Design of a reconfigurable state transition algorithm for fuzzy automata

Abstract

This paper presents a reconfigurable state transition architecture for the Hybrid Fuzzy-Boolean Finite State Machine (HFB FSM). Building modules with parameters are used to provide flexibility for the architecture in terms of the number of fuzzy inputs and the number of states. Flexible architecture is deemed as a very attractive property to implement virtual fuzzy automata for supervisory controllers of complex systems. Relevant clusters of states with respect to fault detection, identification and recovery vary a great degree along the goal path. The reconfigurable architecture allows a supervisory controller to reset the fuzzy automaton to model a particular state cluster of interest. The paper first outlines the problem of fault detection, identification and recovery from faults in the context of ontological control. Next the key properties of a reconfigurable fuzzy automation model are given along with the concepts to address the issues of fault detection, identification and recovery. Then the various aspects of implementing the reconfigurable fuzzy automation using programmable logic devices (FPGAs) are discussed in detail. The hardware was designed in VHDL and the simulations were done using ModelSim.

Keywords:

• fuzzy automata
• reconfigurable hardware implementation of fuzzy automata
• fuzzy logic hardware accelerator
• intelligent decision support system