Nonlinear signal processing with minimization of spectral distortion for embedded systems

Abstract Numerous digital signal processing algorithms originate from the implementation of analog circuits. This approach works well in the case of linear systems. However, there are many nonlinear signal processing methods which digital implementations are accompanied by undesirable spectral distortions. There is a great need to design alias-free and computational efficient algorithms for their microprocessor implementation dedicated to industrial applications. The scope of the paper is to propose an algorithm which employs the methods of analysis and synthesis of the signals in order to minimize the spectral distortion associated with the nonlinear processing of digital signals. The paper presents also the implementation of the proposed method using an inexpensive embedded system based on the ARM Cortex-A53 processor.