中文摘要：

在非合作接收系统信号识别中，针对通信信号的调制类型识别时信息同步性差的问题，提出了一种改进的不需要先验知识的调制识别算法。由于现有算法主要针对复基带信号进行调制类型识别，识别系统需要严格的载波同步和码元同步等预处理过程．处理过程繁琐，需要先验知识较多。为简化识别流程，提高识别效率，在对信号模型深入分析的基础上，提出采用核密度估计结合曲线特征的识别算法，直接在中频阶段完成对信号的调制类型识别。仿真结果表明，提出的算法不需要先验知识，避免了复杂的系统预处理过程，优化了识别流程，且具有良好的识别效果。

英文摘要：

For the problem of the communication signals modulation recognition in non-cooperative receiving, an improved and don＇ t need prior knowledge modulation recognition algorithm is proposed. The existing algorithm is aimed at baseband signal, need complex and verbose carrier synchronization and symbol synchronization process. To simplify the recognition process, improve the recognition efficiency, we firstly analyze the signal model, then use the Kernel Density Estimation and extract curve feature algorithm to identify the signals, complete identification of signal modulation type in Intermediate frequency signal： Experiments show this method doesn＇ t require prior knowledge, optimize the recognition process, avoiding the signal pretreatment process, and has good recognition effect.