Communications Systems using 1-Bit Quantization and Oversampling: Aspects on Synchronization and Runlength-Limited Coding
Buch in deiner Nähe kaufen
Beschreibung
Future wireless communication systems must accommodate increasingly diverse use cases: Connecting billions of devices, supporting ultra-high data rates in the 100 Gbit/s range, and ensuring ultra-low latencies — all while remaining energy-efficient. A key challenge in achieving energy efficiency lies in the power consumption of analog-to-digital converters (ADCs), which scales quadratically with the input bandwidth in wideband systems. One promising solution to this problem involves reducing ADC amplitude resolution. In the extreme case, lowering the resolution to 1 bit not only minimizes power consumption but also enhances robustness against nonlinearities in the radio frequency (RF) front end. To counteract the loss in rate that comes from the reduction of amplitude resolution, a particularly effective approach is shifting resolution from the amplitude domain to the time domain by combining 1-bit quantization with temporal oversampling. This method aligns well with modern semiconductor technology, which favors fast switching capabilities while providing limited voltage headroom for amplitude processing. However, employing temporally oversampled 1-bit quantization in communications systems necessitates the development of an adapted modulation scheme — called zero-crossing modulation (ZXM) — and the redesign of receiver algorithms. In this regard, we study three topics to deepen the understanding and enhance the practicability of systems employing ZXM: First, we derive bounds on phase noise estimation accuracy under 1-bit quantization. Considering a finite transmission time, we formulate the Bayesian Cramér-Rao bound (BCRB) for a system where blocks of data and pilot symbols are transmitted in a periodic fashion. Further, we derive an extension for infinite transmission time. Second, we investigate the limits of digital synchronization techniques for ZXM systems by exemplarily looking at constant timing and constant phase offsets. Since the 1-bit quantized samples do not constitute a sufficient statistic of the received signal, the theoretical foundation for digital synchronization is not given, which raises the question to which extent digital synchronization can still be applied to ZXM systems. We answer this question by investigating a lower bound on the achievable rate as a function of the channel parameters. Moreover, we derive two practical algorithms for timing synchronization and one algorithm for phase synchronization. We also evaluate the end-to-end system performance, considering the synchronization algorithms and practical transmitter and receiver implementations. The considered ZXM waveform relies on a technique called runlength-limited (RLL) coding for the creation of the transmit signal and the RLL codes presently used have been shown to exhibit a suboptimal performance in the low signal-to-noise ratio (SNR) regime. Therefore, the last part of this thesis deals with the investigation of means to improve the performance of RLL codes under channel noise. We investigate the reasons for the suboptimal performance of the present RLL codes and derive a new class of RLL codes that allow for optimization of their performance in the low SNR regime.
Autor*in
Stephan Zeitz