A Low Power Dynamic Reconfigurable Processor using Logarithmic Number System for Software Radio Equalizers

This paper presents the design of a dynamic reconfigurable processor using an array of two dimensional logarithmic numbering system (LNS) processing elements and registers. By programming the processor, the array configures dynamically during operation and executes the required task with different structures in each phase. The proposed coarse grain array architecture is suitable for implementation of software radio baseband equalizers. Since redundant elements found in fine grained structures are reduced, the design consumes less chip area and power. Several different equalizer algorithms including the CMA for QPSK and BPSK signals, the finite interval CMA and the sliding window CMA equalizers are implemented and programmed on the proposed processor array and successive operation is shown. The architecture is designed in a 0.13 ¿m CMOS technology and simulated for extraction of chip specifications. Power consumption, operation speed, gate count and chip area of the design are reported.

[1]  Philip Schniter,et al.  Blind equalization using the constant modulus criterion: a review , 1998, Proc. IEEE.

[2]  A. Lee Swindlehurst,et al.  Fixed window constant modulus algorithms , 1999, 1999 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings. ICASSP99 (Cat. No.99CH36258).

[3]  A. Benveniste,et al.  Blind Equalizers , 1984, IEEE Trans. Commun..

[4]  Henry Hoffmann,et al.  The Raw Microprocessor: A Computational Fabric for Software Circuits and General-Purpose Programs , 2002, IEEE Micro.

[5]  E. Buracchini,et al.  The software radio concept , 2000, IEEE Commun. Mag..

[6]  Jürgen Becker,et al.  A dynamically reconfigurable system-on-a-chip architecture for future mobile digital signal processing , 2000, 2000 10th European Signal Processing Conference.

[7]  Bertil Svensson,et al.  Two-level Reconfigurable Architecture for High-Performance Signal Processing , 2004, ERSA.

[8]  Phillip A. Regalia A finite-interval constant modulus algorithm , 2002, 2002 IEEE International Conference on Acoustics, Speech, and Signal Processing.