Small defects on the rails develop fast under the continuous load of passing trains, and this may lead to train derailment and other disasters. In recent years, many types of wireless sensor systems have been developed for rail defect detection. However, there has been a lack of comprehensive reviews on the working principles, functions, and trade-offs of these wireless sensor systems. Therefore, we provide in this paper a systematic review of recent studies on wireless sensor-based rail defect detection systems from three different perspectives: sensing principles, wireless networks, and power supply. We analyzed and compared six sensing methods to discuss their detection accuracy, detectable types of defects, and their detection efficiency. For wireless networks, we analyzed and compared their application scenarios, the advantages and disadvantages of different network topologies, and the capabilities of different transmission media. From the perspective of power supply, we analyzed and compared different power supply modules in terms of installation and energy harvesting methods, and the amount of energy they can supply. Finally, we offered three suggestions that may inspire the future development of wireless sensor-based rail defect detection systems.

Hydrogen has a wide range of energy applications, but hydrogen energy systems can suffer from high-concentration leaks that pose security risks, therefore making the measurement of high-concentration hydrogen very important. Traditional ultrasonic gas-detection methods are based mainly on ultrasonic time-of-flight measurements and can be divided into threshold-detection and phase-difference techniques. Threshold detection suffers from low resolution and a complex structure in gas detection, while the phase-difference technique has high resolution and a simple structure but can only measure the time of flight within one period of the ultrasonic signal. In this study, a dual-frequency phase-difference technique is proposed that solves the problem of multi-period phase detection with the phase-difference technique and can be used to detect high-concentration hydrogen. Simulation analysis and an experiment show that the proposed technique can measure the multi-period phase difference accurately. The maximum hydrogen concentration can reach 50% with an uncertainty of less than 5%, which meets commercial requirements.

To accurately recognize ordinary handwritten Chinese characters, it is necessary to recognize the normative level of these characters. This study proposes methods to quantitatively evaluate and recognize these characters based on their similarities. Three different types of similarities, including correlation coefficient, pixel coincidence degree, and cosine similarity, are calculated between handwritten and printed Song typeface Chinese characters. Eight features are derived from the similarities and used to verify the evaluation performance and an artificial neural network is used to recognize the character content. The results demonstrate that our proposed methods deliver satisfactory evaluation effectiveness and recognition accuracy (up to 98%~100%). This indicates that it is possible to improve the accuracy in recognition of ordinary handwritten Chinese characters by evaluating the normative level of these characters and standardizing writing actions in advance. Our study can offer some enlightenment for developing methods for the identification of handwritten Chinese characters used in transaction processing activities.

No abstract

In this paper, a top-k pseudo labeling method for semi-supervised self-learning is proposed. Pseudo labeling is a key technology in semi-supervised self-learning. Briefly, the quality of the pseudo label generated largely determined the convergence of the neural network and the accuracy obtained. In this paper, the authors use a method called top-k pseudo labeling to generate pseudo label during the training of semi-supervised neural network model. The proposed labeling method helps a lot in learning features from unlabeled data. The proposed method is easy to implement and only relies on the neural network prediction and hyper-parameter k. The experiment results show that the proposed method works well with semi-supervised learning on CIFAR-10 and CIFAR-100 datasets. Also, a variant of top-k labeling for supervised learning named top-k regulation is proposed. The experiment results show that various models can achieve higher accuracy on test set when trained with top-k regulation.