Abstract With the advantages of flexibility and mobility, unmanned aerial vehicles (UAVs) have been widely used in the wireless rechargeable sensor networks (WRSNs) to collect data and supply energy for ground sensor nodes. Due to the limited battery capacity of UAVs and the continuity requirement of WRSN, mobile unmanned vehicles (MUVs) are introduced as mobile charging stations to ensure the energy supply for UAVs and mitigate energy wastage. This paper investigates the problem of Joint Optimization Mission Allocation and Cooperative Trajectory Planning for data collection in WRSNs. The goal is to maximize the minimum energy efficiency by optimizing mission allocation including UAV trajectory and MUV travel. This problem is proved to be NP-hard and solved by two proposed algorithms. The first algorithm incorporates the clustering utilizing the K-Means algorithm and genetic algorithm. The second algorithm is a self-attention architecture based on the reinforcement learning framework and formulate an actor-critic algorithm for training. The simulation results show the feasibility and efficiency of the proposed algorithms, which achieve better performance. The first algorithm has more advantages when the distribution of sensor nodes is relatively concentrated; and the second algorithm may be more suitable when more comprehensive global path planning optimization is required.

Invasive Alien Species (IAS) pose a significant threat to global biodiversity and economies, necessitating accurate and efficient identification for effective management. Traditional methods are often constrained by time, cost, and expertise while existing unimodal deep learning approaches struggle with species phenotypic diversity, environmental complexity, and long-tailed data distributions. This study introduces InvaderDefender, a novel “Visual Recognition-Semantic Enhancement” two-stage multimodal fusion framework to address these challenges. The first stage utilizes an EfficientNetV2-L visual backbone, enhanced with a novel wavelet fusion module for fine-grained feature extraction, and a combined strategy of class-aware data resampling and gradient-aware focal loss to address long-tailed distributions. The second stage inputs the Top-k probability outputs from the visual model, along with user-provided text, into the GLM-4 LLM enhanced by our proposed Vision-Guided Targeted Retrieval Augmented Generation (VGT-RAG). This GLM-4 leverages a specially constructed expert knowledge base for 20 target IAS to perform in-depth inference and output the final identification. To support this research, the first multimodal benchmark dataset specifically for these IAS was constructed. InvaderDefender achieves a Top-1 accuracy of 99.31%, significantly outperforming mainstream unimodal models. Compared to the visual-only baseline, the full multimodal framework boosts Top-1 and Mean Accuracy by 0.90 and 2.15 percentage points, respectively, ensuring a more balanced performance across all species, particularly rare ones. This work establishes an efficient and robust framework for IAS identification, providing a novel methodology for integrating multimodal learning and knowledge-driven AI in biodiversity monitoring.

This paper investigates the framework of wireless rechargeable sensor network (WRSN) assisted by multiple mobile unmanned vehicles (MUVs) and laser-charged unmanned aerial vehicles (UAVs). On the basis of framework, we cooperatively investigate the trajectory optimization of multi-UAVs and multi-MUVs for charging WRSN (TOUM) problem, whose goal aims at designing the optimal travel plan of UAVs and MUVs cooperatively to charge WRSN such that the remaining energy of each sensor in WRSN is greater than or equal to the threshold and the time consumption of UAV that takes the most time of all UAVs is minimized. The TOUM problem is proved NP-hard. To solve the TOUM problem, we first investigate the multiple UAVs-based TSP (MUTSP) problem to balance the charging tasks assigned to every UAV. Then, based on the MUTSP problem, we propose the TOUM algorithm (TOUMA) to design the detailed travel plan of UAVs and MUVs. We also present an algorithm named TOUM-DQN to make intelligent decisions about the travel plan of UAVs and MUVs by extracting valuable information from the network. The effectiveness of proposed algorithms is verified through extensive simulation experiments. The results demonstrate that the TOUMA algorithm outperforms the solar charging method, the base station charging method, and the TOUM-DQN algorithm in terms of time efficiency. Simultaneously, the experimental results show that the execution time of TOUM-DQN algorithm is significantly lower than TOUMA algorithm.

Unmanned Aerial Vehicle (UAV) can be used as wireless aerial mobile base station for collecting data from sensors in UAV-based Wireless Sensor Networks (WSNs), which is crucial for providing seamless services and improving the performance in the next generation wireless networks. However, since the UAV are powered by batteries with limited energy capacity, the UAV can not complete data collection tasks of all sensors without energy replenishment when a large number of sensors are deployed over large monitoring areas. To overcome this problem, we study the Real-time Data Collection with Laser-charging UAV (RDCL) problem, where the UAV is utilized to collect data from a specified WSN and is recharged using Laser Beam Directors (LBDs). This problem aims to collect all sensory data from the WSN and transport it to the base station by optimizing the flight trajectory of UAV such that real-time data performance is ensured It has been proven that the RDCL problem is NP-hard. To address this, we initially focus on studying two sub-problems, the Trajectory Optimization of UAV for Data Collection (TODC) problem and the Charging Trajectory Optimization of UAV (CTO) problem, whose objectives are to find the optimal flight plans of UAV in the data collection areas and charging areas, respectively. Then we propose an approximation algorithm to solve each of them with the constant factor. Subsequently, we present an approximation algorithm that utilizes the solutions obtained from TODC and CTO problems to address the RDCL problem. Finally, the proposed algorithm is verified by extensive simulations.

Wireless Rechargeable Sensor Networks (WRSNs) have been widely utilized and have played an important role in many surveillance application scenarios. The optimization of the charging process is beneficial for guaranteeing continuous coverage and enhancing the charging efficiency of WRSNs. And there are several influence factors of the charging process, like the sensors’ battery consumption mode, the chargers’ charging pattern and the environmental factors, which should be considered into the charging model. Based on the charging model via assigning sensors’ charging priority weights, we introduce the spatio-temporal optimization for charging scheduling (STO-CS) Problem in WRSNs for the goals of meeting the on-demand charging requirements and saving the charging consumption. We prove the NP-hardness of the problem and propose two algorithms to solve it. The first algorithm is based on two-phase dynamic programming and is proved to find the optimal solution when the charging ability is sufficient; the second algorithm adopts the clustering idea with K-Means Algorithm which has better time complexity. A series of simulation experiments are performed to compare the performance of the proposed algorithms in terms of the charging cost and the running time, whose results are analyzed to conclude that they can be applied to the application scenarios with the accuracy requirements and the real-time requirements respectively.

The development of wireless energy transmission technology has significantly propelled the advancement of wireless rechargeable sensor networks (WRSNs). Energy constraint is one of the most critical challenges in application of WRSNs. Integrating unmanned aerial vehicle (UAV) with wireless energy transmission technology has emerged as a promising approach to overcome the energy constraint problem in WRSNs, leveraging the advantages of UAV such as flexibility and maneuverability. In this paper, we consider the system of WRSN assisted by UAV and mobile utility vehicle (MUV), where the UAV serves as a mobile charger for replenishing energy of sensors and the MUV serves as a mobile base station for replacing the battery of UAV with insufficient energy. In the system, we focus on minimizing the death time of sensors and optimizing the energy consumption of UAV. To address this problem, a multi-objective deep Q-network (DQN) algorithm is employed, where the UAV makes online charging scheduling decisions based on real-time network status and utilizes experience replay for optimization. Experimental results demonstrate that the proposed algorithm significantly reduces the sensors’ death time and effectively decreases the energy consumption of UAV. Specially, the performance of proposed algorithm outperforms the three other classical algorithms: genetic algorithm, greedy algorithm, and Q-learning algorithm.

Edge computing is a computational paradigm that brings resources closer to the network edge, such as base stations or gateways, in order to provide quick and efficient computing services for mobile devices while relieving pressure on the core network. However, the current computing power of edge servers are insufficient to handle the high number of tasks generated by access devices. Additionally, some mobile devices may not fully utilize their computing resources. To maximize the use of resources, we propose a novel edge computing system architecture consisting of a resource-constrained edge server and three computing groups. Tasks from each group can be offloaded to either the edge server or the corresponding computing group for execution. We focus on optimizing the computation offloading of devices to minimize the maximum overall task processing latency in the system. This problem is proved to be NP-hard. To solve it, we propose a DQN-based resource utilization task scheduling (DQNRTS) algorithm that has two desirable characteristics: 1) it effectively utilizes the computing resources in the system and 2) it uses deep reinforcement learning to make intelligent scheduling decisions based on system state information. Experimental results demonstrate that the DQNRTS algorithm is capable of reducing the processing latency of the system by converging to optimal solutions.

In this study, the authors focus on the task of leaf segmentation under different imaging conditions (e.g. backgrounds and shadows). A new method ‐ robust random walk (RW) is proposed to propagate the prior of user's specified pixels. Specifically, they first employ RWs to take the relationship of pairwise pixels into consideration. A superpixel‐consistent constraint is added to make the edges of segmentation smooth. Owing to the effect of illumination, some parts of a leaf surface are brighter than others and it may further harm the subsequent label propagation. To address this problem, they learn a common subspace by taking into account the illumination of local and non‐local pixels. By doing so, it has good adaptability to process noise interfering and non‐uniform illumination. In addition, since RW only considers the pairwise relationship of pixels, it will be sensitive to the specified and connected pixels. Thus, they further employ a log‐likelihood ratio to predict the probability of a pixel belonging to the background and use it to guide the label propagation. Based on the proposed method, they can obtain a smoothed and robust leaf segmentation. Experimental results on unconstrained leaf images demonstrate the efficiency of their algorithm.

Plant leaf recognition is a computer vision task used to automatically recognize plant species. It is very challenging since rich plant leaf morphological variations, such as sizes, textures, shapes, venation, and so on. Most existing plant leaf methods typically normalize all plant leaf images to the same size and recognize them at one scale, resulting in unsatisfactory performances. In this letter, a multiscale fusion convolutional neural network (MSF-CNN) is proposed for plant leaf recognition at multiple scales. First, an input image is down-sampled into multiples low resolution images with a list of bilinear interpolation operations. Then, these input images with different scales are step-by-step fed into the MSF-CNN architecture to learn discriminative features at different depths. At this stage, the feature fusion between two different scales is realized by a concatenation operation, which concatenates feature maps learned on different scale images from a channel view. Along with the depth of the MSF-CNN, multiscale images are progressively handled and the corresponding features are fused. Third, the last layer of the MSF-CNN aggregates all discriminative information to obtain the final feature for predicting the plant species of the input image. Experiments show the proposed MSF-CNN method is superior to multiple state-of-the art plant leaf recognition methods on the MalayaKew Leaf dataset and the LeafSnap Plant Leaf dataset.