Towards Approximate Computing with Quantum-Dot Cellular Automata

In this paper, we inspect the approximate computing concept in quantum-dot cellular automata (QCA) as one of the most promising emerging nanoelectronic paradigms for replacing CMOS technology. Many applications can tolerate errors and imprecision of digital systems, so approximate computing is widely used in such cases. In this work, the fundamental QCA structure required for realization of an efficient approximate adder is proposed consisting of 17 QCA cells occupying 0.02 μm² area. Accordingly, the proposed structure is leveraged to implement ripple carry adder up to 20-bit with different degrees of approximation. A comprehensive structural and power consumption analyses over the proposed approximate and exact adders are performed. Then, discrete cosine transform computation of digital image processing architecture is used to show efficacy of the proposed approximate adder at the system level.