With the rapid increase of high-speed train driving automation, the human-computer interaction of high-speed train driving is changing. It is a major dilemma to maintain a well-interactive process between the driver and the driving system for the intelligent train operation. Nevertheless, the enhancement of the driver’s situation awareness (SA) is an effective way to promote the process of driver-driving system interaction. To achieve this purpose, we proposed an enhanced SA model based on the perceptual cycle model (PCM). The model helps to identify the high-speed train DMI interface elements that affect the different SA level of drivers and provides theoretical support for describing the SA change pattern of driving tasks. We proposed a new display mode (ESA mode) for enhancing driver SA according to this model, and conducted driving simulator experiments to explore the differences of SA change patterns between the ESA mode and the normal mode (NSA mode) for 16 subjects. The experimental results showed that the ESA mode successfully overcame the subjects’ SA recession in the NSA mode. The interaction with enhanced SA can help the driver remain vigilant and further improve the general driving performance during monotonous process of automated high-speed train driving.

Solar-tracking is one of the key issues in underwater polarization navigation. Most existing studies are based on the Angle of Underwater Polarization (AoUP). As water depth increases, however, the AoUP is disturbed by multiple scattering, while the Degree of Underwater Polarization (DoUP) maintains a stable relative relationship. To address the less robustness of polarization navigation, a solar-tracking and integrated orientation method based on DoUP is proposed. In consideration of the refraction, the relationship between the solar position and DoUP is established based on the maximum Degree of Polarization. Moreover, an integrated navigation model is developed aided by the solar position. Results from the static experiment in water tank and dynamic sea trials demonstrate that the proposed method exhibits improved accuracy and robustness compared to the AoUP method. This article presents a potential approach to improve the adaptability of underwater autonomous orientation.

The formation of underwater images is a complex physical process that often suffers from various degradation factors, such as blurriness, low contrast, and color casts, which pose challenges for underwater object detection and recognition tasks. Because of the absence of reference images, learning-based methods that rely on unpaired images have been employed to enhance the underwater images. However, these methods may lose their effectiveness in real-world complex underwater environments. In this paper, we propose a model-driven cycle-consistent generative adversarial network (CycleGAN) model, which is inspired by the underwater image formation model to estimate the background light, transmission map, scene depth, and attenuation coefficient directly. Comprehensive experiments have demonstrated that our approach surpasses the compared underwater image restoration methods in both qualitative and quantitative aspects, providing restored images with satisfactory color saturation and brightness. We also conduct experiments on underwater object detection to illustrate the effectiveness of our CycleGAN in improving the detection accuracy. All our source codes and data are available at https://github.com/Duanlab123/UW-CycleGAN.

MOTIVATION: With the rapid development of precision medicine, a large amount of health data (such as electronic health records, gene sequencing, medical images, etc.) has been produced. It encourages more and more interest in data-driven insight discovery from these data. A reasonable way to verify the derived insights is by checking evidence from biomedical literature. However, manual verification is inefficient and not scalable. Therefore, an intelligent technique is necessary to solve this problem. RESULTS: This article introduces a framework for biomedical evidence engineering, addressing this problem more effectively. The framework consists of a biomedical literature retrieval module and an evidence extraction module. The retrieval module ensembles several methods and achieves state-of-the-art performance in biomedical literature retrieval. A BERT-based evidence extraction model is proposed to extract evidence from literature in response to queries. Moreover, we create a dataset with 1 million examples of biomedical evidence, 10 000 of which are manually annotated. AVAILABILITY AND IMPLEMENTATION: Datasets are available at https://github.com/SendongZhao.

This research proposed a robust Cooperative Adaptive Cruise Control (CACC) to overcome the shortcoming of the existing CACC controllers in dealing with unexpected events, such as malware attack, phishing attack, and DNS tunneling attack, where perception and data received via communication contradict with the reality. The proposed controller combines the advantage of two information flow topologies – All-Predecessor Following (APF) and Predecessor-Leader Following (PLF) control methods – to improve the capability of CACC platoons. The string stability of the proposed CACC controller was proven. The robustness to time delay switch (TDS) attacks was assessed using simulation. The normal cruise was simulated to show the capability of the proposed controller in a wide range of speeds. TDS attacks, which encompassed four different unexpected events, were tested to verify the robustness of the proposed CACC. Results confirmed that the proposed CACC controller is string stable and robust against these TDS attacks without crashing or causing jerks. It also showed that the proposed CACC controller outperformed a state-of-the-art CACC controller.

In the process of intelligent driving technology development, the performance of human–machine systems for high-speed train driving tasks has been scrutinized and challenged. The human cognitive process can be divided into several stages. Within each of these stages, automation can be applied across different levels. However, the impact of different automation levels at different cognitive stages on high-speed train driving performance is still unclear. In this study, different level of automation (LOA) of high-speed train driving functions at different cognitive stages are designed, and 30 participants are recruited to execute orthogonal Latin square experiments. The experimental results show that driving automation in the decision-making stage significantly improved performance in a speed control task but also increased the mental workload and reaction time for monitoring train signals. However, this effect did not occur in a track obstacle observation response time. These results show that driving automation not only improves the performance of primary driving tasks but also has a negative impact on secondary driving-related tasks. This study evaluates the influence of high-speed train driving automation on the performance of human–machine systems and analyzes the advantages and costs of different LOA at different cognitive stages. It provides a theoretical basis for the design of intelligent high-speed train driving systems from the perspective of human–machine collaboration.