Convolutional Codes#

ConvEncoder([gen_poly, rate, ...])

Encodes an information binary tensor to a convolutional codeword.

ViterbiDecoder(*[, encoder, gen_poly, rate, ...])

Applies Viterbi decoding to a sequence of noisy codeword bits.

BCJRDecoder([encoder, gen_poly, rate, ...])

Applies BCJR decoding to a sequence of noisy codeword bits.

This module supports encoding of convolutional codes and provides layers for Viterbi [Viterbi] and BCJR [BCJR] decoding.

While the ViterbiDecoder decoding algorithm produces maximum likelihood sequence estimates, the BCJRDecoder produces the maximum a posterior (MAP) bit-estimates.

The following code snippet shows how to set up a rate-1/2, constraint-length-3 ConvEncoder in two alternate ways and a corresponding ViterbiDecoder or BCJRDecoder. You can find further examples in the Channel Coding Tutorial Notebook.

Setting-up:

encoder = ConvEncoder(rate=1/2, # rate of the desired code
                      constraint_length=3) # constraint length of the code
# or
encoder = ConvEncoder(gen_poly=['101', '111']) # or polynomial can be used as input directly

# --- Viterbi decoding ---
decoder = ViterbiDecoder(gen_poly=encoder.gen_poly) # polynomial used in encoder
# or just reference to the encoder
decoder = ViterbiDecoder(encoder=encoder) # the code parameters are infered from the encoder

# --- or BCJR decoding ---
decoder = BCJRDecoder(gen_poly=encoder.gen_poly, algorithm="map") # polynomial used in encoder

# or just reference to the encoder
decoder = BCJRDecoder(encoder=encoder, algorithm="map") # the code parameters are infered from the encoder

Running the encoder / decoder:

# --- encoder ---
# u contains the information bits to be encoded and has shape [...,k].
# c contains the convolutional encoded codewords and has shape [...,n].
c = encoder(u)

# --- decoder ---
# y contains the de-mapped received codeword from channel and has shape [...,n].
# u_hat contains the estimated information bits and has shape [...,k].
u_hat = decoder(y)

This Page