Rush Hour Capacity Enhancement in 5G Network Based on Hot Spot Floating Prediction

Abstract

Rush hour network capacity enhancement is one of the most hard-to-solve problems in the field of network optimization. With the bust of data traffic requirement in cellular communication, 5G network will face this challenge in future. However, there is few effective solutions to this problem for 5G network optimization. In this paper, a novel solution based on hot spot floating prediction is proposed. Our solution consists of a traffic prediction method for hot spot floating trend estimation, and a semi-dynamic distributed unit (DU) and active antenna unit (AAU) mapping to fit the forecasted high traffic burst with proper DU-AAU mapping. The proposed solution could fit real network irregular gNB distribution. Simulation outcomes indicate that the hot spot floating prediction precision outperforms around 10% in normalized root mean squired error with the existing prediction methods, and our semi-dynamic DU-AAU mapping solution receives a 20% throughput gain on average compared with that without DU-AAU mapping network solution.