Research On Key Technologies Of Resource Allocation In MIMO-OFDM System With Imperfect Feedback

With the rapid development of mobile communication, the demand for multimedia serviceis growing explosively. The available frequency spectrum resource is becoming increasinglytense. So how to use limited resource to obtain greater system capacity, higher transmission rate,more perfect user experience is a more concerned problem. The thesis, aiming at the systemthroughput and users’ QoE(Quality of Experience), was focused on the research surrounded withresource allocation problems in MIMO-OFDM system, such as the feedback delay, feedbackerror and the limited feedback channel bandwidth. The main work is as follows.1. The methods of solving the resource allocation problems in ideal feedback single-usersystem and multi-user system were studied, and the existed limitations were pointed out. Tworesearch hotspots of the thesis were given–the resource allocation problem in imperfectfeedback situation and that based on users’ QoE, and then the lack of existed research results wasdescribed, which laid a foundation for following research.2. Aiming at maximizing the system throughput, considering the problems of feedbackdelay and feedback error, the subcarrier and power allocation algorithm based on feedback delayand feedback error was given. The algorithm uses the confidence coefficient and equivalentvariance of the feedback CSI(Channel State Information) to estimate the actual CSI. Assign thesubcarrier to the best user step by step, and then implement water-filling power allocation. Thetheoretical analysis and simulation results show that the algorithm can take full advantage ofmultiuser diversity, improve the system capacity effectively and is more robust against thefeedback delay and feedback error.3. In order to reduce the feedback overhead, a fair resource allocation algorithm based onlimited feedback was given. The algorithm uses the limited feedback techniques, and then thefeedback clusters CSI are used to estimate the other CSI linearly. Aiming at maximizing thesystem throughput under the constraint of users’ fairness, subcarrier groups, power and bits areallocated by estimative results step by step. The theoretical analysis and simulation results showthat the algorithm can reduce the system feedback cost effectively. It can not only improve thesystem throughput effectively, but also guarantee a good users’ fairness.4. Considering QoS(Quality of Service) can’t tell the user satisfaction with service exactly,a resource allocation algorithm based on users’ QoE was given from the service experience side.The algorithm describes users’ satisfaction by QoE, aiming at maximizing the average of users’QoE, determines the user priority based on QoE utility function, and then implements subcarriergroups and power allocation. The simulation results show that the algorithm can make full use ofsystem resources and improve users’average QoE effectively.