Research On Adaptive Resource Allocation In MIMO-OFDM System

MIMO (Multiple Input Multiple Output) can produce parallel and independent channels in space and transmit multiple data streams. Spectrum efficiency can be improved dramatically under the constraint of constant bandwidth. OFDM has been regarded as an efficient approach to combat the adverse effects of multi-path spread by changing frequency selective channels into flat ones in frequency domain. Therefore, the combination of MIMO and OFDM has been the trend of future wideband wireless communication system.In MIMO-OFDM systems, we use a method to adaptively allocate data and power according to instantaneous channel state information. This method can fully utilize wireless resources when compared with static multiplexing schemes, so it can meet the requirements of high-rate and high-efficiency in the future mobile communication.In this thesis, we briefly introduce the main idea of MIMO-OFDM and summarize key techniques. Then information theory is considered and water-filling algorithm is deduced which is the basis of resource allocation algorithms discussed in the following.Based on what have mentioned above, we consider resource allocation techniques in single-user MIMO-OFDM system. Under the criterion of rate adaptive and margin adaptive, we discuss two basal adaptive methods and propose a new algorithm. Simulation results show that the proposed algorithm can reach the similar performance and reduce computation complexity significantly. It is very easy to implement when compared with original ones.In chapter 4, we put more emphasis on adaptive allocation algorithms in multi-user MIMO-OFDM system. We analyze a rate adaptive algorithm using matrix computation in detail. Because computation complexity of this algorithm is very high, we propose a new algorithm. Simulation results show that proposed algorithm has almost no difference in performance but running time of simulation is reduced greatly. Finally we give an illustration of Margin adaptive algorithms. Simulation results show that the algorithm can reach the capacity of system.