UC San Diego

Due to the large frequency band in which UWB systems may operate, interference mitigation techniques are necessary to reduce the interference caused by UWB systems to narrowband systems operating in the same frequencies. Based on LABI, coding techniques to spectrally shape a UWB signal's spectrum are presented, which insert in-band notches into UWB IR and OFDM based systems. They allow for the arbitrary shaping of the UWB spectrum without significant changes to the hardware. For UWB IR signals, notches can be created by modulating the timing intervals and pulse polarities. By adapting LABI to change one of these aspects of a UWB IR signal, two techniques called LABI TO and LABI PI are presented, and can create notches of -10dB and -18dB respectively. Due to the time hopping nature, the performance of these algorithms may be sensitive to timing jitter, with a rms jitter of 40ps increasing the notch power by 7dB. The effects of timing jitter can be ignored if the rms jitter is below 10ps. Improvements to these algorithms can be used to reduce the computational complexity of these algorithms without reducing the spectral shaping performance. By examining the calculations of the trellis branch metric of the Viterbi algorithm used in LABI, a Gray Code scheme is developed, and can reduce the complexity by approximately the complementary filter block span. An additional decrease in complexity can be achieved by jointly choosing the pulse polarities and timing intervals. The cascaded techniques LABI PITO and TOPI can lower complexity by implementing less complex Viterbi algorithms without reducing performance. For UWB OFDM based systems, LABI is again applied to code the data in the vicinity of a notch created by nulling out subcarriers. In-band notches of - 28dB can be created, and is an improvement of 20dB over an uncoded spectrum. Next, two hardware implementations of a Spectral Encoded system are discussed: SAW-based and monolithic. The former may be challenging to implement due to the high time-bandwidth product required of the SAW filters, while the latter provides an integrated, low-cost solution. A target system is designed using the monolithic solution, and a 5-bit 10GS/s DAC is required. This DAC is implemented using a 0.18[mu]m SiGe BiCMOS process, and consumes 10.2mW with the DNL and INL being within ±0.5LSB and ±1LSB respectively