IntelliSensorNet: A Positioning Technique Integrating Wireless Sensor Networks and Artificial Neural Networks for Critical Construction Resource Tracking

Abstract

The increasing needs for safety and productivity improvement in the field of construction engineering and project management have stimulated research interests in developing cost-effective resource tracking and positioning solutions for challenging indoor or partially covered site environments. This thesis has proposed a robust positioning architecture called IntelliSensorNet that relies on an integrated environment of Wireless Sensor Networks and Artificial Neural Networks for construction resource localization. The wireless sensor network (WSN) based component of the architecture determines the location of mobile sensor nodes (“tags”) by evaluating radio signal strengths (RSS) received by stationary sensor nodes (“pegs”). Only a limited quantity of reference points with known locations and pre-calibrated RSS in relation to the pegs are used to determine the most likely coordinates of a tag. Moreover, to effectively reduce uncertainty and improve accuracy, an on-line error correction approach based on a Radial Basis Function Neural Network (RBF NN) model is embedded in the proposed architecture. In short, this localization technique produces a cost-effective solution to positioning and tracking critical construction resources such as laborers and equipment for challenging indoor environments or partially covered site environments in construction, thus lending itself well to potential deployment in real-world construction sites.