模糊系統應用於信號處理的可實行性探討

Abstract

本文主要在探討模糊系統在信號處理上的可實行性。藉由雜訊消除，通道 等化，影像編碼及影像前級處理等應用來評估模糊系統在信號處理上的可 實行性。目的在於驗證模糊系統的能力與效率，並以最佳化理論的角度來 整合一些已存在的模糊方法。實驗結果顯示，模糊系統在信號處理的使用 上比一般的方法昂貴而且在本論文所探討的應用上，模糊系統並非全比一 般的方法傑出。但是在一些非線性強和複雜度高的問題上，模糊系統有很 好的表現。 This thesis investigates the feasibility of fuzzy system approaches to signal processing. Noise cancellation, adaptive equalization, image coding, and image preprocessing are sample problems used to evaluate the feasibility. The focus of the present exercise is on justifying the capabilities and effetiveness of fuzzy system approaches, and unifying already existing fuzzy system approaches in the problem of optimality. Experimental results demonstrate that fuzzy system approaches are more expensive than conventional approaches and they may not outperform conventional approaches in all of our studies. We illustrate that fuzzy system approaches are effective for problems that are highly nonlinear and complex.