一种最大似然调制识别的快速算法

针对最大似然调制识别算法计算复杂度高的问题，提出了一种可用于实时软件接收机中的离散最大似然算法。通过预存离散似然函数值而后直接查表调用的方式解决耗时的似然函数计算问题，并且算法对载波频率偏差和相位偏移具有鲁棒性。仿真结果表明，该算法与最优最大似然调制识别算法相比，能有效地简化运算复杂度而性能损失较小。

Fast algorithm for maximum likelihood modulation classification

Aiming at the high computational complexity of the maximum likelihood modulation classification method, a discrete maximum likelihood algorithm which can be used in real-time software-defined receivers is proposed. By use of the pre-stored discrete likelihood function value, one can directly look-up a table and pick up the results to eliminate the time-consuming computation of the likelihood function. And this algorithm is robustness to carrier frequency offset and phase offset. Simulation results show that the algorithm can efficiently reduce the computation complexity at a very little sacrifice of performance compared with the optimum maximum likelihood modulation classification method.