Title :
Implementation of pipelined LMS adaptive filter for low-power VLSI applications
Author :
Dukel, Bulent ; Rizkalla, Maher E. ; Salama, Paul
Author_Institution :
Dept. of Electr. & Comput. Eng., Indiana Univ., Indianapolis, IN, USA
Abstract :
In this paper, we described the implementation of a pipelined low-power 6 taps adaptive filter, based on the least-mean square (LMS) algorithm. The process of the power characterization procedure is very efficient and can be easily set in synthesis based design flows without too much additional effort. The architecture shows a novel tradeoff between algorithmic performance and power dissipation. In this work, we also characterized the power with a natural top-down design methodology with iterative improvement.
Keywords :
adaptive filters; circuit optimisation; circuit simulation; integrated circuit design; least mean squares methods; logic design; logic simulation; low-power electronics; parallel processing; pipeline processing; algorithmic performance/power dissipation tradeoff; least-mean square algorithm; low power optimization; low-power VLSI; parallel processing; pipelined LMS adaptive filter; pipelining; power characterization; power consumption; power minimization; relaxed look-ahead; Adaptive filters; Application software; Capacitance; Circuits; Clocks; Energy consumption; Least squares approximation; Pipeline processing; Very large scale integration; Voltage;
Conference_Titel :
Circuits and Systems, 2002. MWSCAS-2002. The 2002 45th Midwest Symposium on
Print_ISBN :
0-7803-7523-8
DOI :
10.1109/MWSCAS.2002.1186916
