In this paper, we study the problem of resource allocation and optimization for multiple-input multiple-output (MIMO) cognitive radio (CR) systems under the assumption of imperfect channel state information (CSI) of the channels between the secondary users (SUs) and the primary users (PUs) at the SUs. We formulate robust design optimization problems for CR systems with one or more SUs communicating over a single or multiple frequency carriers in the presence of multiple PUs. We propose a linear matrix inequality (LMI) transformation that facilitates proper treatment of channel uncertainty at the SU transmitter and we provide solutions to the design problems based on convex optimization and Lagrange dual decomposition techniques. Finally, we demonstrate the importance of the proposed formulations and the implications of ignoring channel uncertainties when designing for CR systems.

In this paper, we study design problems for single and multiple-carrier Multiple-Input Multiple-Output (MIMO) Cognitive Radio (CR) systems. We assume imperfect Channel State Information (CSI) of the channels between the Secondary Users (SUs) and the Primary Users (PUs) at the SUs and propose solutions based on a Linear Matrix Inequality transformation that facilitates proper treatment of channel uncertainty at the SU transmitter. We employ convex optimization tools and use a Lagrange dual decomposition approach to solve the optimization problems efficiently. We also present a number of numerical results which clearly demonstrate the robustness and importance of the proposed algorithms.

In this paper, we introduce a new Differentiated Successive Interference Cancellation (DiffSIC) ordering technique for up-link multi-user Multiple-Input Multiple-Output (MIMO) systems. Unlike classical SIC, DiffSIC is capable of differentiating users according to their priority or class of service by selecting a detection order that best fits the users' service profiles. In addition, DiffSIC is able to achieve the optimal SIC detection order that results in the best overall system performance. In order to develop DiffSIC, we introduce analytical methods towards finding instantaneous symbol error rates (SER) for the Zero Forcing SIC (ZF-SIC) and Minimum Mean Square Error SIC (MMSESIC) detectors. We present a number of numerical results which clearly demonstrate the ability of DiffSIC to accomplish service differentiation and overall performance improvement in general.

In this paper, we introduce a new channel allocation scheme for up-link code division multiple access (CDMA)-based multi-user multiple-input multiple-output (MIMO) systems. In such a system, different users are treated as multiple antennas and space-time multi-user detection algorithms are used at the receiver to separate users sharing the same channel. In our allocation scheme, users are allowed to share multiple up-link codes simultaneously. This improves the overall up-link system capacity and fairly equalizes the bit error rate (BER) performance of different users

The broad spread of packet data networks and the emergence of applications in multimedia communications, has created a driving force towards an improved quality of service (QoS) model for today's Internet. A primary component of this model is packet schedulers. We introduce a new frame-based scheduling technique called mini round robin (MRR) that is primarily designed for providing lower latency bounds, and lower start-up latency bound for low-rate but high-priority flows. This enables applications such as voice-over-IP to demand low delay despite the low reserved bit rate of the voice sessions.

The broad spread of packet data networks and the emergence of applications in multimedia communications, created a driving force towards an improved quality of service (QoS) model for today's Internet. A primary component of this model is packet schedulers. We introduce a new frame-based scheduling technique called mini round robin (MRR) designed primarily for providing lower latency bounds, and lower start-up latency bound for low-rate but high-priority flows. This enables applications such as voice-over-IP to demand low delay despite the low reserved bit rate of the voice sessions.