ارسل ملاحظاتك

ارسل ملاحظاتك لنا







A New Greedy Search Method For The Design Of Digital IIR Filter

المصدر: مجلة جامعة الملك سعود - علوم الحاسب والمعلومات
الناشر: جامعة الملك سعود
المؤلف الرئيسي: Kaur, Ranjit (Author)
مؤلفين آخرين: Patterh, Manjeet Singh (Co-Author) , Dhillon, J. S. (Co-Author)
المجلد/العدد: مج27, ع3
محكمة: نعم
الدولة: السعودية
التاريخ الميلادي: 2015
الصفحات: 278 - 287
DOI: 10.33948/0584-027-003-004
ISSN: 1319-1578
رقم MD: 973650
نوع المحتوى: بحوث ومقالات
اللغة: الإنجليزية
قواعد المعلومات: science
مواضيع:
كلمات المؤلف المفتاحية:
Digital Infinite Impulse Response (IIR) Filters | Binary Successive Approximation Based Evolutionary Search (BSA-ES) | Multi Objective Optimization | Lowest Order | Stability
رابط المحتوى:
صورة الغلاف QR قانون
حفظ في:
المستخلص: A new greedy search method is applied in this paper to design the optimal digital infinite impulse response (IIR) filter. The greedy search method is based on binary successive approximation (BSA) and evolutionary search (ES). The suggested greedy search method optimizes the magnitude response and the phase response simultaneously and also finds the lowest order of the filter. The order of the filter is controlled by a control gene whose value is also optimized along with the filter coefficients to obtain optimum order of designed IIR filter. The stability constraints of IIR filter are taken care of during the design procedure. To determine the trade-off relationship between conflicting objectives in the non-inferior domain, the weighting method is exploited. The proposed approach is effectively applied to solve the multiobjective optimization problems of designing the digital low-pass (LP), high-pass (HP), band pass (BP), and bands top (BS) filters. It has been demonstrated that this technique not only fulfills all types of filter performance requirements, but also the lowest order of the filter can be found. The computational experiments show that the proposed approach gives better digital IIR filters than the existing evolutionary algorithm (EA) based methods.

ISSN: 1319-1578