Search

Compliance with traffic laws is a fundamental requirement for human drivers on the road, and autonomous vehicles must adhere to traffic laws as well. However, current autonomous vehicles prioritize safety and collision avoidance primarily in their decision-making and planning, which will lead to misunderstandings and distrust from human drivers and may even result in accidents in mixed traffic flow. Therefore, ensuring the compliance of the autonomous driving decision-making system is essential for ensuring the safety of autonomous driving and promoting the widespread adoption of autonomous driving technology. To this end, the paper proposes a trigger-based layered compliance decision-making framework. This framework utilizes the decision intent at the highest level as a signal to activate an online violation monitor that identifies the type of violation committed by the vehicle. Then, a four-layer architecture for compliance decision-making is employed to generate compliantly trajectories. Using this system, autonomous vehicles can detect and correct potential violations in real-time, thereby enhancing safety and building public confidence in autonomous driving technology. Finally, the proposed method is evaluated on the DJI AD4CHE highway dataset under four typical highway scenarios: speed limit, following distance, overtaking, and lane-changing. The results indicate that the proposed method increases the vehicle's overall compliance rate from 13.85% to 84.46%, while reducing the proportion of active violations to 0%, demonstrating its effectiveness., Comment: 14 pages, 17 figures

In this paper, we introduce PCR-CG: a novel 3D point cloud registration module explicitly embedding the color signals into the geometry representation. Different from previous methods that only use geometry representation, our module is specifically designed to effectively correlate color into geometry for the point cloud registration task. Our key contribution is a 2D-3D cross-modality learning algorithm that embeds the deep features learned from color signals to the geometry representation. With our designed 2D-3D projection module, the pixel features in a square region centered at correspondences perceived from images are effectively correlated with point clouds. In this way, the overlapped regions can be inferred not only from point cloud but also from the texture appearances. Adding color is non-trivial. We compare against a variety of baselines designed for adding color to 3D, such as exhaustively adding per-pixel features or RGB values in an implicit manner. We leverage Predator [25] as the baseline method and incorporate our proposed module onto it. To validate the effectiveness of 2D features, we ablate different 2D pre-trained networks and show a positive correlation between the pre-trained weights and the task performance. Our experimental results indicate a significant improvement of 6.5% registration recall over the baseline method on the 3DLoMatch benchmark. We additionally evaluate our approach on SOTA methods and observe consistent improvements, such as an improvement of 2.4% registration recall over GeoTransformer as well as 3.5% over CoFiNet. Our study reveals a significant advantages of correlating explicit deep color features to the point cloud in the registration task., Comment: accepted to ECCV2022; code at https://github.com/Gardlin/PCR-CG

Deep learning highly relies on the quantity of annotated data. However, the annotations for 3D volumetric medical data require experienced physicians to spend hours or even days for investigation. Self-supervised learning is a potential solution to get rid of the strong requirement of training data by deeply exploiting raw data information. In this paper, we propose a novel self-supervised learning framework for volumetric medical images. Specifically, we propose a context restoration task, i.e., Rubik's cube++, to pre-train 3D neural networks. Different from the existing context-restoration-based approaches, we adopt a volume-wise transformation for context permutation, which encourages network to better exploit the inherent 3D anatomical information of organs. Compared to the strategy of training from scratch, fine-tuning from the Rubik's cube++ pre-trained weight can achieve better performance in various tasks such as pancreas segmentation and brain tissue segmentation. The experimental results show that our self-supervised learning method can significantly improve the accuracy of 3D deep learning networks on volumetric medical datasets without the use of extra data., Comment: Accepted by MICCAI 2020

The acid deposition has been considered to be a severe environmental issue in China. The pH, electrical conductivity (EC), and the concentrations of the water soluble ions (NO3−, Cl−, Ca2+, K+, F−, NH4+, Mg2+, SO42−, and Na+) in the precipitation samples collected from the 320 cities during 2011–2016 across the whole China were measured. The mean concentrations of F−, NO3− and SO42− were in the order of winter (6.10, 19.44 and 45.74 μeq/L) > spring (3.45, 13.83, and 42.61 μeq/L) > autumn (2.67, 9.73, and 28.85 μeq/L) > summer (2.04, 7.66, and 19.26 μeq/L). The secondary ions (SO42−, NO3− and NH4+), and F− peaked in Yangtze River Delta (YRD) and Sichuan basin (SB). The crustal ions (i.e., Ca2+, Mg2+), Na+, and Cl− showed the highest concentrations in the semi-arid regions and the coastal cities, respectively. The statistical methods confirmed that the mean anthropogenic contribution ratios to SO42−, F−, NO3−, and NH4+ at a national scale were 46.12 %, 71.02 %, 79.10 %, and 82.40 %, respectively. However, Mg2+ (70.51 %), K+ (77.44 %), and Ca2+ (82.17 %) were mostly originated from the crustal source. Both Na+ (70.54 %) and Cl− (60.42 %) were closely linked to the sea-salt aerosols. On the basis of the stepwise regression (SR) analysis, it was proposed that most of the secondary ions and F− were closely related to gross industrial production (GIP), total energy consumption (TEC), vehicle ownership, and N fertilizer use, but the crustal ions (Ca2+ and K+) were mainly controlled by the dust events. The influence of dust days, air temperature, and wind speed on ions increased from Southeast China (SEC) to Central China, and then to Northwest China (NWC), whereas the influence of socioeconomic factors on acid ions (SO42− and NO3−) displayed the higher value in East China.

Taking an asymmetric single-tower cable-stayed bridge as the research object, this paper investigates the effects of seismic mitigation measures for the bridge under the earthquake. Three kinds of damping measures including lead-rubber bearings, viscous dampers and a hybrid seismic mitigation system are suggested. Firstly, parametric analyses have been performed so as to obtain the optimal parameters of the lead-rubber bearings and the damping coefficients of the viscous dampers for the bridge. Then, the hybrid seismic mitigation system combined by lead-rubber bearings and viscous dampers are considered. The effects of these seismic mitigation measures on the bridge have been compared and the results show that these three measures can effectively reduce the seismic responses of the bridge, but the hybrid seismic mitigation system has a better performance than the other two.

In order to further improve the efficiency of flapping wings, inspired by bird flight, a new compliant joint structure is proposed, which is easy to manufacture. The joint has the capacity of passive bending, twisting and sweeping under the aerodynamic effect of the flapping wing. A multi-objective optimization program for the compliant joint is developed, which is combined with the measured joint load. The optimized compliant joint has the effect of asymmetric passive deformation in a flapping cycle. The model wings with compliant joints are studied in the windless flapping experiment, wind tunnel experiment and free flight test. The experimental results show that the model wings have a significant advantage of power consumption compared with the original wings, and also have an advantage in the aerodynamic performance at low flight velocity and low angle of attack.

By reviewing a series of achievements that professional degree education has obtained in China, the author think that China's professional degree education has entered into a new period of strategic transition for comprehensively improving the quality as the core task. By the end of 2012 to early 2013,the Research Center for Graduate Education of Beijing Institute of Technology did an investigation on students’ satisfaction with postgraduate education. Results revealed that the quality of the professional degree education needs to move forward a single step. Improving the curriculum rationality, Changing teaching methods, Adding practice base construction and strengthening the construction of the tutor team are the new horizon for future reform.

According to statistics from 2011 to 2013, in China, approximately 60,000 people were killed in road traffic accidents every year. And about 4% of these accidents have vehicle safety issues, such as brake failure, tire burst. To investigate association between vehicle safety performance and traffic accident casualty risk, 1302 traffic accident samples with detailed information were collected and transformed into data. Base on logistic regression analysis, the quantitative function relationship between vehicle status and driver injury severity was built. This research method represents an important step in the methodological evaluation of vehicle data with respect to traffic safety.

Research on WLAN-3GPP interworking is regarded as a case of implementation of heterogeneous networks which are the tendency of future mobile communication systems. The paper describes a WLAN-3GPP interworking system, describes the issues existing in current networking selection procedures and analyses the disadvantages and advantages of the current solutions in detail. The efficiency of network selection determines the performance of system accessing, especially handover. Finally, two schemes are proposed to improve the procedure to meet the requirements of seamless handover when performing vertical handover between WLAN and UMTS.