Implementation of a modified K-NN Algorithm based 256-QAM using FPGA

Abstract

This paper proposes the design of K-NN algorithm based 256-Quadrature Amplitude Modulation for the fault identification. The main purpose of the proposed method is to use the quadrature amplitude variables of signals in which QAM is used to achieve high levels of spectrum efficiency. Generally, the KNN algorithm is used for both classification as well as regression predictive problems. Though there are several Machine Learning algorithms to classify, the KNN algorithm is an effective way to identify the fault based on the Euclidean distance between variables. The QAM is a modulation technique that combines the phase and amplitude modulation of a carrier wave into a single channel. The QAM may occur with error with respect to the amplitude or phase when fused with Sine and cosine waves. The proposed design utilizes the resolution of 8 bits to analyze 256 unique combinations of signals. The HDL code is developed for the proposed method and real time validated using the FPGA device. The performance characteristics such as power , area and timing are verified using the Xilinx Tool.