Power efficient wireless video communications over CDMA networks

Effective resource management is essential to the success of cellular communications. It not only increases the system capacity but also prolongs the battery life. This thesis focuses on designing power efficient wireless terminals that transmit video signals in the uplink of a code division multiple access (CDMA) cellular system. Since video compression consumes a significant amount of power, this research takes into account signal processing power as well as transmission power.; We first study the video encoder, one principal component in a wireless video terminal. By analyzing key components of a H.263 video encoder, we develop power consumption models that are applicable to most modern video encoders using block-based hybrid video coding structures. Measurement data are used to verify the models.; We then propose an algorithm for optimum video transmission that deals with a single user. Video compression, channel coding and radio transmission are adjusted to minimize the total power consumption subject to video duality constraints. Simulation results demonstrate that a significant power saving is obtained by our adaptation scheme.; Considering the interference caused by the transmission power from different users in a practical multiuser cellular network, the transmitter power determined for one single user may not be appropriate in practical multiuser networks. We first derive analytical solutions based on simplified distortion models and power consumption models. Compared with the existing full search algorithm, this approach not only expedites the optimization process but also provides valuable insights regarding the influence of different operating parameters on the optimization. We also derive a novel fast two-step algorithm for practical encoders with more complicated distortion-rate models. By separating the complex optimization problem into two steps, the fast algorithm reduces the computation burden significantly at the base station and obtains global minimum. This algorithm can be applied to a variety of source coders and channel models.