查詢結果分析
來源資料
頁籤選單縮合
| 題 名 | A Progressive Scheme for FIR System Identification with Short-time Hartley Transform=使用短時哈氏轉換之遞進式有限脈衝響應系統之辨證 |
|---|---|
| 作 者 | 姜皇成; 劉皆成; | 書刊名 | 大同學報 |
| 卷 期 | 25 1995.11[民84.11] |
| 頁 次 | 頁231-237+439 |
| 分類號 | 448 |
| 關鍵詞 | 有限脈衝響應系統; 快速哈氏演算法; |
| 語 文 | 英文(English) |
| 中文摘要 | 現今的文獻中已經有很多方法針對系統辨證及模型化做討論,其中最成功者要算 是時域上之最小平方法。然而當系統之階數很大時,此法常涉及複雜之大型矩陣運算,因此 減低其實際的應用價值。在本文中,針對有限脈衝響應系統,我們應用短時哈氏分析技術於 頻率域上進行系統辨証。即使資料量龐大,在不失準確度的前提下,因為使用了快速哈氏演 算法所以大幅提昇了運算速度。此外,傳統頻率域上之辨證法所無法消除之視窗效應在哈氏 分析技術下也能去除。最後,若不斷有資料輸入系統中,我們也提出了一遞進式演算法,因 應新的資料隨時更新辯證的結果。 |
| 英文摘要 | A wide variety of methods have been proposed for system modeling and identification. The most successful of these methods have been the time domain procedures such as least square analysis. This method, however, involves complicated inverse operation of large scale matrix when the order of the system is very large; thereby reducing the potential applications of this class of techniques. In this paper, we proceed the identification procedure in the transform domain by using the short-time Hartley transform analysis technique. When the amount of data is very large, the fast Hartley transform (FHT) algorithm is utilized and its size can be made as small as possible, hence increasing the speed of the algorithm without sacrificing the accuracy. Besides, the window effect companied with the classical transform domain methods can be removed by linearly combing biased estimates, leading to an unbiased estimate. Finally, we also obtain a progressive scheme allowing the estimate to be recusively updated as new data are available. The method assumes that the system is a finite impulse response system. |
本系統中英文摘要資訊取自各篇刊載內容。