Maritime rescue aims to safeguard individuals and assets amidst maritime emergencies, wherein amphibious aircraft play a proactive role by providing reactive and dependable support for rescue operations. The uncertain maritime emergency environment poses significant challenges to amphibious aircraft maritime rescue (AAMR), necessitating an urgent need for a systematic situation assessment to capture and address the threats and risks involved. Current assessment methods, however, suffer from critical deficiencies in addressing multiple uncertainties in situation information and measuring the impacts of interacting threat factors in the maritime emergency environment. To fill these gaps, this paper proposes a novel situation assessment method using probability linguistic hybrid cloud (PLHC) model and best-worst method (BWM) to delineate the optimal situation level of AAMR. Initially, the situation assessment model for AAMR is developed where the threatening factors are identified through literature reviews and empirical analysis. Then, to facilitate reasonable knowledge utilization, the PLHC model, which combines probability linguistic term sets (PLTSs) and hybrid normal and trapezium clouds, is introduced to address experts’ assessment with various uncertainties such as hesitation, and fuzziness. Moreover, the enhanced BWM method is extended to determine the weights of the threatening factors and their mutual interactions. Finally, validated through vessel-sinking emergency scenarios, the proposed method can offer significant situation information for pilots and maritime authorities to facilitate more effective maritime rescue strategies, highlighting its vital role in enhancing maritime emergency response capabilities.

With the increasing complexity and volume of task demands in high-concurrency IoT applications, UAV swarm systems must scale up to meet these requirements, inevitably introduces challenges related to computational efficiency and performance, as well as a lack of theoretical analysis on solution convergence and optimality. To address these issues, this paper proposes a novel optimization model for coalition formation and a hierarchical task allocation method. The approach combines a semi-centralized clustering with distributed coalition formation scheme, where multi-dimensional contribution clustering decomposes tasks and platforms for complexity reduction. Moreover, by modeling sub-cluster allocation as an Overlapping Coalition Formation (OCF) game, our approach integrates marginal utility criteria with search algorithms featuring adaptive resource matching and random exit mechanisms to accelerate the search and avoid suboptimal solutions. Theoretical proof confirms the Nash equilibrium attainment through iterative coalition adjustments while ensuring low complexity. Simulation results show that the method significantly reduces decision-making complexity while ensuring task utility and overall coalition efficiency, demonstrating its effectiveness in UAV swarm-based civilian disaster relief systems.

Attention Deficit Hyperactivity Disorder (ADHD) is a common psychiatric disorder in childhood, and its pathogenesis may be related to abnormalities in brain network connectivity.In this study, we employ minimum spanning tree and hierarchical clustering algorithms to analyze the differences in brain functional network structures between ADHD patients and normal individuals.Initially, we compute Pearson correlation coefficients to construct functional connectivity matrices, delving into the mean matrices for both groups to pinpoint regions marked by significant discrepancies.Subsequently, we construct the minimum spanning trees for each group, assessing their average leaf scores and examining the variations in network topology.In the final phase, we apply hierarchical clustering to the minimum spanning trees of both cohorts, with an analysis of the community structure conducted through homogeneous and heterogeneous slicing.The results show significant differences in functional connectivity networks in the ADHD group, revealing unique features in the neural mechanisms.

The target trajectory prediction method can assist pilots in situation awareness and provide support for decision-making to improve the capacity to gain the advantage during high dynamic within-visual-range air combat. Aiming at the problem of traditional methods based on the Cartesian frame, such as the low utilization of training data and the weak generalization, and the low accuracy of the existing time series prediction models, a target trajectory prediction method in air combat based on Wavelet-Attention-GRU under the Frenet frame is proposed. In this method, the air combat features are described via the spatial trajectory curve based on the Frenet frame; the target trajectory prediction model is established by combining the GRU network with the improved multi-head self-attention mechanism by adding wavelet transform. Finally, the one-to-one within-visual-range air combat dataset obtained via the high-fidelity air combat simulator is applied to train and test the trajectory prediction model.

Clustering is one of the most important methods to discover the intrinsic grouping in a set of unlabeled data. As ways of getting data are more various and easier, the amount of data processed is increasing exponentially and the data is more likely to be located at different clients. Traditional clustering methods cannot process the large dataset one time due to the limit of memories. In this paper, an Image Scaling Density-based Clustering (ISDC) algorithm is proposed. ISDC can process data by a client alone as well as process in parallel by several clients to deal with data located at different clients. The ISDC algorithm does not need any parameters to be designated manually. The parameters are determined by the algorithm based on the statistical features of dataset. In Parallel ISDC or PISDC, each data block located at different client is clustered alone to form intermediate clusters. By border detection algorithm, representative clusters are formed by the points that are at the edge of intermediate clusters. Then, in global clustering, representative clusters from all clients are merged by the server. The border detection algorithm reduces the communication cost between clients and the server, as well as increases the efficiency of global clustering. At last, the server feeds back the clustering information to clients to complete clustering. Our experimental results verified the effectiveness and efficiency of PISDC and ISDC.