NICOP - Analysis of real scatter data using the point cloud method

Abstract

In [1] the `point cloud method is proposed for analyzing acoustic scattering data. The scatterer is represented with a (convex or non-convex) hull of a point cloud and the far field data is used to perform Bayesian inference to establish the unknown scatterer. As the points in the cloud move using MCMC the possible hull borders adapt to and the most likely scatterer shapes. The Bayesian approach permits the usage of prior information on the scatterer shapes, and Uncertainty Quantification (UQ) on the possible scatterer forms, given the sensed data. As explained in [1] the point cloud method results in a flexible approach for many types of scatterers. However, only synthetic data was considered in [1]. The present research proposal concentrates on exploring the use of our point cloud method on real, sonar type, data, including methodological and practical additions/extensions, to allow for its possible application in real case scenarios.

Document Details

Document Type
DoD Grant Award
Publication Date
Jul 10, 2018
Source ID
N629091812098

Entities

People

  • Jose Christen Gracia

Organizations

  • CIMAT Center for Mathematical Research
  • Office of Naval Research
  • United States Navy

Tags

Readers

  • Distributed Systems and Data Platform Development
  • Graph Algorithms and Convex Optimization.
  • Radar Systems Engineering.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference