People:
Lingkai Kong, Prof. Elad Alon
To support future cognitive or software defined radio systems,
or simply to ease the task of integrating many different radio
standards onto the same die, there is a growing need to realize
radios capable of operation over a broad band of possible frequencies.
Although filtering can be applied at the baseband to select the
current frequency band of interest, the front-end of such a radio
must be capable of processing the frequency range of interest – making
the front-end inherently broadband in nature.
In order to enable such a broadband front-end while remaining
energy-efficient, this research focuses on applying feed-forward
linearization and distortion cancellation techniques to blocks biased
at relatively low current levels. Since many of these techniques rely on
careful matching of several circuit parameters, this project will also
explore low-overhead calibration schemes required to robustly ensure
optimal front-end performance.
In addition to the second and third order distortion issues commonly
encountered in narrow-band receivers, higher order distortion terms may
directly impact the RF front-end’s performance due to its wide-band nature.
Thus, one of the major challenges in implementing such a design is that
there is no clearly appropriate linearity metric (such as IIP3) that can
be used for practical design analysis and optimization. This research
therefore also explores simplified distortion metrics for these systems
that can be used in both the design and calibration of the front-end.
Fig. 1: Proposed front-end architecture with
feed-forward linearization
