This paper proposes an FPGA based fuzzy computation accelerator (FCA) for fixed point microcontroller implementation of fuzzy systems. It supports optimizable and hardware implementable fuzzy membership functions proposed previously. The algorithm for low-end hardware implementation has been tested using Intel 8/spl times/C196KC fixed-point microcontroller [2]. The proposed FCA improves the processing speed by performing complex time-consuming operations. The hardware and software features of the FCA are presented. The percentage improvement in speed for a typical fuzzy structure is computed and reported. The advantages, versatility and limitations of the proposed FCA are summarized The FCA supports fuzzy systems whose membership functions are negative powers of 2, irrespective of the structure of the system. This would greatly aid medium range applications that require high computational power at low cost.
[1]
Alessandro Gabrielli,et al.
A fast digital fuzzy processor
,
1999,
IEEE Micro.
[2]
Donald L. Hung,et al.
Dedicated digital fuzzy hardware
,
1995,
IEEE Micro.
[3]
Valentina Salapura,et al.
A fuzzy RISC processor
,
2000,
IEEE Trans. Fuzzy Syst..
[4]
S. Himavathi,et al.
Implementation of Nonlinear Fuzzy Models Using Microcontrollers
,
2002,
AFSS.
[5]
S. Himavathi,et al.
New membership functions for effective design and implementation of fuzzy systems
,
2001,
IEEE Trans. Syst. Man Cybern. Part A.
[6]
Alessandro De Gloria,et al.
Fuzzy logic microcontroller
,
1997,
IEEE Micro.
[7]
Raoul Velazco,et al.
Digital fuzzy control: a robust alternative suitable for space application
,
1998
.