An FPGA-Based Approach to Eliminate the End Effect of EMD Using Grey Prediction Model

Empirical mode decomposition (EMD) is a widely used algorithm for the analysis of nonlinear and nonstationary signals. The output obtained by EMD is known as intrinsic mode function (IMF). The IMFs of the EMD method have a problem which is known as end effect. Due to the presence of end effect the accuracy of the IMFs is decreased. In this paper, in order to eliminate the end effect grey prediction model (GPM) is used with EMD. A novel FPGA based design for implementing grey prediction model with EMD is proposed in this work. For the implementation of the proposed design, an Artix-7 FPGA is used. To show the effectiveness of the proposed design, root mean square error (RMSE) values are calculated for the output of the simple EMD design and GPM with EMD design. The RMSE values in case of GPM with EMD design are lesser, which means the outputs of the GPM with EMD design are more accurate.