Search

In this paper we consider the application of spatial estimation techniques to a groundwater aquifer and geological borehole data. We investigate the adequacy of these techniques to reliably develop contour maps from various data sets.

We have applied state-space methods to the analysis and measurement of the dynamic response of mechanical structures. Our basic approach has been to use finite element structural dynamics models and convert the geometric description of the structure to produce a set of 2nd order matrix differential equations. The equations are then transformed into state-space form for input to a large-scale discrete-time linear Kalman filter. The filter has been implemented on the CRAY-1 computer which has sufficient capacity for a system of 500 states.

The response of an object subjected to high energy, transient electromagnetic (EM) fields called an electromagnetic pulse (EMP) is an important issue in the survivability of electronic systems (e.g., aircraft), especially when the field has been generated by a high altitude nuclear burst. The characterization of transient reponse information is a matter of national concern. In this paper we apply identification techniques to : (i) improve signal processing at a test facility, and (ii) parameterize a particular object response.

The enclosed space detection problem is concerned with the detection of an individual or individuals present in an area that is not accessible visually but only through sensors deployed about the perimeter under investigation. Devices developed to detect these individuals using only sensor information are called enclosed space detectors. In this paper we develop a technique to enhance signal estimates which will eventually be used to solve the enclosed space detection problem. We develop a model-based approach that leads to a multichannel noise cancelling/deconvolution solution.