With the increasing demand for self-healing techniques in mobile cellular networks, the fault detection method which is the first step of self-healing is studied. As user actions and the wireless environment greatly influence the key performance indicators (KPIs), most of the existing detection methods need to build multiple models to fit different operating scenarios of the network. In this paper, a novel detection model is presented that can automatically adapt to the normal variation of the KPIs caused by the change in environment and/or user actions, and accurately detect the abnormality caused by real system faults. The detection model is based on a linear prediction algorithm and the normalization process of the prediction deviation makes the model more simple and flexible to use. The proposed detection model has been tested in a simulated LTE environment, and the results show that the model can indeed detect real system faults while tracking the normal variations of the KPIs of the network.

No abstract

No abstract

No abstract

No abstract