diff options
| author | Tom Rondeau <tom@trondeau.com> | 2015-06-17 16:24:19 -0400 |
|---|---|---|
| committer | Tom Rondeau <tom@trondeau.com> | 2015-10-15 10:40:24 -0400 |
| commit | 7c0ff8a410528ca3eed769c595d91e2b89f30a0f (patch) | |
| tree | ce6a7aea5a6e48803546e415e811558111cc4449 | |
| parent | 8f717de2ffb3b5f5a9d1aab5cd2876010b631432 (diff) | |
fec: Updated docs for Forward Error Correction section in manual.
Adds information on the different coders, especially on the LDPC
implementations.
| -rw-r--r-- | gr-fec/doc/fec.dox | 108 |
1 files changed, 92 insertions, 16 deletions
diff --git a/gr-fec/doc/fec.dox b/gr-fec/doc/fec.dox index 9d48214d9e..fc514286b4 100644 --- a/gr-fec/doc/fec.dox +++ b/gr-fec/doc/fec.dox @@ -227,6 +227,7 @@ fec/fecapi_async_encoders.grc and fec/fecapi_async_decoders.grc. GNU Radio currently has a minor subset of coders available: + Coders: \li gr::fec::code::dummy_encoder @@ -234,8 +235,9 @@ Coders: \li gr::fec::code::cc_encoder \li gr::fec::code::ccsds_encoder \li gr::fec::code::polar_encoder -\li gr::fec::code::ldpc_R_U_encoder +\li gr::fec::ldpc_encoder \li gr::fec::code::ldpc_gen_mtrx_encoder +\li gr::fec::code::ldpc_R_U_encoder Decoders: \li gr::fec::code::dummy_decoder @@ -243,30 +245,103 @@ Decoders: \li gr::fec::code::cc_decoder \li gr::fec::code::polar_decoder_sc \li gr::fec::code::polar_decoder_sc_list +\li gr::fec::ldpc_decoder \li gr::fec::code::ldpc_bit_flip_decoder -When building a new FECAPI encoder or decoder variable, the dummy -encoder/decoder block would be a good place to start. This coding set -does no processing on the data. For the encoder, each bit is simply -passed through directly. For the dummy decoder, the input data are -floats, so -1's become 0 and 1's stay as 1, but nothing else is done -to the data. Mainly, these blocks are used for references and to make -it easy to compare implementations with and without codes by easily -dropping in these objects instead of restructuring the entire -flowgraph. The ber_curve_gen.grc example file uses the dummy codes to -show the curve to compare against the actual codes. + +\subsubsection fec_dummy Dummy Encoder/Decoder + +When building a new FECAPI encoder or decoder variable, the +gr::fec::code::dummy_encoder / gr::fec::code::dummy_decoder blocks are a good +place to start. This coding set does no processing on the data. For +the encoder, each bit is simply passed through directly. For the dummy +decoder, the input data are floats, so -1's become 0 and 1's stay as +1, but nothing else is done to the data. Mainly, these blocks are used +for references and to make it easy to compare implementations with and +without codes by easily dropping in these objects instead of +restructuring the entire flowgraph. The ber_curve_gen.grc example file +uses the dummy codes to show the curve to compare against the actual +codes. + + +\subsubsection fec_repetition Repetition Encoder/Decoder + +The simplest example of FEC is the repetition code in +gr::fec::code::repetition_encoder and +gr::fec::code::repetition_decoder. The basic idea is to repeat the +information several times so that even if parts of the received +message are corrupted, the majority of the data is received correctly +and the original message can be discerned. The repetition decoder is +not particularly sophisticated and other coders offer better +performance, but it is useful for comparison. + + +\subsubsection fec_cc Convolutional Encoder/Decoder Although mentioned in the convolutional coder and decoder classes, it -is worth another mention. The cc_encoder is a generic convolutional -encoder that can take any value of K, rate, and polynomials to encode -a data stream. However, the cc_decoder is not as general, even though -it is technically parameterized as such. The cc_decoder block +is worth another mention. The gr::fec::code::cc_encoder is a generic +convolutional encoder that can take any value of K, rate, and +polynomials to encode a data stream. However, the +gr::fec::code::cc_decoder is not as general, even though it is +technically parameterized as such. The gr::fec::code::cc_decoder block currently <i>only</i> uses K=7, rate=2, and two polynomials (because the rate is two). We can, in fact, alter the polynomials, but a default of [109, 79] is typically. Eventually, we will make this block more generic for different rates and constraint lengths and take this particular code implementation as the set CCSDS decoder, much like we -have the ccsds_encoder class. +have the gr::fec::code::ccsds_encoder class. + + +\subsubsection fec_ldpc LDPC Encoders and Decoders + +There are a few different LDPC encoders and decoders available in +gr-fec. The LDPC decoder used does not necessarily depend on the LDPC +encoder used; different decoders can be used with a given +encoder. Python scripts are included in gr-fec for constructing parity +check matrices with specified properties. + +The simplest LDPC encoder is +gr::fec::code::ldpc_gen_mtrx_encoder. From a generator matrix in +systematic form G=[I P], the codeword x is generated from the +information word s via simple matrix multiplication: x=G*s. This +encoder can be used if either a parity check matrix H or a generator +matrix G is input. (If a H matrix is provided, the class +gr::fec::code::ldpc_HorG_mtrx will convert it to the corresponding G +matrix for use with this encoder.) + +The gr::fec::ldpc_encoder uses the same method to generate a codeword, +x=G*s. However, the provided matrix file must be for a parity check +matrix. + +The gr::fec::code::ldpc_R_U_encoder uses a reduced complexity +algorithm. Compared to the gr::fec::code::ldpc_gen_mtrx_encoder, this +requires orders of magnitude fewer operations at each encoding +step. This is accomplished by completing a significant amount of the +complex matrix manipulation (including inverse, multiplication, and +Gaussian elimination operations) during preprocessing. The +disadvantage of this encoder is that it requires a specially formatted +matrix. However, GNU Radio includes Python scripts to format a +standard parity check matrix appropriately for this encoder, as well +as a small library of encoding-ready matrices for use. More details +for using this encoder are in the gr::fec::code::ldpc_R_U_mtrx class +reference in the manual. + +The simplest LDPC decoder is probably the +gr::fec::code::ldpc_bit_flip_decoder, a hard decision decoding +scheme. The decoder seeks to find the codeword that was most likely +sent, which must satisfy Hx'= 0. If the received codeword does not +satisfy this parity check, then the decoder computes the parity checks +on all of the bits. The bit(s) associated with the most failed parity +checks are flipped. The process repeats until a valid codeword is +found, or a maximum number of iterations is reached, whichever comes +first. + +The gr::fec::ldpc_decoder is a soft-decision decoder that uses belief +propagation (also known as message passing). Designed for a memoryless +AWGN channel, it assumes a noise variance entered in as 'Sigma' in the +block. This is a suboptimal, yet efficient method of decoding LDPC +codes. + \subsection fec_parallelism Parallelism @@ -324,6 +399,7 @@ parallelism discussed above with the <b>None</b>, <b>Ordinary</b>, and <b>Capillary</b> models of threading. + \section fec_api The API of the FECAPI The FECAPI defined by the parent generic_encoder and generic_decoder |