论文部分内容阅读
为解决信号处理中非因果自回归（ＡＲ）系统的自适应辨识问题，本文提出了一种利用倒谱进行ＡＲ系统辨识的新方法。这一方法在倒谱域内把非因果ＡＲ系统辨识问题转化为非最小相位有限冲激响应（ＦＩＲ）系统辨识问题，因此可以利用现有的基于倒谱的ＦＩＲ系统辨识算法构造非因果ＡＲ系统辨识算法，从而实现了非因果ＡＲ系统的自适应参数辨识，并解决了以往文献中未曾解决的非因果ＡＲ系统阶次自适应确定问题。另外，本算法可以保证估计出的ＡＲ模型是稳定的。数值仿真结果证明了本算法具有很高的估计精度。
In order to solve the problem of adaptive identification of non-causal autoregressive (AR) systems in signal processing, a new method to identify AR systems using cepstrum is proposed in this paper. This method transforms the non-causal AR system identification problem into the non-minimum phase finite impulse response (FIR) system identification problem in the cepstral domain, so the existing cepstral FIR system identification algorithm can be used to construct the non-causal AR system identification Algorithm to realize the adaptive parameter identification of non-causal AR system and solve the unscheduled adaptive non-causal AR system problem that has not been solved in the prior literature. In addition, the algorithm can ensure that the estimated AR model is stable. Numerical simulation results show that this algorithm has high estimation accuracy.