Efficient Modulation Classification for Adaptive Wireless OFDM Systems in TDD Mode

This paper investigates a novel efficient automatic modulation classification (AMC) algorithm in wireless time division duplex (TDD) orthogonal frequency division multiplexing (OFDM) systems with adaptive modulation (AM). Starting from the well-known maximum-likelihood (ML) classification metric for digital quadrature amplitude modulation (QAM) schemes, partial channel reciprocity in TDD-based adaptive OFDM systems is exploited by applying the maximum-a-posteriori (MAP) principle. A heuristic approach to find the required occurrence probabilities of the selected modulation schemes in a time-variant radio channel is analyzed. In order to reduce the overall computational complexity, a simplified metric is presented that is suitable for the implementation in real-time operating systems. Numerical results in terms of classification error probabilities confirm the effectivity of the complexity reductions. Moreover, it is demonstrated that the influence of the performance degradation on the overall packet error ratios (PER) due to imperfect modulation classification is small in a typical wireless communication scenario.