Transform Domain LMS/F Algorithms, Performance Analysis and Applications

Transform Domain LMS/F Algorithms, Performance Analysis and Applications. Masters thesis, King Fahd University of Petroleum and Minerals.


Download (3MB) | Preview

Arabic Abstract

من المعروف أن فلاتر التكيف الذاتي تعاني من بطء التقارب في حالة الدخل شديد الالتفاف. في الجزء الاول من الاطروحة يتم تقصي تقنية تحويل المجال لحل مشكلة بطء التقارب في حلبة استخدام خوارزميات الدالة امن الدرجة الرابعة. من المعروف ان هذه الخورزميات توفر أداء متميز في حالة الضجيج غير الغاوسياني مقارنة مع الخوارزميات المشتقة من الدالة ذات الدرجة الثانية للخطأ. في الجزء الثاني يتم دراسة الخوارزميات من الدرجة الرابعة تحت النظم المتناثرة ذات الاعداد القليلة من المفاتيح في دالة الاستجابة. تم تقديم خوارزمية و التي تحل مشكلة النظم المتناثرة في حلبة الضجيج غير الغاوسياني. وتم تقديم خليط من الفلاتر لحل المشكلة في حالة النظم المتناثرة المترددة.

English Abstract

Adaptive Filtering Algorithms are known to suffer in terms of convergence under highly correlated inputs. In this work we investigate the Transform Domain LMF algorithm. The Fourth Moments family of adaptive algorithms is known to have better steady state and convergence properties compared to the LMS family, under Non-Gaussian noise environments. The second scope of the thesis is to study the problem of sparse system identification, that is, when the number of active elements in the linear system under consideration is small. We proposed the Zero-Attractor (ZA-TD-LMF) algorithm to solve the problem under the highly correlated environments and non-Gaussian noise. Furthermore to deal the prob- lem of variable sparsity, we employed the convex combination filter to design an algorithms that can solve the problem of moving sparse elements. The first convex filter, called ZA-LMF and LMF convex, is designed to alleviate the choice of which algorithm we should use (LMF or ZA-LMF) if we do not have previous information about the level of sparsity of the system under consideration. Similarly, a transform Domain LMS and ZA-LMS convex is proposed and studied to endow the designer with more freedom in system designing. The Computer Simulations confirm that the convex is universal and always offer the best performance available and chooses the best performing algorithm.

Item Type: Thesis (Masters)
Subjects: Electrical
Department: College of Engineering and Physics > Electrical Engineering
Committee Advisor: Zerguine, Azzedine
Committee Members: Alnaffouri, Tareq and Abu-Alsaud, Wajih
Depositing User: MURWAN MOHAMED ELMAHDI BASHIR (g201304570)
Date Deposited: 22 May 2016 10:26
Last Modified: 01 Nov 2019 16:33