Search

Cross‐domain object detection aims to generalize the distribution of features extracted by an object detector from an annotated domain to an unknown and unlabelled domain. Although one‐stage cross‐domain object detectors have significant advantages in deployment than two‐stage ones, they suffer from two problems. First, neglect of category features and inaccurate alignment between multiple category features would lead to decreased domain adaptation efficiency. Second, one‐stage detectors are more sensitive to imbalance of samples and negative samples severely affect the alignment process of domain adaptation. To overcome these two problems, an innovative category‐related attention domain adaptive method that refines discrimination for each category's feature has been proposed in this paper. In the proposed method, a group of domain discriminators is assigned to each category to refine the fine‐grained features between categories. The discriminators are trained via an adversarial discriminant framework to align the fine‐grained distributions cross different domains. A category attention alignment (CAA) module is proposed to navigate more attention to the foreground regions in instance‐level, which effectively alleviates the negative migration problem caused by the positive and negative sample imbalance of the one‐stage detector. Specifically, two sub‐modules in the CAA module are developed: a local CAA module and a global CAA module. These modules aim to optimize the domain offsets in both the local and global dimensions. In addition, a progressive global alignment module is designed to align image‐level features, offering prior knowledge of migration for the CAA module. The progressive global alignment module and CAA module collaboratively engage in benign competition with the backbone network across various levels. Extensive transferring experiments are conducted among cityscapes, foggy cityscapes, SIM10K, and KITTI. Experimental results show that the proposed method has much superior performance than other one‐stage cross‐domain detectors. [ABSTRACT FROM AUTHOR]

The research and development of electromechanical integrated textile machinery generally has the problems of heavy field debugging workload, long cycle and high cost. Aiming at the shortcomings of the current mainstream simulation software, such as poor 3D display effect, insufficient real-time performance, or inability to connect mechanical and electronic software for comprehensive simulation, a set of semi-physical simulation system is designed. The simulation platform is built based on Unity3D virtual reality engine, and the simulation circuit board is used as the carrier to realize the real-time data interaction between the virtual machine and the control system. The lawn carpet automatic repair robot is taken as the simulation object, to prove that the semi-physical system has good real-time performance and 3D display effect. By outputting the motion sequence diagram and space trajectory diagram of the key structure of the repair robot, the process simulation feasibility of the mechanical and electrical system of the repair robot is verified, which can replace part of the field debugging, reduce the workload, shorten the development cycle, save costs, and improve the research and development efficiency of the equipment. [ABSTRACT FROM AUTHOR]

The study of the strain of nano-friction is of great significance to improve our understanding of the friction behavior of two-dimensional (2D) nanomaterials, realize the regulation of tribological properties and support the application of nano-friction in various industries. The electronic structures and the atomic-scale friction behaviors of graphene/ZrS2 heterostructure bilayers systems under the compressive and tensile strain and defect effects are simulated through first-principles methods. The heterogeneous graphene/ZrS2 heterostructure under the 4% and 8% applied strain exhibit almost the similar friction and shear strength as the intrinsic system without strain. However, a significant increase in friction and shear strength under compressive strain conditions. At the condition of the applied strain of − 4%, 0%, 4% and 8%, the maximum values of the potential energy surfaces (PES) are maintained at a very low value of 0.00121–0.00326 eV. In point defect systems, it was found to increase the frictional barrier. And compared to the intrinsic system, the system of the C atoms substituted with S atoms lead to an increase in potential energy of nearly 2–3 orders of magnitude. Applying strain and introducing point defects are important and effective methods to tune the interlayer friction of 2D materials. The low friction is primarily due to the greater electrostatic repulsion caused by the interface of the asymmetric bilayer system and the small change in the charge density throughout the sliding process. [ABSTRACT FROM AUTHOR]