| In order to meet the explosive growth of the demand of data traffic in the future wireless network, the next generation of mobile communication network has been considered from the very start with higher requirement of network capacity. Among the candidate key technologies of the next generation(5G) of wireless communication, dense networks have been recognized as an effective way to enhance the network capacity and coverage. At the same time, in order to reduce network operator's capital expenditure and operating expense in dense networks, study on self-organizing network(SON) has greatly aroused industrial and academic's enthusiasm. Capacity and coverage optimization(CCO) is an important use case defined in SON by 3GPP(The 3rd Generation Partnership Project), which aims at enhancing the network overall performance and cell edge users' performance through self-optimization method under the network deployment.Firstly, the thesis describes the overall architecture and functional modules of the simulation platform, including wireless channel module, data flow module, MAC(Media Access Control) layer resource scheduling module, mobility module, AMC(Adaptive Modulation and Coding) module, etc.Secondly, based on the system level simulation platform, capacity and coverage optimization related modules have been designed and realized. Considering the importance of the realistic channel model for capacity and coverage performance, a 3-dimension(3D) channel model is developed according to the standard of 3GPP TS36.873 based on the basic system level simulation platform. The 3D channel model mainly includes 3D pathloss model, shadow fading model and 3D antenna gain model and is well tested and calibrated afterwards. On the other hand, in order that users can connect to the target cell which will provide more stable coverage and better quality of service to guarantee the network performance, a RRC(Radio Resource Control) layer handover process is modeled under the guideline of the standard of 3GPP TS36.331, which is very crucial and essential for system level simulation. The handover modules are mainly consist of TTT(Time To Trigger) module, measurement report trigger and reporting module,handover preparation and execution module, and etc. After completing capacity and coverage related system platform, a capacity and coverage optimization SON entity has been designed and added. The SON entity main includes data collection module, evaluation and analysis module, optimization and processing module and execution module. At last, on the sound foundation of the simulation platform, system performance is simulated and analyzed under heterogeneous dense network deployment and reasonable suggestions are given for dense small cell deployment in city scenario.Finally, In order to solve the capacity and coverage problem caused by irregular small cell dense deployment in heterogeneous cellular networks, this thesis proposed a reinforcement learning control system for joint capacity and coverage self-optimization through small cell power control. The proposed algorithm was based on reinforcement learning control system, which was a combination of fuzzy logic and Q-learning algorithm. By taking a comprehensive consideration of network average user performance, edge user performance and the influence among network environment, this thesis developed a cooperative reward for capacity and coverage joint optimization algorithm. Simulation results show that the proposed method has the ability to achieve capacity and coverage optimization in small cell dense deployment scenario, which effectively improves the average user throughput and edge user throughput. |
PDF Full Text Request