Research On Coverage And Capacity Optimization In LTE System Under High-speed Railway Environment

Depending on the huge advantage of running speed, high-speed railway gradually becomes the first choice of people's travel, meanwhile, people enjoy the convenience of high-speed railway also hope enjoy high quality communication services in the course of journey. Currently, as the second generation narrow-band communication system, GSM-R technology has been used in the railway communication system, which is unable to meet people's growing demand for broadband services, therefore LTE technology with the advantages of wide spectrum, low time delay, strong coverage, has becoming the best technical means to solve the railway mobile communication problem. However, due to the special nature of high-speed environment, coverage and capacity issues are still exist along high-speed railway, so the research on coverage and capacity optimization under high-speed railway environment has important theoretical and practical value.Firstly, this thesis introduces the features of LTE communication system and the function of self-organization network, and analyzes the wireless network coverage and capacity performance. On this basis, the fuzzy Q learning(FQL) algorithm and cooperative fuzzy Q learning(co-FQL) algorithm both of which applied in the optimization of coverage and capacity are introduced.Secondly, this thesis focuses on the coverage and capacity optimization algorithm under high-speed railway environment. Due to the high running speed will lead to penetration loss, Doppler frequency shift and other issues, this issues may deteriorate the communication environment, so that users cannot get the continuous high quality experience in the aspect of communication needs. Therefore, on account of the superiority that the location of base station and rail train can be predicted in railway environment, and the traditional fuzzy Q learning algorithm which single adjusts the antenna downtilt angle, this thesis introduces the concept of multi-agent system, and constructs multi-agent cooperative fuzzy Q learning(ma-FQL) algorithm. By jointly adjusts antenna downtilt angle of two base station in the optimization area, ma-FQL algorithm can achieve optimization in a faster rate. In order to achieve a more ideal network performance, this thesis introduces the concept of coordination mechanism and proposes coordination multi-agent cooperative fuzzy Q learning(coordination FQL) algorithm. Based on ma-FQL algorithm, this algorithm adjusts the transmit power with downtilt angle coordinately, which not only improves the optimal rate but also enhances the probability of overall network achieving global optimal performance.Finally, this thesis set two simulation scenarios that weak coverage and over coverage for analysis. By changing train's speed, traditional algorithm and improved algorithm are simulated and analyzed in terms of spectral efficiency and throughput. Simulation results of weak coverage scenario show that, compared to FQL algorithm, co-FQL algorithm and ma-FQL algorithm, coordination FQL algorithm can effectively improves the spectral efficiency and throughput performance of overall network. Meanwhile, in order to further verify the coordination FQL algorithm's ability of dealing with different wireless environment, simulating it in the over coverage scenario. The results show that under the over coverage scenario, coordination FQL algorithm also has a good optimization effect in terms of spectral efficiency and throughput. In addition, under the two kinds of simulation scenarios, with the growth of train speed, coordination FQL algorithm also increases the improvement of whole network performance, indicating that it can be well adapted to high speed environment.