Implementation of logic circuits using a novel design of a reconfigurable feedforward artificial neural network approach

No doubt the digital design and implementation of any logic circuit suffer much from the difficulty. On the other hand, the analog design and realization of analog devices have advantages of both economy and easy to implement compared with the digital design. In the present work, we adopted the reconfigurable artificial neural network approach for the of the 7 traditional logic functions in both the bit- and word-level. A new design for XOR and XNOR logic functions have been proposed using this approach. Comparison between the proposed and traditional networks shows that, the proposed network has a minimum number of neurons and synaptic weights. The proposed neural network for bit-level has been realized by a hardware means and the result have been checked using Pspice computer program. Both the actual and computer results are found to be very close to the correct results.