To significantly enhance the performance of point cloud semantic segmentation, this manuscript presents a novel method for constructing large-scale networks and offers an effective lightweighting technique. First, a latent point feature processing (LPFP) module is utilized to interconnect base networks such as PointNet++ and Point Transformer. This intermediate module serves both as a feature information transfer and a ground truth supervision function. Furthermore, in order to alleviate the increase in computational costs brought by constructing large-scale networks and better adapt to the demand for terminal deployment, a novel point cloud lightweighting method for semantic segmentation network (PCLN) is proposed to compress the network by transferring multidimensional feature information of large-scale networks. Specifically, at different stages of the large-scale network, the structure and attention information of the point features are selectively transferred to guide the compressed network to train in the direction of the large-scale network. This paper also solves the problem of representing global structure information of large-scale point clouds through feature sampling and aggregation. Extensive experiments on public datasets and real-world data demonstrate that the proposed method can significantly improve the performance of different base networks and outperform the state-of-the-art.

In this letter, a three-dimensional finite-difference time-domain (FDTD) method that is second-order in time and fourth-order in space on a face-centered cubic (FCC) grid, termed as FCC-FDTD(2,4), is proposed to solve the time-domain Maxwell's equations. The proposed method shows a much lower numerical dispersion error compared with the errors of other FDTD schemes, such as the FCC-FDTD(2,2) method, the FDTD(2,2) method, and the FDTD(2,4) method. This improvement in the numerical dispersion results from the FCC grid and the fourth-order scheme in space used in the proposed FCC-FDTD(2,4) method. Then, the Courant-Friedrichs-Lewy condition is given through spectral analysis, and an analytical numerical dispersion relation is obtained and comprehensively studied through numerical results with the FCC-FDTD(2,2) method, the FDTD(2,2) method, and the FDTD(2,4) method. Compared with the other three FDTD methods, the proposed method has much smaller numerical dispersion errors. Finally, the accuracy of the proposed method is further validated with other numerical examples.

To enhance the scalability and performance of the traditional finite-difference time-domain (FDTD) methods, a three-dimensional summation-by-parts simultaneous approximation term (SBP-SAT) FDTD method is developed to solve complex electromagnetic problems. It is theoretically stable and can be further used for multiple mesh blocks with different mesh sizes. This paper mainly focuses on the fundamental theoretical aspects upon its three-dimensional implementation, the SAT for various boundary conditions, and the numerical dispersion properties and the comparison with the FDTD method. The proposed SBP-SAT FDTD method inherits all the merits of the FDTD method, which is matrix-free, easy to implement, and has the same level of accuracy with a negligible overhead of runtime and memory usage. Four numerical examples are carried out to validate the effectiveness of the proposed method.

A finite-difference time-domain (FDTD) subgridding method is proposed to efficiently and accurately solve three-dimensional electromagnetic problems. Based on the reciprocal and symmetric interpolation operators between coarse and subgridding meshes, spatial coupling matrices are carefully designed to guarantee long-time stability. To further enhance its capability of handling multiscale structures, arbitrary grid ratios and nested subgridding meshes are extended to be supported in the proposed method. In addition, the rigorous analysis shows that the proposed FDTD subgridding method is theoretically stable. Five numerical examples including a simple cavity with perfect electric conductors, a rectangular TEM waveguide, a dielectric resonator, a single-layer substrate integrated waveguide, and a large airplane platform with a dipole are carried out to validate its effectiveness. Results show that it is stable, accurate, efficient, and easy to model complex structures.

Frequency-modulated continuous-wave (FMCW) radar plays a pivotal role in the field of remote sensing. In low-altitude environments, where UAV-borne FMCW radars are increasingly used, mutual interference poses significant challenges to radar performance. In this paper, a novel FMCW radar interference mitigation (RIM) method, termed as RIMformer, is proposed by using an end-to-end Transformer-based structure. In the RIMformer, a dual multi-head self-attention mechanism is proposed to capture the correlations among the distinct distance elements of intermediate frequency (IF) signals. Additionally, an improved convolutional block is integrated to harness the power of convolution for extracting local features. The architecture is designed to process time-domain IF signals in an end-to-end manner, thereby avoiding the need for additional manual data processing steps. The improved decoder structure ensures the parallelization of the network to increase its computational efficiency. Simulation and measurement experiments are carried out to validate the accuracy and effectiveness of the proposed method. Extensive simulations and empirical measurements demonstrate that RIMformer significantly improves interference mitigation and signal recovery, which advances the reliability and effectiveness of UAV-borne FMCW radars in complex environments.

The finite-difference time-domain (FDTD) method based on nonuniform face-centered cubic (FCC) grids, termed as NU-FCC-FDTD method, is proposed in this article. The FCC-FDTD method with uniform grids has a smaller numerical dispersion error than the traditional FDTD methods. To further enhance its efficiency and scalability, the NU-FCC-FDTD method is proposed to solve complex electromagnetic problems. Unlike the traditional FDTD methods based on Yee’s grids, for which the nonuniform grid technique is relatively straightforward, instability may occur when trying to apply nonuniform grids into the FCC-FDTD method. To ensure stability, cell-size-based correction factors are carefully defined in the time-marching formulations after taking cell sizes into consideration. Through comprehensive numerical dispersion analysis, the NU-FCC-FDTD method shows numerical dispersion advantage over the FDTD methods and the FCC-FDTD method with uniform grids. Furthermore, the stability condition is also given by using the parameter scanning method and verified in numerical examples. Four examples, including electromagnetic wave propagation, a dielectric rod (DR) resonator, a composite metallic-dielectric radome and the electromagnetic energy absorption of human head, are carried out to demonstrate its performance. Results show that the proposed NU-FCC-FDTD method can significantly improve efficiency compared with that of the FCC-FDTD method without compromising accuracy.