Finding high quality paths within a limited time in configuration space is a challenging issue for path planning. Recently, an asymptotically optimal method called fast marching tree (FMT*) has been proposed. This method converges significantly faster than its state-of-the-art counterparts when addressing a wide range of problems. However, FMT* appears unable to solve the narrow passage problem in optimal path planning, since it is based on uniform sampling. Aiming at solving this problem, a novel region-based sampling method integrating global scenario information and local region information is proposed in this paper. First, global information related to configuration space is extracted from an initial sample set obtained via hybrid sampling. Then, local regions are constructed and local region information is captured to make intelligent decisions regarding regions that are difficult and need to be boosted. Finally, the initial sample set is sent to FMT* using a modified locally optimal onestep connection strategy in order to find an initial and feasible solution. If no solution is found and time permits, the guided hybrid sampling will be adopted in order to add more useful samples to the sample set until a solution is found or the time for doing so runs out. Simulation results for six benchmark scenarios validate that our method can achieve significantly better results than other state-of-theart methods when applied in challenging scenarios with narrow passages.
High‐performance convolutional neural networks (CNNs) stack many convolutional layers to obtain powerful feature extraction capability, which leads to huge storing and computational costs. The authors focus on lightweight models for hyperspectral image (HSI) classification, so a novel lightweight criss‐cross large kernel convolutional neural network (LiteCCLKNet) is proposed. Specifically, a lightweight module containing two 1D convolutions with self‐attention mechanisms in orthogonal directions is presented. By setting large kernels within the 1D convolutional layers, the proposed module can efficiently aggregate long‐range contextual features. In addition, the authors effectively obtain a global receptive field by stacking only two of the proposed modules. Compared with traditional lightweight CNNs, LiteCCLKNet reduces the number of parameters for easy deployment to resource‐limited platforms. Experimental results on three HSI datasets demonstrate that the proposed LiteCCLKNet outperforms the previous lightweight CNNs and has higher storage efficiency.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.