Skip to Main Content
Currently, combining turbo or low-density parity-check (LDPC) codes with bit-interleaved coded modulation (BICM) is the most common coding scheme for bandwidth-limited channels. However, in order to be capacity achieving, shaping is necessary, which is difficult in conjunction with iterative processing. An alternative is interleave-division multiplexing (IDM), which can be either treated as a coded modulation scheme or a multiplexing scheme. In IDM, coded data sequences are linearly superimposed, thus avoiding the necessity of active signal shaping. The characteristics of IDM can be controlled by power and phase allocation. With typical parameter settings, IDM is non- bijective. In that case, the main task of channel coding is aiding to resolve the data sequences ("layers") at the receiver side, whereas noise mitigation is a secondary task. In this paper, we demonstrate that classical channel codes (like LDPC and turbo codes) fail when applied with IDM. In fact, repetition coding is quite useful. It is shown by means of an EXIT chart analysis that a concatenation of a high-rate convolutional code with an irregular repetition code matches well with the characteristics of an IDM a posteriori probability (APP) demapper.