| Novel adaptive strategies for maintaining quality of service (QoS) in mobile communications while maximizing efficiency are presented in this thesis. The unifying theme is that communication systems should adapt to the environment: when the channel is hostile, robust modes should be used to maintain good QoS; however, when benign channel conditions are encountered, higher-throughput modes should be invoked for efficiency. First, a time-varying channel model explicitly taking into account changing environments is developed. Under this channel model, methods for tracking channel stationarity are proposed based on analysis of variance, correlation and information-theoretic criteria. Tracking enables constructing variable-size bursts for efficiency. Next, adaptive modulation for improving throughput is proposed for uncoded and coded systems, respectively using mean-square error and bit-flipping metrics. Finally, performance of the proposed methods is evaluated in a semi-blind framework. Obtained results show feasibility of adaptive methods for improving efficiency compared to the non-adaptive counterparts. |
