通过分析典型卫星通信信号功率谱、平方谱、四次方谱以及包络平方谱特性,针对卫星通信随参信道造成信号噪声谱均值与方差随频率变化这一问题,采用最小二乘法对信号谱图进行预处理,定义偏离比作为单频分量检测值,在无任何先验知识条件下提取一组鲁棒性强的特征参数,实现信号自动调制识别。实验证明,改进后的算法复杂度较低、过程简单易实现,所提取参数对数字信号的调制指数与滚降系数不敏感,具有较好的可分性与稳健性,并同样适用于随参信道的信号调制识别,具有一定的实用意义。
By analyzing the power spectrum,power spectrum of the second and forth spectrum of the common used modulated signals,square with the influence of satellite communication channel,the mean and variance of noise spectrum vary with frequency. By preprocessing the signal spectra with the method of the least square,defined deviation than as a single frequency component values,under the condition of without any prior knowledge. Simulation results show that the method has high practicability,owing to being more resolvable and steady which is not sensitive to the roll-off factor and modulation index of the digital signals. The improved method can also achieve recognition accuracy in the adaptive channel and has certain practical significance.