Boltryk, P.J., Harris, C.J. and White, N.M.,
Prosser, S. J., Yan, Y. and Lewis, E.(eds.)
Intelligent sensors—a generic software approach
Journal of Physics: Conference Series, 15, (1), . (doi:10.1088/1742-6596/15/1/026).
Full text not available from this repository.
Designating a sensor as intelligent is a long-standing term implying that it provides more functionality than merely providing an output measurement. Since there is some discrepancy governing what makes a given sensor intelligent, this paper defines the features required for improving confidence in sensor measurements, from the sensor management perspective. We describe a software framework used to implement tasks such as condition monitoring onboard the sensor itself, rather than at the traditional supervisory level. The algorithms include data-based models, which allows for modelling of non-linear effects and estimation uncertainty, which is a prerequisite for data fusion. Density estimation for novelty detection is demonstrated for an accelerometer that is purposely damaged in an environmental chamber.
Actions (login required)