支持向量回归盲检测16PSK信号

Blind Detection of 16PSK Signal Using Support Vector Regression

鉴于目前信号盲检测算法主要依赖二阶或高阶统计量,均需要大数据量作为基础。而基于小数据量的高阶PSK信号盲检测算法迄今未见。研究了仅依赖小数据量的盲检测16PSK信号算法,采用支持向量回归方法并采用线性ε-不敏感损失函数将16PSK信号盲检测问题转化为无约束优化问题,并采用迭代算法求解该优化问题,进而盲检测16PSK信号。算法验证中采用经典移动通信复数信道进行仿真和分析该方法的性能。

Most algorithms of blind detection are based on second-order or high-order statistics, all of which need a large data block. Blind detection of high-order PSK（phase shift keying） signal based on short data block hasn＇ t been presented according to references we found. This paper focuses on studying the algorithm of 16PSK signal blind detection depends on a short data block. The problem of 16PSK signal blind detection is transformed to solve unconstrained optimization problem by using support vector regression and adopting linear ε-insensitive loss function. Moreover, an iterative method is used to solve the optimization problem of blind detection of 16PSK signal. Finally, a classic mobile communications complex channel is adopted to simulate and analyze the performance of this novel algorithm and its characteristics.