中文摘要：

在对多进制正交幅度调制（M-QAM，M-ary Quadrature Amplitude Modulation）信号星座图的不同聚类识别算法进行比较与分析的基础上，并针对信号码元数目不足的情况，提出了一种利用星座图对称特性与减法聚类相结合的改进算法。算法首先从接收数据中估计出码元速率，并根据码元速率估计值对其匹配滤波及抽样判决来恢复出信号星座图，然后将星座图中关于原点对称的两个象限中的星座点横坐标值与纵坐标值都投影到横轴或纵轴上，并对投影点进行聚类，从而，实现怍QAM调制方式识别。仿真结果证明：该识别算法能正确有效地识别JI~-QAM信号调制类型，而且减少了计算量，易于实际应用。

英文摘要：

By comparing and analyzing different clustering algorithms for M-ary quadrature amplitude modulation （ M-OAM ） signals recognition, and for the problem of insufficient symbols, an improved algorithm based on symmetrical characteristics of constellation and subtractive clustering is proposed. Symbol rate is estimated first form the received data, and based on the estimation, and through matched filtering and sampling judgment the signal constellation is recovered; then abscissa and ordinate values of the constellation points in two quadrants which are symmetric about the origin are projected onto the axes; and finally, the M-QAM signals are identified through clustering the projected points. Simulation results show that the improved method could identify the M-QAM signals correctly and effectively while saving the computing time. So this algorithm is easy and convenient in practical application.