This letter is concerned with the implementation of iterative decoding algorithms in analog integrated circuits. We study the convergence speed and the throughput of analog decoders for low-density parity-check codes, and show that they depend on the code, the decoding algorithm, the signal-to-noise ratio, and the average time constant of the analog circuit interconnections. However, they are not a function of the variance of the time constants. The analysis presented here can be used for selecting suitable codes and decoding algorithms for analog decoding. Furthermore, it can be used to estimate the throughput of an analog decoder, if the average time constant of the analog circuit is known.

Additional Metadata
Keywords Analog iterative decoder, Belief propagation (BP), Convergence speed, Low-density parity-check (LDPC) codes, Min-sum (MS) decoding, Throughput, Turbo codes
Persistent URL dx.doi.org/10.1109/TCOMM.2007.895978
Journal IEEE Transactions on Communications
Citation
Hemati, S. (Saied), & Banihashemi, A. (2007). Convergence speed and throughput of analog decoders. IEEE Transactions on Communications, 55(5), 833–836. doi:10.1109/TCOMM.2007.895978