降低单通道盲分离PSP算法复杂度的方法

A Method for Reducing the Complexity of Per-Survivor Processing Algorithm on Single Channel Blind Separation

随着电磁环境的日益恶化,越来越多的通信系统面临着多信号混合干扰的问题,这对解调信号造成了一定的困难。解决这个问题使用PSP算法对混合信号进行单通道盲分离是一种有效的手段。本文针对目前PSP算法在单通道盲分离中随着调制阶数上升所带来的算法复杂度提高的问题,利用PSP算法的复杂度随串扰长度L增加而上升的特点,提出了一种改进型的PSP盲分离算法。仿真结果表明,在对误码率影响不大的基础上此种改进型算法跟原算法相比,在BPSK解调中能够加速大约3倍,QPSK解调中能够加速大约18倍。若采取另一种不降低串扰长度L的改进方式,则BPSK解调与原算法相比最高能有2 d B的增益,并且即使降低L仍有1.5 d B的增益。QPSK解调与原算法相比最高能有1.5 d B的增益,降低L仍有1 d B的增益。

With the worsening electromagnetic environment, more and more communication system is facing the mixed sig- nal interference problem which cause some difficulties to demodulation. PSP （ Per-Survivor Processing） algorithm is an ef- fective solution for single channel mixed signal blind separation. To solve the current problem that the complexity of PSP algorithm will increase with the higher modulation order in single channel blind separation, an improved PSP blind separa- tion algorithm was proposed based on the characteristic that the complexity of PSP algorithm increases with the rise of the length of ISI. Simulation results showed that, comparing with the original algorithm, this improved algorithm can speed up approximately three times in the BPSK demodulation and about 18 times in QPSK demodulation without much impact on Bit Error Rate. If we took another method using the unchanged length of ISI, the BPSK demodulation only had 2 dB gain at most compared with the original algorithm. Even if we reduce the duration of ISI, the result still has 1.5 dB gain. While the QPSK demodulation had 1.5 dB gain at most. Even if we reduce the duration of ISI, the result still has 1 dB gain.