Channel Simulation and Tracking Performance Comparison of LMS, RLS and Kalman Filter

我們想要在最佳情況下比較最小均方法(LMS)、遞迴最小平方法(RLS)和卡爾曼濾波(Kalman Filter)的追蹤效能。但是我們很難從真實的資料得到正確的通道脈衝響應特性和訊雜比(SNR)。因此，我們使用模擬的資料以確保我們可以在最佳情況下追蹤頻道。我們也希望模擬的資料可以符合真實的情況。首先，我們從真實的資料拿到模擬參數。然後我們使用這些參數和頻道模型方法模擬接收到的訊號。因此，在各種頻道脈衝響應變化起伏下，我們比較三種追蹤方法LMS、RLS和Kalman Filter的效能。結果顯示，在不同頻道脈衝響應變化起伏下，Kalman Filter的效能比LMS 和RLS 還要好。且在最佳情況，RLS效能又比LMS好。We want to compare the performances of LMS, RLS and Kalman filter on the optimum condition. But we may not get the proper characteristics of channel impulse response (CIR) and the information of SNR from real data. Hence, we use simulation data to guarantee that we track channel on the optimum condition. We also expect that the simulation data is similar to the real condition. First, we get the simulator parameters from the real data. Then we use the parameters and channel model method to simulate the received signal. Therefore we can compare the performances of three tracking algorithms, LMS, RLS, Kalman filter on the condition of different fluctuations of CIR. It shows that the performance of Kalman filter is much better than LMS and RLS in sundry fluctuations of CIR. And the performance of RLS is better than LMS on the optimum condition