Complexity reduction is one of the major issues in today’s digital system designfor many obvious reasons, e.g., reduction in area, reduced power consumption,and high throughput. Similarly, dynamically adaptable digital systems requireﬂexibility considerations in the design which imply reconﬁgurable systems, wherethe system is designed in such a way that it needs no hardware modiﬁcationsfor changing various system parameters. The thesis focuses on these aspects ofdesign and can be divided into four parts.
The ﬁrst part deals with complexity reduction for non-frequency selectivesystems, like diﬀerentiators and integrators. As the design of digital processingsystems have their own challenges when various systems are translated from theanalog to the digital domain. One such problem is that of high computationalcomplexity when the digital systems are intended to be designed for nearly fullcoverage of the Nyquist band, and thus having one or several narrow don’t-carebands. Such systems can be divided in three categories namely left-band systems,right-band systems and mid-band systems. In this thesis, both single-rate andmulti-rate approaches together with frequency-response masking techniques areused to handle the problem of complexity reduction in non-frequency selectiveﬁlters. Existing frequency response masking techniques are limited in a sensethat they target only frequency selective ﬁlters, and therefore are not applicabledirectly for non-frequency selective ﬁlters. However, the proposed approachesmake the use of frequency response masking technique feasible for the non-frequency ﬁlters as well.
The second part of the thesis addresses another issue of digital system designfrom the reconﬁgurability perspective, where provision of ﬂexibility in the designof digital systems at the algorithmic level is more beneﬁcial than at any otherlevel of abstraction. A linear programming (minimax) based technique forthe coeﬃcient decimation FIR (ﬁnite-length impulse response) ﬁlter design isproposed in this part of thesis. The coeﬃcient decimation design method ﬁndsuse in communication system designs in the context of dynamic spectrum accessand in channel adaptation for software deﬁned radio, where requirements can bemore appropriately fulﬁlled by a reconﬁgurable channelizer ﬁlter. The proposedtechnique provides more design margin compared to the existing method whichcan in turn can be traded oﬀ for complexity reduction, optimal use of guardbands, more attenuation, etc.
The third part of thesis is related to complexity reduction in frequencyselective ﬁlters. In context of frequency selective ﬁlters, conventional narrow-band and wide-band frequency response masking ﬁlters are focused, where variousoptimization based techniques are proposed for designs having a small number ofnon-zero ﬁlter coeﬃcients. The use of mixed integer linear programming (MILP)shows interesting results for low-complexity solutions in terms of sparse andnon-periodic subﬁlters.
Finally, the fourth part of the thesis deals with order estimation of digitaldiﬀerentiators. Integral degree and fractional degree digital diﬀerentiators areused in this thesis work as representative systems for the non-frequency selectiveﬁlters. The thesis contains a minimax criteria based curve-ﬁtting approach fororder estimation of linear-phase FIR digital diﬀerentiators of integral degree upto four.
Linköping: Linköping University Electronic Press, 2012. , 47 p.
2012-03-08, Visionen, Hus B, Campus Valla, Linköpings universitet, Linköping, 20:03 (English)
Coleman, Jeffery, Dr.
Håkan Johansson, HåkanGustafsson, Oscar, Dr.