A Study on Likelihood Estimation Method Taking Account of Mutual Information in Multi-Level Symbol ~ A Proposal of Twin Turbo Decoder ~

Aiming at high spectrum efficiency, this paper proposes the twin turbo decoder. The received symbol carried by multi-level modulation or MIMO (multi-input multi-output) transmission consists of two or more bits, and these bits share mutual information. However, the conventional likelihood estimation neglects this mutual information and deteriorates the estimation accuracy. This paper presents an alternative estimation method taking into account the above mutual information and applies our proposal to an iterative MAP (maximum a posteriori probability) decoder in order to improve its error correction capability. Computer simulations verify that the proposed scheme improves the PER (packet error rate) performance with the conventional method, and the required E _SN₀ obtaining a PER of 1 % is improved by about 1.0 dB at maximum. Note that our proposed method is called the twin iterative decoder (twin turbo decoder) because it has double feedback loops