Theoretical Analysis And Resource Allocation Algorithm Of Ultra Dense Network With Dynamic TDD

Efficiently adaptting traffic asymmetry and high communication capacity,dynamic TDD Ultra Dense Network(UDN)has been regarded as one of the important technologies in future communication networks.In dynamic TDD UDN,theoretical analysis and resource allocation are two important research directions.The former can statistically present system performance and its change trend under specific parameters through the derived performance expression,which provides a reference for system design and algorithm optimization,while resource allocation suppresses conventional interference and cross-link interference by allocating spectrum resources,thereby improving system performance.However,most of the current theoretical analysis research adopts the Homogeneous Poisson Point Process(HPPP),rather than the Poisson Cluster Process(PCP)which is more in line with the actual scenario,as the network model,and thus their derived results deviate from the actual situation.Moreover,existing resource allocation algorithms require accurate channel information and global collaborative optimization,resulting in high algorithm complexity and signaling overhead that limit their practical applications.In response to these problems,this thesis studies the theoretical analysis of dynamic TDD ultra-dense network under the PCP model and proposes the low complexity and low signaling overhead resource allocation algorithm.The theoretical analysis part is the focus of this thesis.This thesis firstly uses PCP to model the locations of base stations and users in UDN,and models users' service as a Bernoulli process.Based on above models,using stochastic geometry and probability theory,this paper derives the distribution of the spatial relationship between the users and the serving base station under the PCP model,including user connection condition probability,service distance condition distribution,and service distance and angle joint condition distribution,and provide theoretical support for subsequent research.Based on the above conclusions,this thesis completes the performance analysis of the two-tier dynamic TDD UDN in the full buffer scenario.In this scenario,considering two dynamic TDD modes(cluster-based dynamic TDD and base station-based dynamic TDD),this paper derives the coverage probability and area spectral efficiency of ordinary users and hotspot users under the general channel model.The accuracy of the derivation results is verified by simulation and then those analytical results is compared with that of the HPPP model.Then,in order to better fit the actual scenario,this thesis derives the single-tier dynamic TDD UDN performance expression of non-full buffer.Based on a more accurate LOS/NLOS channel model,this paper sequentially completes the derivations of the distribution of the numbers of service users of one base station,the two dynamic TDD time slot ratios,time slot utilization,and transmission probability,and completes the approximation of Rice fading,and then completes the expression derivations of user coverage probability and mean packet throughput.The accuracy of the derived results is verified by simulation.The above theoretical results under the PCP model can directly give the results and change trends of system performance metrics under specific parameters,and provide a more accurate reference for system design.Finally,this paper models the problem of non-cooperative intra-cluster resource allocation as a nonlinear integer programming problem through building an intra-cluster interference graph,then converts the problem into an linear integer programming problem,and obtains its optimal solution by branch and bound method.Simulation results prove that compared with random allocation algorithm,the algorithm proposed in this thesis can effectively improve the coverage performance of users with low complexity and low signaling overhead.