Non-Euclidean Robustness Measures for Communications and Signal Processing

In this paper we develop a new robustness measure which is based on techniques from differential geometry. Modeling imperfectly known quantities as elements of a non-Euclidean manifold, we show how the tangent space can be used to define a method of measuring robustness for a variety of communications and signal processing algorithms. We then illustrate the technique by example and show that the analysis can be greatly enhanced through the use of stereo imaging.