Search

The implementation of Convolutional Neural Networks (CNNs) and Transformers has significantly accelerated the booming advances of human pose estimation. However, challenges persist in accurately estimating target details, and performing pose estimation in complex environments where targets may be blurred or chaotic. In this paper, we exploit TokenPose as the baseline, revisiting the design of CNNs and Transformers. TokenPose utilizes CNNs and Transformers sequentially, combining the advantages of CNNs in extracting low level features and Transformers in establishing long-range dependencies. On this basis, we modify the CNNs and Transformer related structures to further enhance expression ability of the model. The modification is made to the original CNNs section: we alter BasicBlock to AtBlock to maintain high-resolution information exchange to further excavate details of objects. The two following modifications are applied to the subsequent Transformer section: (1) we replace the self-attention layer in each encoder block with the Local Enhancement Self-Attention(LESA) module to capture local information. (2) we propose a Local Perception Feed-Forward Network(LPFFN) to substitute for the feed-froward network layer in each encoder block, which employs the depth-wise convolution to enhance the correlation of neighbour information in the spatial dimension. Our modifications contribute to analyzing poses that are arduous to estimate due to occlusion. Our method combines self-attention and depth-wise convolution, which is named CSDNet. We demonstrate a better performance over the TokenPose baseline on both the COCO2017 and MPII datasets. Additionally, in complicated environments such as poor lighting conditions, our method can more accurately estimate fuzzy keypoints.

Uncontrollable lithium (Li) dendrite growth has hampered the application of lithium metal batteries (LMBs). Herein, a novel plastic crystal composite polymer electrolyte (nh-CPE) consisting of multifunctional two-dimensional (2D) molybdenum disulfide (MoS2) and ion-conducting one-dimensional (1D) surfactant oxidized cellulose nanocrystal (OCNC) is developed to overcome this challenge. The nh-CPE possesses good Li-dendrite-suppressing ability, which is attributed to the well-designed three-dimensional (3D) ion conducting network constructed by homogeneously dispersed MoS2 and OCNC. This structure facilitates Li-ion conductivity and enables uniform Li deposition on the anode. Moreover, the impermeable MoS2 nano-flakes with excellent mechanical strength serve as physical barriers to hold the electrolyte-electrode interface stability and consequently block Li dendrite growth and enhance fire retardancy. When applied in LMB, the nh-CPE exhibits excellent cycle performance and rate capability, with 90% capacity retention after 200 cycles at 0.5 C and 60 °C, and a discharge capacity of 128 mAh g−1 at 1 C. Notably, owing to high room-temperature ionic conductivity (0.8 mS cm−1), the cell with nh-CPE cycled at room temperature delivers an initial discharge capacity of 144 mAh g−1. The outstanding performance and simple process offer great opportunity to promote Li-dendrite-suppressing LMBs for practical application.

The existing comparative analysis of the unbalanced moment mainly compares the differences between the calculated value and the experimental value of the unbalanced moment in different national codes and different calculation methods. There is less analysis on the change trend of the important parameters in the unbalanced moment calculation methods in national codes. A comparison of the provisions for unbalanced moment of reinforced concrete slab-column connections was studied based on Chinese code (GB 50010-2010), American code (ACI318-14) and European code (EN1992-04). The effects of concrete strength, longitudinal reinforcement ratio, the use of shear reinforcement and gravity-shear ratio on the unbalanced moment were studied through the calculation example and experimental results of 186 specimens. The results indicate that the calculation values of Chinese and American codes are conservative and dispersive, because the longitudinal reinforcement ratio and gravity-shear ratio on distribution coefficient of unbalanced moment were not considered, while the calculation results of European code are the closest to the test values.

A new structure of self-centering buckling restrained brace was proposed. The finite element model was established by ABAQUS software. The mechanical failure mechanism of this kind of self-centering buckling restrained brace was studied. The influence of the parameters such as the area of self-centering reinforcement and the sectional area of core tube on its hysteretic performance and self-centering ability was analyzed. The results show that with the increase of the area of the reset reinforcement, the hysteretic curve of the brace is more pinched, and the self-centering ability is increased by 46%; it was suggested that the weakening degree of the cross-sectional area of the core tube should not exceed 50%.

In order to study the main influencing factors of the ultimate bearing capacity of the PBL connector, the large-scale nonlinear finite element analysis software ABAQUS was used to establish the solid model of the PBL connector, and the accuracy of the model was verified by comparison with the test results. Based on this, the main influencing factors of the ultimate bearing capacity of PBL connectors are analyzed. The results show that it is feasible to apply ABAQUS to the finite element analysis of the mechanical properties of PBL connectors, which can be assisted by experiments and empirical formulas; The concrete strength, the diameter of rebar and the number of holes have a significant effect on PBL connectors’ force, but the diameter of holes has little effect.