This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Computationally Efficient Active Rule Detection Method: Algorithm and Architecture Mahdi Hamzeh a , Hamid Reza Mahdiani a , b , * , Ahmad Saghafi a , Sied Mehdi Fakhraie a , Caro Lucas c , a Silicon Intelligence and VLSI Signal Processing Lab., School of Electrical and Computer Engineering, University of Tehran, IRAN b Computer and Electronics Department, Sh. Abbaspour University of Technology, IRAN c Center of Excellence for Control and Intelligent Processing, University of Tehran, and School of Cognitive Science, IPM, IRAN Abstract In this paper, a new active rule detection algorithm is proposed which is efficiently implemented in dedicated fuzzy processors. Here, its advantages are analytically attested. A novel realization architecture is proposed that has higher performance and uses lower hardware resources in comparison to the other reported architectures. The structure of the proposed active rule detection unit is scalable in terms of the number of inputs, the number of membership functions, and their bit widths. The proposed architecture is flexible in term of membership function shape as well. Key words: Active rule detection, algorithm, architecture, fuzzy processor, scalable. * Corresponding author. Silicon Intelligence and VLSI Signal Processing Labora- tory of School of Electrical and Computer Engineering, University of Tehran, North Kargar Ave., Tehran 14395-515, Iran. Tel.: +98-912-1907099; Fax: +98-21-88006064; email@example.com. Email addresses: firstname.lastname@example.org (Mahdi Hamzeh), email@example.com (Hamid Reza Mahdiani), firstname.lastname@example.org (Ahmad Saghafi), email@example.com (Sied Mehdi Fakhraie), firstname.lastname@example.org (Caro Lucas). Preprint submitted to Elsevier 22 April 2008 1 Introduction Fuzzy logic [1,2] is used in an increasing number of applications. These appli- cations include process control , decision making support systems  and signal processing . The time constraints of various fuzzy systems may differ considerably according to the demands of different applications. In washing- machine controllers and auto focus imaging devices, for instance, the required inference speeds are quite low, while in real-time applications, they are ex- tremely high. Among them, we can mention of the applications reported in [8–11,13,14]. There are some successful efforts to propose simpler algorithms to reduce the amount of computations in fuzzy processes  which result in higher speeds  and smaller hardware . However, there is still need for higher performances. Thus, the designers have to use either high speed special purpose fuzzy hardware [8,18] or very high performance general purpose pro- cessors [19–21] and digital signal processors [22,23] to provide the necessary computational power....
View Full Document
- Spring '11