Adaptive techniques for data conversion and coding with application to power control in CDMA wireless systems

Digital communication systems transmit information in digital form. In TV and radio broadcasting, for example, the source of information is analog. The information would then need to be transformed into digital form in order be suitable for transmission or storage. This process is performed by analog source encoding. Such source encoders achieve high transmission efficiency decreasing the bit rate needed to represent the information while maintaining acceptable accuracy and robustness against channel distortions.; Existing source encoders can suffer from limited SNR and dynamic range performance. This dissertation develops a variety of structures for adaptive delta and sigma delta modulators that address these difficulties. This is achieve by introducing a new scheme for adapting the step size of the one-bit quantizer in the modulation loop. In order to get further performance improvement, these systems are extended to the multi-bit case, where multi-bit quantizers are allowed. It is shown that the resulting systems maintain the same attractive properties single-bit case, namely, improved SNR and superior dynamic range. All systems are analyzed both analytically and by simulations.; The ideas developed for delta and sigma delta modulation turn out to be relevant for power control schemes in wireless communications. As is well known, the requirement of power control in the uplink channel of DS-CDMA system is a critical limitation. Power control is needed because all users share the same bandwidth to transmit data and thus inter-users interference will occur. The delta modulation-like behavior of closed loop power control, however, makes it suffer from slow tracking in the presence of fast and deep fading of the wireless channel condition. This dissertation analyzes the causes of such limited performance and develops several algorithms to compensate for these causes. The resulting schemes show improved power control performance.