GFDM Pulse Shaping Optimization Based Genetic Algorithm

Abstract

Generalized Frequency Division Multiplexing (GFDM) is one of the candidate schemes for the 5G and beyond. Its multicarrier modulation scheme and structure are independent blocks, and each block contains sub-carriers and sub-symbols. Sub-carriers are filtered with a prototype pulse shaping that shifts by time and frequency domain. This work uses Genetic Algorithm (GA) to assign the best parameters of the pulse shaping filter, which uses the error as a cost function. The algorithm assigns its parameters based on a minimum error by iteratively processing until reaching the pulse values that give the minor error. This method gave the advantage of reducing the error generated due to orthogonality and thus gave improved performance. This method initially depends on searching for filter values that provide the lowest error value and then adopting these values in building the GFDM transmitter and receiver. This method reduced the BER to 0.0107 at 10 SNR and 0.0033 at 25 SNR compared with the traditional method. Thus, this is a new method for building a GFDM transceiver system.