Research On Link Scheduling Algorithm Based On MIMO Wireless Network

With the popularization of wireless networks,various wireless terminal devices require the network with high transmission rate and high real-time performance.The introduction of Multiple-Input Multiple-Output(MIMO)technology in the physical layer enables the wireless network to obtain both the diversity gain and the multiplexing gain of the channel at the same time.It can exponentially increase the system throughput and improve the quality of wireless signal transmission without increasing the bandwidth.The main research of link scheduling is to allocate and use channel resources among different links,which are the basic problems and hot issues of wireless networks.The problem of link scheduling is closely related to the interference model which reflects the interference relationship.It is particularly important to choose a suitable interference model to describe the interference relationship among MIMO links.Instead of choosing the widely used protocol interference model and SINR model,this thesis chooses the MIMO-pipe model that is more suitable for MIMO networks.The model not only captures the tradeoff among MIMO links,but also simulates the interference among links more accurately.The main contributions of this thesis are as follows.Firstly,MIMO-based distribution link scheduling(MDLS)is studied.The algorithm redefines the constraints of the MIMO-pipe model-Strong-SINR(S-SINR)coexistence constraints and filters out the sets of links that meet the requirements.The screened link sets obtain the link scheduling set by participating jointly in the MDLS algorithm,and these links will transmit data during the data transmission phase.The innovation of this algorithm lies in the selection of the interference model and the redefinition of model constraints.Through theoretical analysis and simulation,it is proved that the MDLS algorithm can achieve about half of the system throughput in the ideal network model and better than the throughput performance of the CSMA algorithm based on the MIMO-pipe model.The second major part of the thesis is to propose Minimum-Delay CSMA(MD-CSMA)algorithm.The algorithm simulates a single physical channel into several virtual channels,and simultaneously calculates multiple feasible link scheduling sets so as to avoid frequent switching of algorithms among different schedulers.Finally,the problem of large delay of the CSMA algorithm is improved.The main innovation of MD-CSMA algorithm is to propose the concept of virtual channel.In addition,window-based traffic control is added as congestion control to the algorithm to further reduce the delay of the CSMA class algorithm.The upper bound of the expected packet delay for MD-CSMA algorithm is 1/(l,?)rlmin,where rimin has a constant lower bound of 1/4K(? + 2).In addition,the algorithm optimizes the average throughput of link l.In the thesis,both algorithms achieve the expected results that MDLS algorithm improves system throughput,and MD-CSMA algorithm reduces the delay of CSMA algorithm.