Economic aspects of sensor selection optimization of finite and infinite dimensional dynamical systems

Abstract

The economic aspects as a new factor in the selection of sensors for improved filtering of dynamical systems are introduced. By using the price of a single sensor, reflected by high values of the associated covariance, an economic aspect of the sensor optimization for optimal filtering is introduced. Both the unit price and the total price of a network of inexpensive noisy sensors are used as an alternative to the performance of a single expensive and highly accurate sensor. Algorithms for the integrated sensor optimization for both finite and infinite dimensional systems are presented and examples are provided to demonstrate these effects.