Abstract: This paper discusses the Extended Binary Phase Shift Keying（EBPSK） modulation communication system by applying detection technology to achieve high spectral efficiency while reducing energy consumption. The SIF can transform phase modulation into amplitude changes, so the tiny difference of the received waveforms can be separated. The nonlinear detection with Support Vector Machine （SVM） technique is used in EBPSK communication system, due to the difficulties for achieving the optimal performance by Threshold Decision （TD）. The union selection methods are presented to get the vector with fewer dimensions for SVM detector. Also, proper kernel selecting with parameter optimization and complexity reduction are analyzed. Simulation results show that the SVM nonlinear detector can obtain the maximum SNR gain at 7 dB under low dimension feature vector and low sampling rate. Thus, EBPSK modulation system can not only enhance the performance of detection by using SVM technique, but also has the ability of anti-interference and robustness to sampling rate.

Key words: nonlinear detection, support vector machine, extended binary phase shift keying, green communication

摘要： 主要讨论扩展的二元相移键控（EBPSK）调制通信系统中的检测技术，以实现高频谱利用率的同时降低能耗。针对高效调制信号码元间差异性较小的特点，接收机利用特殊的冲击滤波器（SIF）将相位变化转化为寄生调幅，增加码元间的差异性，提高能量利用率。同时，引入了支持向量机（SVM）非线性检测方法，解决EBPSK系统中门限检测性能难以达到最优的问题，并给出了维数较少的统一特征向量选取方法。为了减少计算量，选择适用于EBPSK系统的SVM核函数并优化计算方法。仿真结果表明，SVM非线性检测在较少的特征向量维数和低采样率的条件下，可获得最多7 dB的信噪比增益。因而，在EBPSK调制通信系统中选择SVM方法不仅可以提升检测性能，而且抗干扰能力更强，对采样率的鲁棒性更好。

关键词: 非线性检测, 支持向量机, 扩展的二元相移键控, 绿色通信