Abstract:

To deal with the huge burden to the function node (FN) in each decoding iteration in sparse code multiple access (SCMA) by using message passing algorithm (MPA), this paper first derives some statistical observation for the probability density function of received signals at various signal-to-noise ratios (SNRs) by using the Monte Carlo method. Then based on a detailed analysis between the probability density function value of received signals and external information values, a threshold decision scheme is introduced into MPA, therefore a partial extrinsic information transmission of MPA (PEIT-MPA) is proposed. Simulation results show that the proposed PEIT-MPA efficiently relieves the implementation complexity of MPA, especially at the case of high SNR, meanwhile keeps the bit error ratio and convergence rate from degrading noticeably.