中文摘要：

考虑到直接应用最大似然序列检测实现多调制指数连续相位（Multi-h CPM）解调需要大量的匹配滤波器和网格状态,难以用于工程实践,故对Multi-hCPM信号的通用PAM分解算法加以研究。分析各分解脉冲的幅度特征,提出合并同等幅度脉冲以减少匹配滤波器个数、忽略小脉冲以减少网格状态的低复杂度解调算法,并针对M=2和M=4两类典型的Multi-h CPM信号通过计算机仿真验证该解调方法的BER性能。结果显示：相比于最优解调,新算法信噪比损失分别小于0.5dB和2.5dB,且匹配滤波器个数和网格状态数可大大减小。更多还原

英文摘要：

Considering the Multi-h CPM demodulation adopted maximum likelihood sequence detection requires a large number of matched filters and trellis states,it is too complex to implement for practical application.Therefore,the universal decomposition of Multi-h CPM signals has been studied.By analyzing the amplitude character of each decomposed pulse,a low-complexity demodulation scheme of merging equal-amplitude pulses to reduce the number of matched filters and ignoring low amplitude pulses to reduce trellis states is presented.The BER performance is verified via computer simulation for the two typical Multi-h CPM signals of M=2 and M=4.The results of the research show that the SNR loss are lower than 0.5 dB and 2.5 dB respectively comparing with optimal demodulation.Moreover,the number of matched filters and trellis states are reduced greatly.