Short Term Link Performance Modeling for ML Receivers with Mutual Information per Bit Metrics

Link performance abstraction with simplified models that accurately capture the decoded receiver performance under fading channels has several applications. Such models help a receiver estimate a channel and map them to a performance metric, which allows it to select and recommend a preferred mode of transmission. In addition, these models capture the short term performance characteristics of link layer which can be used in system level studies. Mutual information metrics based on bit channels have been shown to give good performance prediction . In this paper, this approach is generalized by considering conditional PDFs of log-likelihood ratios (LLRs) of these bit channels. A novel and key contribution of this paper is deriving mutual information per bit (MIB) metrics that capture the performance of non-linear receivers like maximum-likelihood (ML) receiver. These are expressed as simple parameterized functions of the channel matrix, which can be evaluated with low complexity in a system simulation. Numerical results are presented verifying the prediction accuracy of the proposed models. The corresponding methodology and the functions used are provided, which are applicable to WiMAX communication system evaluation.