Publication

Abstract : The performance of a wireless system is directly dependent on the accuracy of retrieving the transmitted message from the noisy signal received at the demodulation end. Over the years many methods have been developed to reduce the Bit Error Rate(BER) at the receiver. Recently, machine learning algorithms have been explored as an alternative to traditional symbol detection methods to enhance overall performance. M-ary QAM is a popular digital modulation scheme used in Wireless communication owing to its high data rate. Nevertheless, for a given signal-to-noise ratio, the BER for QAM is higher than the BPSK or QPSK systems, which provide a lower data rate. The BER can be reduced using ML classifiers such as the K-Nearest Neighbours Algorithm (KNN) to leverage the high data rate of QAM and have more efficient symbol detection. This paper describes the implementation of KNN to QAM symbol detection and the resultant decrease in BER is observed to be around 4-6 dB.