CAREER: Translinear Adaptive Filtering

Cornell University, Ithaca NY

Investigators

Abstract

Portable computer and communication systems are all around today and they will only become more prevalent in the days to come. The design of such mobile electronic systems involves a number of significant engineering challenges. Because they operate from batteries, they must both consume very little power and operate on a low power-supply voltage. Moreover, because they are always on the move, such systems must be able to dynamically adapt their performance to an ever-changing environment. This research involves the development of a unified framework for systematically designing and implementing efficient adaptive signal-processing systems that consume very little power and can operate on a low power-supply voltage with applications in the areas of mobile communication systems and enhanced human-computer interfaces. The class of dynamic translinear circuits implements the feed-forward signal processing functions needed for continuous-time linear, polynomial, and rational-function adaptive filters with very few devices. Such circuits are capable of operating with low power consumption and on a low power-supply voltage. This research involves the design an implementation of continuous-time linear and nonlinear adaptive filtering algorithms with dynamic translinear circuits with applications in the areas of mobile communication systems and enhanced human-computer interfaces. The primary focus of this research is on the efficient hardware implementation of such adaptive filters, assessing how device nonidealities and noise impact system performance and developing robust architectures and adaptive algorithms.