Computing Platforms for Software-Defined Radio

Software-Defined Radio (SDR) has been a hot topic both in academic and industrial research already for several years, yet no final definite solution to the implementation architecture for it has emerged. As SDR is a key enabling technology for taking the next step, implementing a true Cognitive Radio, it is high time to overview and discuss the proposals for SDR architecture. The computing platforms for SDR have not been addressed thoroughly and holistically in the literature by now. There are a few books that analyse from a theoretical point of view algorithms, software and hardware techniques to implement SDRs, but a clear point of reference for computer architects is missing. The approaches in this book can be divided into three main categories: Architectures exploiting parallelism by extending single-processor environment (such as VLIW, SIMD, TTA approaches); Multi-core platforms distributing the computation to either a homogeneous array or a set of specialized heterogeneous processors; and, Reconfigurable architectures exploiting fine-grain, coarse-grain, or hybrid reconfigurability.