No abstract

Weigh-in-motion technology is an effective tool that has been extensively used to monitor traffic on highways. Pavement-based weighin-motion systems usually have poor durability and will cause traffic interruption during their installation and maintenance process. The recently developed bridge weigh-in-motion technology provides a more convenient and cost-effective alternative to the pavement-based weigh-in-motion technology. Bridge weigh-in-motion systems can be installed without interrupting the traffic. Also, bridge weigh-in-motion systems have the potential to deliver better accuracy than pavement-based weigh-in-motion systems. Due to these significant advantages, the bridge weigh-in-motion technology has been playing an increasingly important role in bridge health monitoring and overweight truck enforcement, and many studies have been conducted to continuously improve the bridge weigh-inmotion technology. In this review, the common algorithms for bridge weigh-in-motion are discussed in detail, and the typical instrumentation of bridge weigh-in-motion systems is also introduced. Meanwhile, much effort is made to identify the remaining issues in the application of bridge weigh-in-motion technology, and the corresponding future research is proposed.

Out-of-domain (OOD) detection for lowresource text classification is a realistic but understudied task. The goal is to detect the OOD cases with limited in-domain (ID) training data, since we observe that training data is often insufficient in machine learning applications. In this work, we propose an OODresistant Prototypical Network to tackle this zero-shot OOD detection and few-shot ID classification task. Evaluation on real-world datasets show that the proposed solution outperforms state-of-the-art methods in zero-shot OOD detection task, while maintaining a competitive performance on ID classification task.

No abstract

Bridge weigh-in-motion (BWIM) technique uses an instrumented bridge as a weighing scale to estimate vehicle weights. Traditional BWIM systems use axle detectors placed on the road surface to identify vehicle axles. However, the axle detectors have poor durability due to the direct exposure to the traffic. To resolve this issue, a free-of-axle-detector (FAD) algorithm, which eliminates the use of axle detectors, was proposed. As a further improvement to simplify the BWIM systems, the concept of nothing-on-road (NOR) BWIM was recently introduced. The axle identification method proposed in this paper is an attempt to achieve the NOR BWIM, i.e., using bridge global responses to identify vehicle axles. Wavelet analysis is applied to extract the axle information from the global responses. This allows the BWIM technique to be achieved with only weighing sensors. Numerical simulations are conducted using three-dimensional vehicle and bridge models and the effect of several parameters, including sampling frequency, road surface condition and measurement noise on the identification accuracy is investigated. The results demonstrate that the proposed identification method using wavelet analysis can accurately identify vehicle axles, except for cases where the road surface condition is rough or measurement noises exceed certain levels.