Application Of Cognitive Base Station Spectrum Management Algorithm In High Speed Railway

At present,the available frequency band of GSM-R is 885-889 MHz and 930-934 MHz.The quality assurance and service development of Railway Integrated Private Digital Mobile Communication System are severely restricted by the strategy of only 4 MHz spectrum resources and static spectrum allocation.As an important part of realizing the Chinese Dream,China's railways are steadily advancing towards the goals of safety,efficiency,warmth and convenience under the call of "transportation power and railway first".The requirements for the types and quality of wireless communication services in high-speed environments are bound to be higher and higher,which makes the traditional way of fixed spectrum allocation.It is no longer able to meet the increasing demand of railway spectrum.Cognitive radio technology is considered as an important method to solve the spectrum problem.Based on the above situation,this paper proposes an algorithm for spectrum management of cognitive base stations in high-speed railway.It is hoped that the frequency of spectrum switching of cognitive users can be reduced by reasonable planning and decision-making,and the problem of spectrum limitation and enhancement can be solved.The main works of the thesis are as follows:The Enhanced Learning Algorithm(Q-leaming)is applied to the spectrum management of the cognitive base station.The q-leaming algorithm is used to study the behavior of the main user near the base station with the appropriate learning rate and damping coefficient,and the channel is evaluated by the cumulative Q value,and the appropriate channel is allocated dynamically for railway users.The simulation experiment of simple scene and complex scene shows that the frequency-hopping frequency based on Q-learning spectrum management algorithm is less than that based on polling,which proves that the spectrum algorithm based on q-leaming is more advantageous.Fusion HMM channels prediction and q-learning spectrum management algorithm.In the cognitive base station,the channel model is established by the HMM algorithm based on the historical information and the forecast information.HMM is combined to Q-learning algorithm,and to improve the accuracy of channel evaluation.The simulation results show that Fusion HMM channel prediction and q-learning spectrum management algorithm can effectively reduce the frequency of frequency hopping,and thus help to improve the communication quality.Q-leaning algorithm based on cooperative improvement of multiple base stations is proposed.Railway communication along railway is composed of several base stations,the switch between base station and base station is a problem to be solved in railway communication.An information cooperation of multiple cognitive base stations is shared to each other to improve the accuracy of channel evaluation.The simulation results show that Q-learning algorithm based on cooperative improvement of multiple base stations can reduce the frequency of spectrum switching.In the process of high-speed train movement,thousands of users will initiate the handover request simultaneously.At the same time,networks of different will exist simultaneously in the development of railway.The papers propose a cognitive base station model based on heterogeneous wireless network is proposed to solve the spectrum resource allocation problem in mufti-user switching.In eognitive base stations of heterogeneous wireless networks,public networks and railway private networks are regarded as sub-base stations and cognitive base stations as main base stations.When the arrival of trains causes multi-user handover,other business users are transformed from normal state to cognitive state by combining packet spectrum management algorithm.