Abstract
Accurately estimating the weight of a moving vehicle at normal speed remains a challenging problem due to the complex vehicle dynamics and vehicle–pavement interaction. The weighing technique based on multiple sensors has proven to be an effective approach to this task. To improve the accuracy of weigh-in-motion (WIM) systems, this paper proposes a neural network-based method integrating identification and predication. A backpropagation neural network for signal classification (BPNN-i) was designed to identify ideal samples acquired by load sensors closest to the tire-pavement contact area. After that, ideal samples were used to predict the gross vehicle weight by using another backpropagation neural network (BPNN-e). The dataset for training and evaluation was collected from a multiple-sensor WIM (MS-WIM) system deployed in a public road. In our experiments, 96.89% of samples in the test set had an estimation error of less than 5%.
Highlights
Overloading has always been a major concern for road traffic [1]
We presented a neural-network-based method to estimate gross vehicle weight for an multiple-sensor WIM (MS-WIM) system
The main contribution of this work is the integration of ideal sample identification into weight estimation
Summary
Overloading has always been a major concern for road traffic [1]. To date, several weighing techniques have been used for overweight vehicle detection [2,3]. Sensors in a WIM system measure dynamic load [6,7] rather than the static weight due to vehicle dynamics [8,9]. In the past few decades, the studies on MS-WIM systems have mainly focused on two aspects: sensor layout and vehicle weight estimation algorithms The former concerns how to arrange sensors to precisely measure the dynamic response along the travelling direction. The mean value of the Gaussian distribution is approximately linear with the dynamic load applied by the vehicle, and occurs at the centerline of the wheel path [29,30] Inspired by these studies, we propose a novel gross vehicle weight estimation method for MS-WIM systems. The proposed method is implemented for an MS-WIM system deployed on a public road, and the data obtained from real road traffic are used to train and test our identification network and prediction network
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
