The study of self organized networks is considered a key element in developing the new generation of cellular networks. Mobile service providers are adopting new technologies such as small cells and precoding to deploy their networks according to 5G standards. As these type of technologies gain more importance and popularity, network self organization become a more essential function in operating the network resources. Self organized networks have been classified into three main categorizes: self optimization, self configuration, and self healing. The first two categories showed further development and research interest when compared to self-healing. Due to the high density nature of small cells, using massive antennas in the network, and their susceptibility to failures for various reasons, the demand for a clear self-healing process and architecture is deemed urgent. In this paper the current self-healing process in addition to the fault tolerance aspects for the future 5G are studied and a new process model for organizing the self-healing process is proposed. The new process model is meant to map the different functionalities needed to perform a successful self healing process. The model with the proposed descriptive network architecture aims to identify the different functions within the self-healing process model. In order to test the operation ability of the model, a new big data aspect is added to the network architecture to aid in analyzing the huge amount of data needed to efficiently perform the self-healing process. Results show that the proposed precoding technique in conjunction with the machine learning algorithm based on a decision tree model that uses empirical data collected from the network can identify the status of cells (healthy, congested or failing) and suitable self-healing procedures can be triggered to recover the cell accordingly.

自组织网络的研究被认为是发展新一代蜂窝网络的关键要素。移动服务提供商正在采用新技术，例如小型蜂窝和预编码，以根据5G标准部署其网络。随着这些类型的技术变得越来越重要和流行，网络自组织成为操作网络资源中更为重要的功能。自组织网络已分为三大类：自我优化，自我配置和自我修复。与自我修复相比，前两个类别显示出进一步的发展和研究兴趣。由于小型小区的高密度特性，因此在网络中使用大型天线，并且由于各种原因它们容易发生故障，对清晰的自愈过程和体系结构的需求被认为是迫在眉睫的。本文除了研究未来5G的容错能力外，还研究了当前的自愈过程，并提出了一种用于组织自愈过程的新过程模型。新的过程模型旨在映射执行成功的自我修复过程所需的不同功能。具有建议的描述性网络体系结构的模型旨在识别自我修复过程模型中的不同功能。为了测试模型的操作能力，网络体系结构中增加了新的大数据方面，以帮助分析有效执行自我修复过程所需的大量数据。