| Wideband code-division multiple-access (WCDMA) is a key technology to satisfy the ever increasing demand for high-speed wireless data. Another critical modern application is that of sensor networks, which consist of a large number of sensors with a limited power supply and computation, and who communicate through a wireless link. In this dissertation, we address technical challenges of each of these.;The first two parts of the dissertation consider the high peak-to-average power ratio (PAPR) issue of WCDMA systems, although the results are more generally applicable. High-level clipping is a rare event and hard to simulate. This dissertation analyzes the spectrum of these rare excursions for the complex Gaussian envelope, which is found to have properties that are extensions of the properties of real Gaussian processes. Furthermore, simulation results show that the statistical parameters derived from relatively common events for a low crossing level event can be used to estimate the spectrum in the rare-event high-level case. In addition, time domain PAPR reducing algorithms are simulated and compared.;A sensor network consists of a large number of nodes densely distributed across a region. Each sensor has wireless communication capability and some level of intelligence for signal processing and networking of the data. However, a limit power supply and a hostile environment in some cases constrain the data transmission. In this dissertation, we discuss group transmission of correlated information over the physical channel. The power of one node is too limited to reach the collector, but transmitting simultaneously with its neighbors, the effective signal-to-noise (SNR) is increased enough to allow detection. However, this cooperation can cause a reduction in spatial resolution. Our goal is to find the proper tradeoff between the mean squared error (MSE) increase caused by the spatial resolution deterioration and the MSE improvement caused by the SNR enhancement. |
