To tackle the challenges of path planning for unmanned aerial vehicle (UAV) in complex environments, a global–local balanced whale optimization algorithm (GLBWOA) has been developed. Initially, to prevent the population from prematurely converging, a bubble net attack enhancement strategy is incorporated, and mutation operations are introduced at different stages of the algorithm to mitigate early convergence. Additionally, a failure parameter test mutation mechanism is integrated, along with a predefined termination rule to avoid excessive computation. The algorithm’s convergence is accelerated through mutation operations, further optimizing performance. Moreover, a random gradient-assisted optimization approach is applied, where the negative gradient direction is identified during each iteration, and an appropriate step size is selected to enhance the algorithm’s exploration capability toward finding the optimal solution. The performance of GLBWOA is benchmarked against several other algorithms, including SCA, BWO, BOA, and WOA, using the IEEE CEC2017 test functions. The results indicate that the GLBWOA outperforms other algorithms. Path-planning simulations are also conducted across four benchmark scenarios of varying complexity, revealing that the proposed algorithm achieves the lowest average total cost for flight path planning and exhibits high convergence accuracy, thus validating its reliability and superiority.

Text-image cross-model matching is a core challenge in multimodal machine learning, aiming to enable efficient retrieval of images and texts across different modalities. The difficulty in this task stems from the inherent gap between text and image representations, which can lead to suboptimal retrieval performance. Traditional approaches attempt to learn a shared representation space where both image and text can be directly compared. However, they often fail to account for the varying levels of semantic information captured in different layers of the encoders, resulting in inadequate alignment between the modalities. To address these limitations, we propose a novel approach called Progressive Multi-Subspace Fusion, dubbed PMSF for text-image matching. Our model reduces the model gap by using a progressive learning process, starting with shallow representations and moving to deeper layers. We use a dual-tower structure to encode multi-level features for both image and text, which are then mapped to corresponding auxiliary subspaces. These subspaces are fused through an adaptive GPO pooling strategy, enabling joint learning of a shared representation space. Experimental results on benchmark datasets, including Flickr30K and MSCOCO, show that PMSF significantly improves retrieval performance, achieving a Rsum score of 516.9 and 510.7, outperforming 23 state-of-the-art methods.

Proactive beam management is essential for high-mobility millimeter-wave communication systems, where transmission latency may cause beams selected from current observations to become mismatched with the optimal beams at future transmission times. This paper investigates radar-only causal multi-horizon proactive beam prediction, aiming to directly rank future communication beams from past radar observations. To this end, we propose Shared-Beam-GCN-TCN, a multi-task framework that learns a shared predictive sensing-communication representation and performs structured decoding in the beam space. Specifically, four consecutive range-angle radar maps are encoded by a frame-wise convolutional neural network and a causal temporal convolutional network to capture short-term motion dynamics. The shared temporal representation is jointly optimized for future beam prediction and short-term radar prediction, encouraging the extraction of predictive sensing dynamics relevant to future communication states. A beam-space graph convolutional decoder then exploits the adjacency structure among 64 beam classes to model correlations between neighboring beams and improve future beam ranking. The framework is evaluated in two high-mobility mmWave communication scenarios, including a trajectory-disjoint DeepSense 6G vehicular scenario and a simulated unmanned aerial vehicle (UAV) scenario. In the vehicular scenario, Shared-Beam-GCN-TCN achieves the best overall performance across different prediction horizons, improving the average Top-3 and Top-5 accuracies by 8.2 and 6.9 percentage points over Seq-CNN-TCN, while reducing the mean beam-index error. In the UAV scenario, the proposed framework achieves an accuracy of 93.4% and a beam-index error of 1.022. These results demonstrate the effectiveness of shared predictive representation learning and structured beam-space decoding for causal multi-horizon beam prediction in high-mobility environments.

Ultra-high frequency (UHF) signal detection is one of the most effective methods for spatial orientation and partial discharge fault diagnosis. However, when the background noise, especially Gaussian white noise and narrow-band interference, is very high, or the detection region is remote, location accuracy may decrease. To improve this, a location system based on the improved higher-order cumulant time delay estimation (TDE) method combined with energy accumulation is proposed. In the system, the UHF waves are received by a four-antenna array. Improved fourth-order cumulants with a smoothed coherence transform (SCOT) window are used in the TDE, by which Gaussian white noise and narrow-band interference can be efficiently suppressed. The energy accumulation algorithm is applied to the cross-correlation results, by which the accuracy of TDE can be enhanced. The applicability of the proposed localization algorithm is evaluated by simulation and experiment. The simulation results show that the improved fourth-order cumulant TDE-parameter accumulation algorithm is superior in accuracy to traditional location methods. In the experimental measurements, two partial discharge measurement points located in a complex electromagnetic environment are tested. The results illustrate that the proposed method can effectively suppress Gaussian white and narrow-band noise, and the location results can satisfy accuracy requirements when the measurement point is within 70 m.

This article first elaborates on the characteristics and classification of big data for maritime law enforcement, and analyzes the current application status of big data in the field of maritime law enforcement. Subsequently, based on the research situation of big data application in the marine field both domestically and internationally, discussions were conducted on the direction of big data application towards maritime law enforcement. Finally, in response to the shortcomings in the application of big data for maritime law enforcement, and drawing on existing research experience at home and abroad, a method based on the Apriori algorithm for data mining of maritime illegal activities is proposed.