This paper presents an inertial system for railway track diagnostics. The key element of the system is a set of inertial measurement units (IMUs) based on MEMS gyros and accelerometers, which are mounted directly on the axle boxes (bearing covers) of the wheel pairs. The system made it possible to investigate how the car-track dynamic interaction affects measurements of geometrical deformations and to determine parameters, such as defects of rail treads.
The paper presents the results of development of the Optical-Inertial System for Railway Track Diagnostics. It is demonstrated that in order to implement the solution at a speed of up to 430 kmph (used for example in South Korean high-speed train HEMU-430X, standing for High-Speed Electric Multiple Unit 430 km/h experimental) while satisfying the accuracy of 0.1…0.5 mm during measurement of longitudinal level, cross level, twist, curvature, rail profile, etc., it is needed to combine the optical scanners of the inner profile of the rail line with the strapdown inertial navigation system (SINS) in a single block. Supplying of odometer and Global navigation satellite system receiver (GNSS) into the system structure allows to determine measurement point position.Thanks to our a priori knowledge of the semipermanent nature of the railway track, and also to the fusion of the odometer data and satellite navigation system reception equipment data, it is possible to use fiber-optic gyros as the sensitive units of the SINS (both open-loop and closed-loop configurations of FOG can be used).The distinctive feature of the system's algorithm is that it solves both the navigation/orientation task (i.e. it fuses odometer data, satellite navigation system data and inertial navigation system data), and the task of measuring the inner surface profile of the rail line.The use of a sole odometer to localize the found rail flaws does not provide satisfactory results because of its errors. Integration of the odometer, SINS and GNSS receiver data offers highly accurate referencing of diagnostic results to the traversed track coordinate.Odometer readings are updated using the navigation system data. The system provides measuring of the track geometry and accurate localization of the measurement point using the geographical coordinates (latitude and longitude) and orientation parameters (roll, pitch and course angle).The possibility of using SINS based on fiber-optic gyros (FOG) for railway applications is considered in the article. Some practical results are given.
Active learning (AL) is a prominent technique for reducing the annotation effort required for training machine learning models. Deep learning offers a solution for several essential obstacles to deploying AL in practice but introduces many others. One of such problems is the excessive computational resources required to train an acquisition model and estimate its uncertainty on instances in the unlabeled pool. We propose two techniques that tackle this issue for text classification and tagging tasks, offering a substantial reduction of AL iteration duration and the computational overhead introduced by deep acquisition models in AL. We also demonstrate that our algorithm that leverages pseudo-labeling and distilled models overcomes one of the essential obstacles revealed previously in the literature. Namely, it was shown that due to differences between an acquisition model used to select instances during AL and a successor model trained on the labeled data, the benefits of AL can diminish. We show that our algorithm, despite using a smaller and faster acquisition model, is capable of training a more expressive successor model with higher performance. 1
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.