The advantages of Orthogonal Frequency Division Multiplexing (OFDM) and Multiple-Input-Multiple-Output (MIMO) systems in wireless communications are now well-established. These techniques have found application in many current systems, including WiFi and WiMax, and the consensus opinion is that fourth generation (4G) systems will almost certainly be based on OFDM and MIMO. In the proposed work, we consider algorithms for the PHY and MAC layers of multiuser OFDM and MIMO systems when various system parameters are not perfectly known at the transmitter; in addition to standard performance measures, such as bit error rate (BER), computational complexity, robustness, and fairness are also considered. In particular, the link adaptation problem in an OFDM system, the resource allocation and uplink synchronization problems in a multiuser OFDM context (OFDMA) are investigated, and, in the multiuser MIMO context, techniques are described for designing transceivers when spatial multiplexing is used.; We first investigate the link adaptation in OFDM systems when time variations in the channel cause the channel state information to be outdated, therefore degrading the performance. A robust adaptive modulation scheme with transmit and receive diversity is introduced and its performance is studied. Then, we study the resource allocation problem in an OFDMA system with multiple antennas. Specifically, we introduce a low complexity and robust algorithm which optimally assigns the subcarriers and power to different users given a fixed power budget and QoS constraints (for example, BER and user data rate). Later, we study timing synchronization for the uplink of an OFDMA system and analyze the performance loss when the guard interval is not long enough to absorb the propagation delay and the impulse response of the channel. A timing offset estimator which is robust to additive white Gaussian noise and interchannel interference is introduced and the optimality of the estimator is demonstrated.; We also study the design of precoding and decoding matrices for the downlink of a multiuser MIMO system when spatial multiplexing is used. We investigate the block diagonalization approach for multiuser MIMO systems and propose a random precoding technique that schedules users for transmission with only limited feedback from the receivers. We then introduce a new optimization criterion for designing a linear transceiver which tries to minimize the maximum mean-squared error among all users or data substreams; we show that this approach results in fairness among users and improved average BER.
展开▼
