Font Size: a A A

Constrained Least Square Design Of FIR Digital Filter With Reduce Group Delay

Posted on:2012-08-17Degree:MasterType:Thesis
Country:ChinaCandidate:B YuanFull Text:PDF
GTID:2178330335962643Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of science and technology, digital signal processing technology is widely applied in many kinds of fields. Digital filtering technology plays a very important role in the digital signal processing technology. Because the finite impulse response (FIR) digital filters can be inherent stability and exactly linear-phase, the design of FIR digital filters have attracted extensive research efforts. For exactly linear-phase FIR digital filters, the group delay is about 1?2 of its length. When the transition band is narrow, the group delay will be large. However, in many areas of practical applications, such as microwave and radar communications, optical communications, medical imaging, real-time video processing and so on, which require the constant group delay of digital filters as low as possible to avoid the large signal delay, and to minimize the distortion of the transmitted signal. Therefore, the research on FIR digital filters with reduce group delay has very important practical significance.This thesis considers the design of digital filters, mainly concentrates on the techniques for nonlinear-phase FIR digital filter with reduced group delay. Firstly, the concept and mathematical model of FIR digital filter are disserted. The basic theory of FIR digital filter and optimization is widely disserted. The methods that design FIR digital filter are researched. The reason to group-delay error that is large near the bandedge and the approaches of reducing the group-delay error are analyzed. Based on the above theoretical analysis, the following tasks are completed:1. The constrained least square design of nonlinear-phase finite impulse response digital filter with reduce group delay is considered. By minimizing the complex error energy and imposing the constraints on the complex approximation error and the phase error, the convex feasible region can be obtained, and the magnitude error and phase error can be independently controlled. The method in [16] for low-pass filter with given phase error, in which the right-sided sigmoid upper-function is introduced to constrain the phase error, is extended to the design of high-pass and band-pass filters. The left-sided and double-sided sigmoid upper-bound functions are introduced to constrain the phase error of high-pass and band-pass filters in order to reduce the group-delay error near the bandedges.2. Since the sigmoid function has tow parameters to be determined by the user, the manual selection of these parameter values makes it not easy to use in the design of FIR digital filters with small group-delay error. To solve this problem, this thesis presents an iterative scheme to adaptively update the phase-error upper-bound function, which uses the envelop of the group-delay error obtained from an old design to construct a new phase-error upper-bound function, such that the resulting filter has near-uniform group-delay error, and thus to reduce the maximum group-delay.The above two methods of reducing group-delay error are realized by Matlab and a large number of design examples have been simulated in this thesis. Design examples demonstrate that the proposed methods are effective and practical to the reduction of group-delay error of the designed digital filter.
Keywords/Search Tags:FIR digital filter, constrained least square design, group-delay error, phase error
PDF Full Text Request
Related items