A study of a radix-2 Signed-Digital fuzzy processor using the logic oriented neural networks

A novel Signed-Digital fuzzy processor using the logic oriented neural networks is proposed. Since the signed-digit number system has a redundant property to represent the binary numbers, the high speed adder in the processor can be realized in the signed-digit system without a delay of the carry propagation. In this paper the novel circuits of the main two systems of inference engine and defuzzifier in the fuzzy processor are constructed with the logic oriented neural networks. The advantages of the logic oriented neural networks are the simple construction in high speed operation. Emphasis is placed on the logic oriented neural networks to perform the function of defuzzifier based on the novel algorithm in the signed-digit number system. In order to investigate how the fuzzy processor in the signed-digit system achieve its function, computer simulations for the key circuits of the processor are made. The results show that the proposed circuits can perform the operations in higher speed than those of common fuzzy processor.