Abstract
Software radio is a deceptively simple idea: Identify all the features that specialize an RF communications device to a particular waveform (e.g. GSM cell phone or FM walkie talkie), and implement these features in flexible software on a generic platform, rather than in fixed-function hardware. This change should provide significant advantages compared to legacy hardware radios: the ability to support multiple waveforms on the same device, to upgrade the waveforms on the device through software downloads, and to dynamically adapt modulation or other physical layer parameters to a wide range of channel conditions. However, taking full advantage of software radio turns out to require a broad change in communications systems. Affected hardware components include antennas, filters, A/D converters, and power amplifiers. Affected device-level software components include signal processing, timing control, inter-layer APIs, and security. The hardware architectures optimal for this software are not GPPs, DSPs or FPGAs, but an interesting hybrid among these approaches. At the network level, new MAC algorithms and topology management are needed to exploit the flexibility of individual nodes. In this talk I present a slice through these varied and interrelated systems and research challenges, taking a total systems view of software radio.