Investigations On Self-healing Techniques For Ultra Dense Wireless Networks

With the development of the super dense heterogeneous network,the density and the variety of base stations grows rapidly,and results in more complex network management which is difficult for artificial management.Hence,Self-Organizing Network(SON)is adopted in LTE-A networks to meet the requirements of the network management.As an important part in SON,Self-Healing technology has attracted much attention.Self-Healing technology of the mobile communication networks aims to recover from the network failure without manual intervention after a network failure.Self-Healing technology mainly includes three parts,the cell-outage detection,cell-outage diagnosis and cell-outage compensation.The core of the cell-outage detection and cell-outage compensation is to collect the history parameters of the system to setup a model for parameters adjusting,and then by monitoring the change of network parameters,possible cell outage can be identified.Then,the network automatically generate and execute the compensation operations to recover the services of the outage cell.In this thesis,cell outage detection is investigated,and a TCM(Transductive confidence machines)based automatic cell-outage detection algorithm is proposed.TCM algorithm is a transductive algorithm.It makes decision on the class that data belongs to based on the training data.Before running the algorithm,the training data need to be collected to build the detecting model.During the detecting period,the P value relative to each category for the testing data is evaluated.The P value means the probability of the appearance of the testing data under the relative category.Then according to the P value,the final judgment is made and the decision confidence is given.TCM can guarantee high detecting accuracy under the premise of greatly reduced false positive rate.More importantly,in the presence of noise data of the training set and small training set,it can guarantee high detecting accuracy.Then,an improved TCM based algorithm incorporating hypothesis test with the Neyman-Pearson criterion is proposed to improve the detection accuracy of the algorithm.Cell-outage compensation algorithm is investigated.Considering the coverage rate,overlap coverage rate of the system and the system capacity simultaneously,we model the compensation algorithm as an optimal problem related to the coverage threshold.By adjusting the base station downlink transmission power,optimal compensation scheme is obtained using PSO algorithm.The simulation results show that proposed cell outage compensation algorithm can compensate the performance loss of the outage cell by a large extent.