the method comprising:
- subjecting said coded signal to a cascade of a first decoding step, an interleaving step and a second decoding step;
- and
- subjecting said coded signal to a log-likelihood ratio computation of coded symbols associated with said coded signal, wherein said coded signal is generated by means of cascaded first and second convolutional coding steps, wherein at least one of said first and second convolutional coding steps generates a respective convolutional coded signal to be transmitted with a required code rate and further includes at least one rate-matching step implemented by subjecting said respective convolutional coded signal to puncturing to determine what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate, whereby certain log-likelihood ratios are missing in said coded signal subject to said log-likelihood ratio computation, the decoding method further including the step of replacing the missing log-likelihood ratios by means of dummy log-likelihood ratios.
A method of decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the method comprising:
- subjecting said coded signal to a cascade of a first decoding step, an interleaving step and a second decoding step;
- and
- subjecting said coded signal to a log-likelihood ratio computation of coded symbols associated with said coded signal, wherein the log-likelihood ratios out of said second decoding step are fed to said first decoding step to improve the log-likelihood ratio computation performed therein according to an iterative process, wherein the serial concatenated convolutional coded signal was generated by means of cascaded first and second convolutional coding steps, wherein said first coding step generates a first convolutional coded signal to be transmitted with a required code rate and further includes at least one of:
- a rate-matching step implemented by subjecting said respective convolutional coded signal to puncturing to determine what symbols of said first convolutional coded signal are to be transmitted in order to match said required code rate;
- and
- an interleaving step of said first convolutional coded signal, the decoding method involving the step of subjecting said log-likelihood ratios to at least one of a rate-matching function and an interleaving function matching said rate matching step and said interleaving step performed on said first convolutional coded signal, respectively.
An apparatus for encoding an input digital signal, comprising
- a cascade of a first convolutional coding module, an interleaving module and a second convolutional coding module to generate a serial concatenated convolutional coded signal;
- and
- a modulation module for subjecting said serial concatenated convolutional coded signal to modulation by means of a two-dimensional modulation scheme, wherein at least one of said first and second convolutional coding modules generates a respective convolutional coded signal to be transmitted with a required code rate and further including at least one rate-matching module for subjecting said respective convolutional coded signal to puncturing to determine what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate.
An apparatus for encoding an input digital signal, comprising
- a cascade of a first convolutional coding module, an interleaving module and a second convolutional coding module to generate a serial concatenated convolutional coded signal;
- and
- a modulation module for subjecting said serial concatenated convolutional coded signal to modulation by means of a two-dimensional modulation scheme, wherein:
- said first convolutional coding module generates an outer code having an outer code rate-selected as r0=(rc)γ,
- said second convolutional coding module generates an inner code having an inner code rate-selected as ri=(rc)1−γ,
- wherein rc is the product of said inner and outer rates and γ is an optimized rate splitting parameter.
An apparatus for decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the apparatus comprising:
- a cascade of a first decoding module, an interleaving module and a second decoding module, wherein the coded signal was subjected to row-column interleaving before being modulated by means of said two-dimensional modulation schemes;
- a module for subjecting said coded signal to a log-likelihood ratio computation of coded symbols associated with said coded signal;
- and
- an interleaver for subjecting said coded signal to a complementary row-column interleaving function and wherein said log-likelihood ratios are written row wise and read column wise.
An apparatus for decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the apparatus comprising:
- a cascade of a first decoding module, an interleaving module and a second decoding module;
- and
- a module for subjecting said coded signal to a log-likelihood ratio computation of the coded symbols associated with said coded signal, wherein said log-likelihood ratio computation includes the steps of:
- evaluating log-likelihood values of all modulation signal, said signals being designated si, according to the relationship:
- L i = exp ( - [LeftDoubleBracketingBar] r - s i [RightDoubleBracketingBar] 2 N 0 )
- where r is the received signal and N0 is a factor proportional to the noise variance, and
- subsequently computing the log-likelihood ratios of the individual symbols as follows:
- λ k = log ( ∑ i :
- b k ( s i ) = 1 L i ) - log ( ∑ i :
- b k ( s i ) = 0 L i )
- where bk(.) are the mapping functions that associate to the i-th modulation signal the value of the k-th symbol of its label.
An apparatus for decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the apparatus comprising:
- a cascade of a first decoding module, an interleaving module and a second decoding module;
- a module for subjecting said coded signal to a log-likelihood ratio computation of the coded symbols associated with said coded signal, wherein the serial concatenated convolutional coded signal was generated by means of cascaded first and second convolutional coding steps, wherein at least one of said first and second convolutional coding steps generates a respective convolutional coded signal to be transmitted with a required code rate and further includes at least one rate-matching step implemented by subjecting said respective convolutional coded signal to puncturing to determine what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate, whereby certain log-likelihood ratios are missing in said coded signal subject to said log-likelihood ratio computation;
- and
- an inserter module for replacing the missing log-likelihood ratios by means of dummy log-likelihood ratios.
An apparatus for decoding a serial concatenated convolutional coded signal modulated by means of a two-dimensional modulation scheme, the apparatus comprising:
- a cascade of a first decoding module, an interleaving module and a second decoding module;
- and
- a module for subjecting said coded signal to a log-likelihood ratio computation of the coded symbols associated with said coded signal, wherein the log-likelihood ratios out of said second decoding module are fed to said first decoding module to improve the log-likelihood ratio computation performed therein according to an iterative process, wherein the serial concatenated convolutional coded signal was generated by means of cascaded first and second convolutional coding steps, wherein said first coding step generates a first convolutional coded signal to be transmitted with a required code rate and further includes at least one of:
- a rate-matching step implemented by subjecting said respective convolutional coded signal to puncturing to determine what symbols of said first convolutional coded signal are to be transmitted in order to match said required code rate, and
- an interleaving step of said first convolutional coded signal,
- the apparatus further comprising at least one of a respective rate-matching module and a respective interleaver for subjecting said log-likelihood ratios to at least one of a rate-matching function and an interleaving function matching said at least one of a rate matching step and an interleaving step performed on said first convolutional coded signal, respectively.
A method of encoding an input digital signal, the method comprising the steps of:
- subjecting said input signal to a cascade of a first convolutional coding step, an interleaving step and a second convolutional coding step to generate a serial concatenated convolutional coded signal;
- and
- subjecting said serial concatenated convolutional coded signal to modulation by a two-dimensional modulation scheme, wherein said two-dimensional modulation scheme is selected from the group consisting of M-PSK and M-QAM, wherein at least one of said first and second convolutional coding steps generates a respective convolutional coded signal to be transmitted with a required code rate and further including at least one rate-matching step implemented by subjecting said respective convolutional coded signal to puncturing to determine what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate.
An apparatus for encoding an input digital signal, comprising:
- a cascade of a first convolutional coding module, an interleaving module and a second convolutional coding module to generate a serial concatenated convolutional coded signal;
- and
- a modulation module for subjecting said serial concatenated convolutional coded signal to modulation by means of a two-dimensional modulation scheme, wherein said two-dimensional modulation scheme is selected in the group consisting of M-PSK and M-QAM, wherein at least one of said first and second convolutional coding modules generates a respective convolutional coded signal to be transmitted with a required code rate and further including at least one rate-matching module for subjecting said respective convolutional coded signal to puncturing to determine what symbols of said respective convolutional coded signal are to be transmitted in order to match said required code rate.
Dependencies are inferred from the claim text.