On Nonlinear Shaping Filters with Minimal Output Peak Value

Wednesday, October 19, 2016 - 4:30pm

Event Calendar Category

LIDS & Stats Tea

Speaker Name

Omer Tanovic

Affiliation

LIDS

Building and Room Number

LIDS Lounge

Abstract

One of the main challenges in modern digital communication systems is high peak-to-average power ratio (PAPR) of signals being transmitted. For example, this is the major drawback of orthogonal frequency division multiplexing (OFDM), a modulation scheme of choice in modern communications standards/systems. High PAPR can significantly degrade energy efficiency and system performance, when power amplifiers (PA) are employed to raise power level of a transmitted signal. Due to nonlinear operation of a PA, and the corresponding induced spectral leakage, additional processing of the transmit signal is needed before it is amplified and emitted by an RF antenna. This processing further increases PAPR of the transmit signal, where the main source of increase comes from so called shaping filters, which impose some necessary spectral mask constraints on communication signals. In this talk, we will discuss the problem of designing nonlinear shaping filters which simultaneously impose spectral mask condition and minimize peak value of the output signal. I will show that this problem can be formulated as an infinite dimensional convex quadratic program with box constraints, and will also present results for a special case when the optimal shaping filter can be analytically found as a series interconnection of a specific linear time invariant system followed by simple saturation non-linearity (clipper). I will discuss some theoretical questions that result from this problem formulation, and are related to ('cheap') implementations of optimal/near-optimal solutions. If time permits, I will also discuss some more general question emerging in modern digital communications pertaining to joint analysis of power efficiency, spectral efficiency and linearity.

Biography