在訊號辨識的領域裡，自動訊號辨識是一門古典的題目。這項技術較常應用在當傳送訊號具有可適性時的情況。針對通過非AWGN通道的訊號進行辨識的工作至今仍是一項困難的挑戰，高階統計法是最常被設計應用於針對此狀況的辨識技術。我們使用高階統計參數來估測通道係數，並使用累計量來設計一個多階層決策架構。我們將討論在靜態和時變通道模型下的演算法差異，並比較在不同接收條件下的辨識率。 Automatic modulation classification (AMC) is a classical topic in signal classification field. This technique is often used when the transmitted signals are adaptive. So far, recognition of signals passing through non-AWGN channels is still a hard task. High-order statistics is the most adopted method of being designed for classification in non-AWGN situations. We use high-order statistical parameters to obtain estimated channel coefficients and design a multiple-layered decision structure with cumulants. We will discuss the difference of algorithms for static and time-varying channel models, and compare the classification rate in different receiving conditions.
