非线性失真场景下基于ELM帧同步改进方法

无线通信系统中存在大量的非线性设备或器件,不可避免地造成非线性失真,引起一些经典与时新帧同步方法的帧同步性能严重恶化。为改善帧同步性能,文章开发帧间相关性,提出一种基于极限学习机(ELM,extreme learning machine)的帧同步改进方法。首先,根据帧间相关性先验信息,提出方法利用加权叠加进行预处理捕获帧同步度量的初步特征;然后,基于预处理后的初步特征,构建ELM网络估计帧同步偏移。在线运行结果表明,相比于经典与时新的帧同步方法,文章方法可改善帧同步错误概率性能;针对不同的参数影响,帧同步错误概率性能的改善具有鲁棒性。

In the wireless communication systems,a large number of nonlinear devices or components inevitably cause nonlinear distortion.For some classical and novel frame synchronization(FS)methods,their performance of FS is seriously degraded due to the nonlinear distortion.In order to improve the FS performance,an extreme learning machine(ELM)-based improved method of FS is proposed in this paper by exploiting the inter-frame correlations.With the exploited inter-frame correlations,a preprocessing is first performed to capture the coarse features of synchronization metric by weighting and superimposing multi-frame signals.Then,an ELM network,which is based on the coarse features of preprocessing,is constructed to estimate the FS offset.Compared with the classical and novel FS methods,the results of online running indicate that our approach could improve the error probability of FS,and this improvements are robust against the impacts of different parameters.