Conference Paper

The strict latency and reliability requirements of
URLLC are particularly difficult to comply with, due to the
inherent trade-off between these two factors. Most applications
falling under the URLLC umbrella rely on the exchange of
small packets. Convolutional codes (CC) exhibit near-optimal
performance when encoding small information blocks. In order
to operate in packet based transmission, it is necessary to
introduce some kind of termination to the CC. Tail-biting CCs
(TBCC) avoid rate loss and maintains equal error protection
for all information bits. However, its ML decoding algorithm
requires a Viterbi algorithm (VA) execution for each possible
initial trellis states, which rapidly becomes unpractical. This
becomes increasingly problematic in a multi-user scenario, where
several URLLC users need fast decoding at the same time. This
work proposes a novel distributed multi-user TBCC (MU-TBCC)
coding scheme, which completely eliminates the ZTCC rate-loss
and reduces decoding complexity for any multi-user scenario. In
fact, the decoding algorithm complexity converges to the one of
the VA as the number of users increase. The results shows that,
as the SNR increases, the error performance of the MU-TBCC
approaches the one of orthogonal transmissions. This scheme can
be also interpreted as non-orthogonal multiple access scheme,
whose structure enables efficient joint detection and decoding.