This technology is a generalized shift-map code decoding apparatus and method that restores original source signals by having multiple sub-decoders process N input values received via a channel, iteratively estimating the quotient of each input value by exchanging reliability information with neighboring decoders, and utilizing a quotient determination unit and a source estimate calculation unit.
Shift-map codes have gained attention as analog error-correction codes due to their simple encoding process—requiring only multiplication and modulo operations—and their graceful degradation in signal-to-noise ratio (SNR). However, decoding has historically been significantly more complex than encoding. In particular, the lack of efficient decoding apparatuses or methods for generalized shift-map codes, which offer flexible error trade-offs between low and high SNR, has limited their practical application.
This technology incorporates a reliability calculation unit, a reliability update unit, a maximum value determination unit, and a reliability selection unit into each sub-decoder to iteratively update reliability using a belief propagation method, finalizing the quotient once reliability converges. Subsequently, individual source estimates calculated by multiple estimators are synthesized to produce a final estimate. Applicable to wireless sensor networks, low-latency video/audio transmission, and multi-terminal communication receivers using analog joint source-channel coding, this method significantly reduces computational complexity compared to maximum likelihood decoding while precisely restoring source signals in noisy environments.
This invention was developed with support from the Ministry of Science and ICT for the development of QKD technology standards in the field of quantum cryptographic communication.
N/A
