- 作者: 黎自奮
- 作者服務機構: 國立中興大學應用數學系
- 中文摘要: 在一個多重疊信號中,本論文估計各個信號的到達時問(Delay time)及它的振幅(Amplitude)。以前多用過濾器(Filter)的技術來估計。最近多用最大可能法(ML)、估計最大法(Estimate Maximize)及最小平方法(LeastSquares)來估計。 本論文用淘汰法(Eliminated Method)來尋找各個信號到達時間,再用最小平方法估它的振幅。本論文的構想是先假定在每取樣時間(Time Simple)都有一個單獨信號,再用淘汰法將錯誤信號淘汰。最後沒有被淘汰者才是真正信號到達時間。用電腦模擬二種不同多重疊信號,所有各個信號本文的淘汰法都準確地找到。
- 英文摘要: Consider a multipath signal whose individual signals are deterministic known pulses. Our problemis to estimate the amplitudes and delay times of individual signals. The traditional methods used to findthe arrival times and the amplitudes were matched-filter and inverse-filter techniques. Later, maximumlikelihood (ML) estimation and the Estimate Maximize algorithm were commonly used. Recently, a leastsquares algorithm with complex-valued amplitudes has been used to improve estimation. In this paper, our main idea is to consider the multipath signal as a multiple regression model, whereeach time sample is considered to be a possible arrival time of a pulse which may not be real. We usean elimination method to eliminate false pulses. Using this method, the number of pulses in a multipathsignal does not need to be known. The results of a Monte Carlo simulation study with sinusoidal andtrapezoidal pulses are presented to demonstrate the accurate estimation of arrival times and amplitudesobtained by this method.
- 中文關鍵字: estimate maximize algorithm; maximum likeihood estimation; multiple regression mode; multipath signal
- 英文關鍵字: --