Efficient Cell Outage Detection in 5G HetNets Using Hidden Markov Model

Next generation 5G wireless systems envision ultra dense networks with a huge number of heterogeneous cells. This makes the management of such heterogeneous networks (HetNets) very complex and practically impossible without any automated procedure. Self-organizing networks (SON) are expected to provide self-configuration, self-optimization, and self-healing functions for automated management of 5G wireless networks. Cell outage detection is identified as a critical problem that requires efficient self-detection process. In this letter, we first classify the 5G base stations (BSs) into four different states. Subsequently, we explore a hidden Markov model to automatically capture current states of the BSs and probabilistically estimate a cell outage. Simulation results on typical, dense 5G HetNets demonstrate that our proposed strategy is capable of predicting the state of a BS at an average of 80% accuracy, as well as correctly detecting a cell outage ~ 95% of the time.