Study of stopping criteria in LDPC coded iterative MIMO OFDM receiver

In this paper, we address the problem of scheduling a LDPC coded iterative MIMO OFDM receiver. Iterative reception is an efficient way to approach optimal performance avoiding exhaustive algorithms. In such receivers using LDPC codes, complexity and delay could be critical since decoding and equalization tasks are repeated at each equalization loop. A LDPC decoder stops once it reaches a maximum number of iterations or fulfills a preset stopping criterion. In this paper we propose and evaluate the use of different stopping criteria based on reliability metrics to reduce the complexity and delay of the receiver by minimizing the total number of iterations while preserving the same performance. We show that, at low signal to noise ratio, the use of reliability based stopping criteria reduces complexity and delay by an early reuse of MIMO equalization. Also, stopping criteria can be chosen dynamically to match with channel conditions and the corresponding outer loop step.