A SYSTEM FOR THE AUTOMATIC SELECTION OF SENSORS

M. J. Korsten, P. E. van der Vet, P. P. L. Regtien

Abstract

In this paper a method is proposed to assist the designer of a measurement system in finding the optimal set of sensors for a given measurement problem. The method will be implemented on a computer system and will contain a framework for optimality, where the user can define, what is optimal in his/her case.<BR />To combat the combinatorial explosion during a search, solutions will be generated on different levels of abstraction. A solution at a high level of abstraction represents a large number of solutions at a lower level. Making choices by cutting possibilities at this high level will therefore enable to go through the search-tree in an efficient way.<BR />Criteria with respect to the performance and costs of the solutions are discussed, together with methods to weigh different criteria against each other.<BR />The user interface should enable the user to formulate his measurement problem. This means in our view that a model must be generated of relevant aspects of the scene, in which the measurement problem occurs. Finding a solution to the measurement problem then means the design of an environment in which a measurement can be done that determines the measurand.

Keywords

measurement science, knowledge-based system, sensor selection

Download

IMEKO-WC-2000-EXP-P527.pdf

DOI

-

Back to the proceedings