Sensor-Based Decision Making in Uncertain Context

Sensor-Based Decision Making in Uncertain Context

Eric Villeneuve, François Pérès, Cedrik Beler, Vicente González-Prida
Copyright: © 2017 |Pages: 24
DOI: 10.4018/978-1-5225-0651-5.ch011
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Decision makers, whether human or computer, using sensor networks to instrument information collecting necessary for decision, often face difficulties in assessing confidence granted to signals transmitted and received in the network. Several organizational (network architecture or nature, distance between sensors ...), internal (sensor reliability or accuracy ...) or external (impact of environment ...) factors can lead to measurement errors (false alarm, non-detection by misinterpretation of the analyzed signals, false-negative …). A system-embedded intelligence is then necessary, to compare the information received for the purpose of decision aiding based on margin of errors converted in confidence intervals. In this chapter, the authors present four complementary approaches to quantify the interpretation of signals exchanged in a network of sensors in the presence of uncertainty.
Chapter Preview
Top

Interest And Background On Multi Sensor Fusion

Multi-sensor fusion can provide more accurate and reliable information than the information given by each sensor individually. In addition, data from multiple heterogeneous sensors have different degrees of uncertainty and confidence. Among the techniques of multi-sensor fusion, Bayesian methods and Theories of evidence such as the theory of Dempster-Shafer (DS), are commonly used to treat the degrees of uncertainty in the merge process. The research carried out takes place within this context. The multi-sensor fusion refers to the combination of sensory data from multiple sensors to provide more accurate and reliable information. The potential benefits of the multi-sensor fusion are redundancy and complementarity of the acquired information. The fusion of redundant information can reduce the overall uncertainty and thus helps to provide more accurate target information. Multiple sensors providing redundant information may also be used to increase reliability in the case of error or failure of the sensors. Additional information from multiple sensors provide environmental characteristics that would be impossible to collect using only each isolated sensor information.

Within the field of science and engineering, data imperfection requires the use of tools to define mechanisms for reasoning with partial knowledge and uncertain information. In the work of Dubois and Prade (2009) several types of imperfections are discerned:

Complete Chapter List

Search this Book:
Reset