Traffic Monitoring Using Strain Measurements and Neural Networks
A strain sensor network is evaluated using artificial neural networks (ANN) to perform traffic monitoring. The approach uses strain profiles to determine the number of vehicles and the weight of each vehicle. In this simulation study, the vehicles travel a constant speed across a three-bay truss bridge. Each vehicle may enter the bridge at independent times and up to four vehicles may travel the bridge. The training and testing data for truss member strain is obtained from a finite element model that incorporates multiple, single-direction rolling loads. Strain profiles from the truss member elements serve as inputs to multi-layered feed-forward ANNs trained using back propagation training algorithms (BPNN). The proposed system can be scaled for a more complex traffic scenario and truss bridge.
R. Dua et al., "Traffic Monitoring Using Strain Measurements and Neural Networks," Intelligent Engineering Systems Through Artificial Neural Networks (ANNIE '06), American Society of Mechanical Engineers (ASME), Jan 2006.
Electrical and Computer Engineering
Article - Conference proceedings
© 2006 American Society of Mechanical Engineers (ASME), All rights reserved.