Modern software-defined radios are large, expensive, and
power-hungry devices and this, we argue, hampers their more
widespread deployment and use, particularly in low-power,
size-constrained application settings like mobile phones and
sensor networks. To rectify this problem, we propose to
put the software-defined radio on a diet by redesigning it
around just two core chips – an integrated RF transceiver
and a Flash-based, mixed-signal FPGA. Modern transceivers
integrate almost all RF front-end functions while emerging
FPGAs integrate nearly all of required signal conditioning
and processing functions. And, unlike conventional FPGAs,
Flash-based FPGAs offer sleep mode power draws measured
in the microamps and startup times measured in the microseconds,
both of which are critical for low-power operation. If
our platform architecture vision is realized, it will be possible
to hold a software-defined radio in the palm of one’s
hand, build it for $100, and power it for days using the energy
in a typical mobile phone battery. This will make software
radios deployable in high densities and broadly accessible
for research and education.