Digital Signal Processing of Nonuniform Sampled Signals

Today the trend towards higher communication channel capacities and increased connectivity is inevitable. Wireless technology plays a vital role in this development. Electronic appliances are supposed to interoperate seamlessly together using self-configuring, short distance ISM band air interfaces like WLAN, Bluetooth, ZigBee and also ultra wideband (UWB), to name only a few. Software defined radio (SDR), cognitive radios and lately sensor networks have attracted a lot of attention. Particularly in those areas where ad hoc sensor networks are gathering environmental parameters, wireless communication remains the only option to establish communication links. Related to this problem is the topic of dynamic spectrum partitioning and selection of frequencies for communication bands (spectrum management).

This text bridges the gap between theory and real system implementation. It complements the vast amount of theoretical material found in the open literature by discussing hardware realisations and hardware architectures, too. It explains how the use of correlated additive random nonuniform sampling, compared to uniform (under) sampling, offers additional freedom in spectrum management. First, a theoretical section establishes the background by introducing relevant nonuniform sampling processes. Second, available algorithms and hardware architectures are presented. Third, novel algorithmic enhancements as well as a discussion of strategies for innovative sampling driver architectures are rounding up the considerations. Actual hardware was built and nonuniform sampled data sets were gathered laying the foundation for the algorithmic enhancements and extensions. Thus, the theory of nonuniform sampling is backed by practical results. Development towards the most efficient nonuniform sampling scheme has been done in a FPGA using VHDL.

A comprehensive work such as this is always embedded into and linked to the work of other groups and persons working in the same field. This work is a result of research and development funded by two European Union (EU) projects, namely EURODASP (IST-1999-21004) and DASPTOOL (IST-2001-34552). The project consortiums were formed by companies as well as R&D institutions with strong interest to promote the strategy of nonuniform sampling and to make novel contributions to this field, respectively.