Measurements design and phenomena discrimination

TL;DR

This paper presents a method for designing measurements that can effectively discriminate between different phenomena by cancelling irrelevant signal components using oblique projectors and adaptive pursuit strategies.

Contribution

It introduces a novel measurement design framework employing oblique projectors and nonlinear adaptive pursuit techniques for phenomena discrimination.

Findings

Effective discrimination of phenomena using designed measurements

Use of oblique projectors for signal component cancellation

Adaptive pursuit strategies improve measurement selection

Abstract

The construction of measurements suitable for discriminating signal components produced by phenomena of different types is considered. The required measurements should be capable of cancelling out those signal components which are to be ignored when focusing on a phenomenon of interest. Under the hypothesis that the subspaces hosting the signal components produced by each phenomenon are complementary, their discrimination is accomplished by measurements giving rise to the appropriate oblique projector operator. The subspace onto which the operator should project is selected by nonlinear techniques in line with adaptive pursuit strategies.