Forty-six Years of Statistical Signal Processing

点击翻译按钮获取中文摘要

Forty-Six Years of Statistical Signal ProcessingThe influence of geometrical ideas in statistical signal processing dates to the inception of the field. Linear subspace ideas have had enormous impact and their success has led to exploitation of increasingly sophisticated geometrical structures, such as Grassmannian manifolds of linear subspaces, as domains for statistical reasoning about signal detection, parameter estimation, and inference. There is a trend in recent statistical signal processing research (and in related areas, such as statistical learning) toward exploitation of low-dimensional structure in data presented in ambient spaces of much higher dimension. This development continues to bring new geometrical machinery into the foundations of the field. In the forty-six years since completing his doctoral studies at the University of Washington, Louis Scharf has brought geometrical insights to almost every aspect of statistical signal processing. Because his work in the discipline has been so widespread, contributing significantly to almost every major topic within the field, building this workshop around the theme of his career?s influence on statistical signal processing provides a natural framework for a snapshot of the history of the field and how key geometrical insights continue to shape it. The organizers anticipate that this workshop will promote awareness of the power of the geometrical perspective advanced by Scharf?s work in shaping the forefront of today?s research. They believe this is a particularly important message to upcoming researchers and practitioners in the field at a time when geometrical ideas beyond linear subspaces are becoming increasingly relevant in signal processing. And holding the workshop in coordination with the Asilomar Conference on Signals, Systems, and Computers, where many Ph.D. students and early-career researchers in statistical signal processing from the western U.S. attend regularly, provides the workshop access to this audience.