Power allocation in UMTS under SNR constraints

In this work, the distributed power allocation problem in universal mobile telecommunication systems (UMTS) is studied under feedback signal-to-noise ratio (SNR) constraints. The sources of uncertainty are assumed to come from the quantization process and measurement noise in the closed-loop system. The power allocation problem is formulated as a reference tracking problem of a pre-defined signal to noise-interference ratio. First, the synthesis problem with SNR constraints is studied as a 2-norm minimization process, which is equivalent to an LQR optimization problem. The solution of the associated Riccati equation in the LQR formulation is completely characterized, resulting in a state-feedback law with a special redundant structure. This control law is re-written in a transfer function format, where it is obtained a simple control strategy which is dependant on the round-trip delay in the feedback system. In addition, the corresponding 2-norm closed-loop performance is also studied. Thus, the selection of the weight in the LQR problem establishes a compromise between robustness to quantization errors and noise, and tracking performance. Finally, the analytical results are validated through simulation by considering time-varying channel gains and quantization in the feedback information.