Analysis And Application Of LTE Network Uplink Interference

LTE (Long Term Evolution) is a long term evolution of the general mobile communication system technology launched by 3 GPP. Recent years, with TD-LTE and FDD-LTE mobile communication system are widely used in commercial. 2G, 3G and 4G network co-exist. Network structure evolves continuously. As a result, the wireless environment LTE network faced is more and more complex, and the network interference problem becomes more and more serious, affecting the network quality.It is an important job for mobile operator network maintenance to detect and analyze the LTE network uplink interference. This paper analyze the current situation for LTE network interference investigation and prove the necessity of designing a fast and automatic interference detection and classification tool to improve the efficiency. The LTE network architecture and the domestic 2G/3G/4G spectrum distribution are induced as a theoretical basis for further study of TD-LTE system uplink interference. Through extracting the PRB waveform characteristics for six kinds of interference in LTE networks, a frequency-region domain based joint interference analysis algorithm is proposed with the region feature of the interference source. This paper also implements a detection and classification software for LTE network uplink interference.The main work of this paper is divided into three parts:(1) Aiming at detecting interference in the whole network efficiently, we analyze the interference generated mechanism, regulation methods and the corresponding interference wave characteristics in detail.The interferences include harmonic interference, intermodulation interference and spurious interference, adjacent channel interference,blocking interference and external interference. A multi-dimensional joint interference analysis algorithm is proposed after designing a PRB interference waveform characteristic computing method and an interference modeling method. The interference classification model is established by random forest algorithms with PRB interference feature data. This model can be used for classifying interference data. After combining theregion feature of interferences, we infer the possible interference source. With the theory of interference generation, we can ascertain the interference type;(2) We describe the implementation steps of the key technologies for the interference analysis. The key technologies include interference cell evaluation, interference feature extraction and modeling, interference classification algorithms based on random forest and the co-location information match and the frequency information match.(3) From the point of view of engineering application, we develop an uplink interference detection and classification software system for the LTE network. This system is based on the above key technologies and is implemented via C# programming language and SQL Server database in the.NET4.0 platform. After testing the proposed key technologies and the software systems with real inference collected from the practical application, we evaluate and analyze the system with the evaluation indexes such as the detection rate, error rate and correct rate.Experimental results demonstrate that the proposed multi-dimensional joint interference analysis algorithm can determine the interferences of the test cell efficiently and classify the interference type accurately. As a result, it can improve the analysis and exclusion efficiency for LTE network uplink interferences.