Sensor Fusion and Civil Infrastructure Systems Monitoring: A Valuation Algebras Analysis of Output Data

This paper proposes the use of valuation algebra as a tool for sensor fusion in civil infrastructural monitoring. The valuation algebra provides a framework for representing knowledge of sensor data and drawing inferences under uncertainty. The solution is based on Shenoy-Shafer architecture and uses binary join tree and local computational method for the calculation of the marginals. The use of binary tree reduces the amount of multiplication needed to calculate the marginals. The approach is Illustrated using an example for pavement condition monitoring.