Research On Self-healing Technology Of The IMT-Advanced System

The research on the Self-Organizing Network (SON) technology has become animportant part of the standardization work in the IMT-Advanced system. Comparedwith the traditional wireless network, SON by means of self-configuration, self-optimization and self-healing technology not only can effectively increase the networkperformance and improve the user experience, but also can reduce the complexity ofnetwork operations and operational costs. Self-healing can maintain networkperformance by detecting, localizing and curing the problems automatically.Around the cell outage self-healing scenario, cell outage detection and cell outagecompensation are studied in depth. The optimization algorithms and programs areproposed and simulated. The main works of the thesis are as follow:Firstly, according to the cell outage detection of the self-healing technology, a celloutage detection method under the hybrid SON architecture is proposed, which is basedon the Channel Quality Indicator (CQI) information reported by the users. The methodutilizes the CQI mechanism existing in the system to detect the cell outage, withoutintroducing the additional signaling overhead and making full use of the CQIinformation. Compared with the existing detection method based on the historicalinformation, the proposed method can detect the cell outage in real time. The resultswhich based on LTE system simulation platform show that the detection algorithm caneffectively detect the cell outage.Secondly, according to the cell outage compensation of the self-healing technology,a compensation method after cell outage detection process is proposed, which is basedon the fuzzy Q-learning (FQL) algorithm for compensation under the hybrid SONarchitecture. Through adjusting collaboratively the antenna down tilt, the adjacent cellsincrease the coverage to compensate the outage cell user performance. The FQLalgorithm fuzzy the states of Q-learning to reduce the number of state and action-statetable and learn from the interaction with the surrounding environment to adjust theparameters in order to accomplish compensation. Compared with the centralizedmethod which requires reporting the data to management center in real-time, the proposed method has better timeliness. At the same time, the method utilizing thereal-time data to adjust parameters can adapt to changes in the surrounding environment.The results which based on LTE system simulation platform demonstrate theeffectiveness and stability of the proposed method and provide technical reference forthe cell outage compensation research.