Dynamic Small-Cell Clustering Using Affinity Propagation Algorithm in Asynchronous 5G NR OFDM Systems

In a joint transmission (JT) system using orthogonal frequency-division modulation (OFDM), performance degradation due to asynchronous interference resulting from the time difference of arrival is unavoidable. We propose an affinity propagation clustering (APC) algorithm for radio remote heads clustering for downlink JT in an asynchronous OFDM system. In particular, by closely connecting the principles of APC and JT with synchronization mismatch, we newly design pairwise similarities and preferences of the APC algorithm. The proposed algorithm provides 1.5 times higher 5-percentile user spectral efficiency than the APC algorithm with a conventional similarity matrix, and it has the same time complexity in term of big O notation.