He Zhichao , You Tao , Yuan Chengce , Li Yingchun , Liu Li , Huang Yudong
2026, 58(8):1-9. DOI: 10.11918/202512118
Abstract:Carbon fiber reinforced polymer composites are increasingly utilized in critical fields such as aerospace and ship building. As a key interface material, the sizing agent is crucial for protecting fibers, bundling filaments, and optimizing the interfacial performance between fibers and resin matrices. Although the Chinese carbon fiber industry has made significant progress, core technologies for high-performance sizing agents still lag behind international counterparts, becoming a “bottleneck” problem restricting the independent control of the industrial chain. This paper summarized two core mechanisms by which sizing agents for carbon fibers enhance the interfacial performance of composites, providing a theoretical basis for the design and preparation of sizing agents. The characteristics of solvent-based, emulsion-type, and aqueous sizing agents were summarized. The differences in composition, preparation methods, interfacial performance, environmental protection, and cost among different sizing agents were compared. It is considered that eco-friendly and low-cost aqueous sizing agents represent the future direction, and their current research status is reviewed in detail. Meanwhile, the challenges currently faced by aqueous sizing agents for carbon fibers are pointed out, including the prevalent stability issues and the lack of specialized high-end products suitable for high-performance thermoplastic resin systems. Finally, the future research directions are prospected, aiming to provide theoretical references and technical ideas for the independent research and development of high-performance sizing agents for carbon fibers in China.
Zhao Huihao , Zhang Yanxiang , Yan Mufu
2026, 58(8):10-26. DOI: 10.11918/202604032
Abstract:Friction and wear are key factors causing component failure, accuracy degradation, and energy loss during service. By regulating the chemical composition, phase constitution, defect structure, and mechanical gradient of the material surface layer, the capabilities of wear resistance, friction reduction, fatigue resistance, and corrosion resistance can be significantly improved while maintaining the comprehensive properties of the substrate. Focusing on five typical surface modification technologies including carburizing, nitriding, carbonitriding, ion implantation, and diamond-like carbon (DLC) coatings, this paper systematically summarized the research progress in improving the friction and wear properties of materials and compared the effects of different technologies on surface microstructural evolution, hardness distribution, friction behaviors, and transitions of wear mechanisms. Existing studies indicate that the above technologies can improve surface load-bearing and damage-resistance capabilities through surface hardening, gradient support, residual stress regulation, interfacial chemical state optimization, and transfer film stabilization, usually reducing the wear rate by 1-2 orders of magnitude and promoting the transition of wear mechanisms from severe adhesive wear, abrasive wear, and plastic deformation to oxidative wear, mild abrasive wear, fatigue wear, or low-damage sliding controlled by stable transfer films. Different technologies exhibit distinct characteristics in modification depth, load adaptability, and engineering applicability: carburizing is applicable to heavy-load deep strengthening; nitriding has advantages of low temperature and low deformation; carbonitriding balances high surface hardness and subsurface support; ion implantation is suitable for near-surface functional design of high-precision components; DLC coatings perform outstandingly in low friction and micron-scale surface functionalization. Finally, this paper prospected the development directions of hybrid modification, service-oriented design, and multi-technology synergy.
Duan Jiao , Shi Jingchen , Qiao Yuchen , Yang Xuemei
2026, 58(8):27-41. DOI: 10.11918/202603056
Abstract:Under the current background of “carbon neutrality”, using light energy to drive the catalytic reduction of carbon dioxide (CO2) is one of the important approaches to achieve green and sustainable development. Metal-organic frameworks (MOFs) are a class of crystalline porous materials composed of metal ions/clusters and bidentate or polydentate organic ligands, which possess high structural stability, structural tunability, and convenience of post-modification. By modifying metal nodes and organic ligands or compositing MOFs with other active materials, their properties can be specifically and conveniently regulated to meet the requirements of CO2 photoreduction. Based on the structural characteristics and modification strategies of photocatalytic MOF materials, this review primarily summarized the latest progress of MOF photocatalyst materials used for CO2 reduction and elaborated on different MOF modification methods, such as the modification of metal sites and organic ligands, the construction of heterostructures, the introduction of single/dual atoms, and the encapsulation of quantum dots; it emphatically discussed the influencing mechanisms of band structure, photogenerated charge separation behavior, and reaction microenvironment in pure MOF systems and finally prospected the development directions for future research of MOF-based photocatalytic materials.
Wang Jing , Li Peng , Li Zhen , Lin Jiayi , Zhou Huihui , Xing Defeng
2026, 58(8):42-54. DOI: 10.11918/202601017
Abstract:Green hydrogen is an effective pathway to achieve Chinas carbon neutrality goals and promote energy transition. Biohydrogen production, with its advantages in synergistic pollutant treatment, environmental friendliness, and energy recovery, shows broad application prospects in this field. Compared with other biohydrogen production technologies, dark fermentation has attracted much attention due to its high hydrogen production rate, broad substrate utilization range, and simple operating conditions. Among these, ethanol-type fermentation has become a research focus and hotspot owing to its high substrate utilization efficiency, low pH tolerance, and synergistic ethanol-hydrogen production. This paper systematically elucidated the ethanol-hydrogen co-metabolic pathway of the representative ethanol-type fermentative bacterium Ethanoligenens harbinense, and revealed its molecular response mechanisms under traditional stressors (pH fluctuation, product accumulation) and emerging pollutant microplastics exposure, deepening the understanding of the adaptive mechanisms of acidogenic fermentative microorganisms. Accordingly, this work proposed dark fermentation enhancement strategies from the perspectives of microbial interaction regulation and electron transfer enhancement, and further clarified the mechanism and application prospects of microbial electrochemical technology in enhancing biohydrogen production. From the application perspective, a multi-source organic solid waste cascade biohydrogen production system oriented towards the “dual carbon” strategy was constructed, achieving synergy among organic waste treatment, resource/energy cascade recovery, and carbon emission reduction, while identifying future research directions centered on techno-economic analysis and life cycle environmental benefit assessment. This work systematically elaborated the development pathway of biohydrogen production technology, providing theoretical and technical guidance for its industrial application.
He Shengyuan , Yu Shuili , Li Lei
2026, 58(8):55-66. DOI: 10.11918/202601128
Abstract:Eutrophication-induced frequent phytoplankton blooms in inland waters have become a global water environment challenge. Relying on macroscopic and rapid advantages, remote sensing technology provides spatiotemporally continuous observation methods for the precise prevention and control of algal bloom outbreaks. This paper reviewed the research progress of remote sensing monitoring and early warning for phytoplankton in inland waters, systematically summarized the technical systems ranging from algal bloom extraction, pigment concentration retrieval, and taxonomic identification to one-dimensional time series prediction and spatiotemporal distribution prediction of algal blooms, and focused on discussing the application advantages of data-driven artificial intelligence models in the above fields, as well as the collaborative monitoring models of multi-source data and multiple platforms. The results show that artificial intelligence methods and multi-source collaborative observations significantly improve the monitoring and early warning accuracy in complex optical environments and to a certain extent compensate for the shortcomings of single sensors, such as insufficient spatiotemporal resolution and susceptibility to cloud occlusion; the deep coupling of long-term remote sensing observations and data-driven methods promotes the development of one-dimensional time series prediction and spatiotemporal distribution prediction. However, existing models still face challenges such as high risks of overfitting with small samples and weak physical interpretability. Future research should further deepen multi-source data synergy, explore technologies such as data augmentation to break through the sample dependence bottleneck, and develop a process-data dual-driven model integrating algal bloom evolution mechanisms to improve robustness and interpretability.
Wang Rui , Liu Jinjin , Qiu Shengjie , Zhang Qiong , Li Xiyao , Peng Yongzhen
2026, 58(8):67-78. DOI: 10.11918/202603093
Abstract:To investigate the feasibility of enhancing advanced nitrogen removal from low C/N municipal wastewater using a fermentation mixture of excess sludge at ambient temperature (15-22 ℃), this paper analyzed the effects of fermentation performance at different temperatures (30±2 ℃, 22±2 ℃, and 15±2 ℃), changes in nitrifying bacteria activity, and the addition of sludge fermentation mixture on the nitrogen removal performance of municipal wastewater through batch and long-term experiments. Batch experiments indicate that although a decrease in temperature inhibits cellular hydrolysis, the ratios of soluble chemical oxygen demand to ammonia nitrogen at 15±2 ℃ and 22±2 ℃ (19 and 20) are higher than that at 30±2 ℃ (17), which is more favorable for nitrogen removal. Meanwhile, the inhibitory effects on nitrite-oxidizing bacteria are significantly higher than those on ammonia-oxidizing bacteria during the fermentation process and after the addition of products at different temperatures, thereby promoting the realization of partial nitrification. Long-term experimental results show that after the addition of the sludge fermentation mixture at ambient temperature, the removal efficiency of total inorganic nitrogen in the municipal wastewater nitrogen removal system increases from 77% to 93%, and the effluent total inorganic nitrogen concentration decreases from 14 to 4 mg/L; the nitrite accumulation rate at the end of the aerobic stage increases from 34% and stabilizes at 51%, and the sludge reduction rate reaches 35%. Economic analysis indicates that compared with maintaining fermentation at a mesophilic temperature of 30±2 ℃, fermentation at ambient temperature can save approximately 37% of the cost. Therefore, the addition of the sludge fermentation mixture at ambient temperature can not only significantly improve the nitrogen removal effect of municipal wastewater and achieve sludge reduction but also possess high economic efficiency.
Zhang Rongzheng , Tao Zheng , Wang Yong , Zhang Junqiu
2026, 58(8):79-87. DOI: 10.11918/202601037
Abstract:The midcourse phase of ballistic missile flight is an important stage for early warning, detection, recognition, and interception. After deployment, heavy conical targets such as warheads typically exhibit complex precession motion, a combination of spin and coning, due to unavoidable lateral torque interference. To address the problems that the angular velocity vector is incompletely observable under monostatic and bistatic radar conditions and that existing parameter estimation algorithms heavily rely on observation geometry and are sensitive to environmental noise, this paper proposed an estimation method for the angular velocity vector of precessing targets based on multi-static radar micro-motion curves. First, by constructing a precessing cone motion model and a multi-static radar observation model, this paper extracted the micro-motion curves of scattering points from the high-resolution range profile sequences of multiple radars. Second, by deeply mining the micro-motion laws of scattering points, this paper established a geometric constraint relationship between the radial displacement of scattering points and the coning angular velocity and realized the precise calculation of the coning angular velocity vector using linear constraints from multiple viewing angles. Finally, combined with the precession period characteristics exhibited by the micro-motion curves, the magnitude and vector of the angular velocity were derived and calculated. Numerical simulations and reconstruction experiments demonstrate that, compared with traditional methods, the proposed method can achieve stable estimation of the angular velocity vector under low signal-to-noise ratio conditions, exhibiting superior anti-noise robustness. Furthermore, the micro-motion curves reconstructed using the estimated parameters effectively verify the physical consistency of the results.
2026, 58(8):88-98. DOI: 10.11918/202601063
Abstract:Soil and water loss induced by subsurface structure leakage is an important trigger for pavement collapse accidents, while traditional numerical methods have limitations in simulating large-deformation problems such as engineering-scale quick sand erosion. Therefore, this paper introduced the material point method and adopted a single-point two-phase coupling algorithm. Firstly, this paper verified the method through comparison with two model test results and then established an engineering-scale quick sand leakage model in shallow-buried pipelines to systematically analyze the influence laws of geometric parameters (opening width, opening angle, and pipeline’s burial depth), soil parameters (internal friction angle and cohesion), and hydraulic conditions (groundwater table depth) on leakage evolution. The results indicate that the soil disturbed zone presents three-stage evolution characteristics, including a rapid expansion stage, a quasi-stationary stage, and a stable growth stage; the plastic shear band extends upward in a “crescent shape” and penetrates the surface, forming a typical funnel-shaped settlement trough; the widening of the pipeline opening, the lateral shift of position, and the increase in burial depth all exacerbate soil loss and surface settlement; soil strength parameters can effectively retard leakage, among which the increase of internal friction angle can narrow the settlement trough, and cohesion has a critical threshold, beyond which the soil arching effect can be triggered to enable the quick sand to self-stabilize and stagnate; a high groundwater table reduces soil’s effective stress and enhances seepage drag force, thereby accelerating soil and water loss and broadening the settlement influence scope.
Chen Yuanhanmin , Song Mengjie , Shi Han , Wang Hongjin , Luo Longxi , Zhang Santai
2026, 58(8):99-112. DOI: 10.11918/202601078
Abstract:Autonomous vehicles are important transport tools in future transportation and on battlefields, and pavement ice accretion seriously threatens their driving stability. To detect the distribution characteristics of pavement ice accretion in low-temperature environments, this paper acquired infrared thermal images of ice accretion under the influence of multiple parameters, adopted an image segmentation algorithm and the technique for order preference by similarity to ideal solution to extract and process experimental data respectively, and determined the numerical mapping relationship between the recognition color threshold and accuracy according to visible light images. It selected the optimal threshold accordingly and proposed and verified a detection technology for the distribution characteristics of pavement ice accretion based on an infrared-visible dual-light image segmentation algorithm. The research results show that unlike the shape factor, the optimal color threshold for recognition is affected by the embedding depth and thickness of the ice, ambient temperature, and the emissivity of the contrast materials; when the embedding depth increases from 0 mm to 20 mm; the thickness increases from 10 mm to 30 mm, and the ambient temperature decreases from -0.98 ℃ to -4.71 ℃, the threshold decreases from 0.20 to 0.04, decreases from 0.18 to 0.05, and increases from 0.10 to 0.25, respectively; to quantitatively measure the distribution characteristics of ice accretion, this paper proposed a detection accuracy comprehensively constructed based on three indicators: effective uniformity ratio, distortion frequency, and effective average deviation, and the detection accuracy based on the optimal color threshold for ice recognition reaches 0.81-1. This paper can provide an underlying algorithm and test data reference for onboard infrared thermal imaging ice-detecting sensing technology and make up for the shortcomings that existing visible light image segmentation methods are limited by illumination and difficult to identify black ice and that infrared sensing methods have fixed installation positions and are unable to identify pavement debris.
Wang Yunan , Luo Jiong , Liu Yuhuan , Hua Yongzhao , Li Xiaoduo , Dong Xiwang
2026, 58(8):113-124. DOI: 10.11918/202510107
Abstract:To solve the problems that front-end path search based on a fixed resolution in hierarchical planning cannot balance efficiency and accuracy, and back-end spatiotemporal joint optimization with multiple constraints has high computational complexity and is difficult to solve, this paper proposed a hierarchical motion planning framework based on an adaptive A*-minimum control (AAM) planner. First, this framework consisted of a front-end adaptive search module and a back-end spatiotemporal joint optimization module and operated cooperatively in a receding horizon manner to achieve full-process planning from geometric path generation to dynamically feasible trajectories. Second, the adaptive A* algorithm designed in the front-end part could dynamically adjust the search resolution according to the environmental structure complexity, thereby maintaining high efficiency in open areas and ensuring high precision in narrow areas. The back-end part constructed a spatiotemporal joint optimization model based on MINCO parameterization, integrating multiple objectives such as trajectory smoothness, time efficiency, dynamic feasibility, environmental safety, and swarm cooperative collision avoidance into a differentiable cost function. Finally, this paper adopted the limited-memory broyden-fletcher-goldfarb-shanno (L-BFGS) algorithm to achieve an efficient solution. Comparative experiments in multiple complex simulation scenarios such as pole arrays, narrow gaps, and long walls indicate that the proposed AAM planner outperforms the baseline methods in planning success rate, computational efficiency, and trajectory quality; under the premise of ensuring a high planning success rate and trajectory smoothness, its average planning efficiency increases by 18.4%, and the total trajectory cost decreases by 15.3% on average.
Wang Wenfei , Ru Le , Lü Maolong , Mo Li
2026, 58(8):125-140. DOI: 10.11918/202511098
Abstract:To address the urgent demand for the autonomous decision-making capability of unmanned aerial vehicles in future intelligent air combat and to solve the bottlenecks of low decision-making efficiency and poor strategy robustness of traditional methods in complex and highly dynamic environments, this paper proposed a novel framework for the autonomous decision-making of unmanned aerial vehicles in air combat combining hierarchical reinforcement learning and a two-phase training paradigm. Firstly, this paper designed a hierarchical decision-making architecture to strategically decouple the vast end-to-end decision space. The upper layer utilized a deep Q-network (DQN) to learn high-level tactical decisions, compressing the original continuous action space into a finite set of discrete tactical commands, which effectively reduced the learning complexity; the lower layer mapped the high-level tactical intentions into specific flight control command sequences in real time through multi-mode control and state-driven mechanisms, which ensured the executability and physical safety of the decisions. Secondly, a two-phase training method of “expert imitation and self-game evolution” was proposed and implemented. In the first phase, an expert opponent based on the SECA rule architecture was constructed, and the agent obtained high-quality initial policies and prior knowledge through imitation learning by fighting against it; in the second phase, the agent transitioned into a virtual self-game environment to continuously fight against its historically optimal versions, which drove the policy to continuously break through the limitations of initial expert knowledge. Finally, a high-fidelity air combat simulation environment was built to experimentally verify the performance of the proposed method. The research results indicate that, compared with the traditional DQN and other baseline algorithms, the proposed method effectively improves the reward convergence performance during the agent learning and training and the final confrontation winning rate and effectively enhances the air combat confrontation level of the unmanned aerial vehicle agent.
Luo Shaojie , Wang Le , Qin Mingxing , Xi Jianxiang
2026, 58(8):141-153. DOI: 10.11918/202511100
Abstract:To address the cooperative target engagement problem of underactuated leader-follower missile swarms, this paper proposed a nonsingular multi-missile time-cooperative guidance law with adjustable impact time under terminal angle constraints. By independently designing the prescribed-time guidance command of the leader missile to realize the fast convergence of its impact time error and designing the prescribed-time guidance commands of the follower missiles to make them rapidly track the leader missile, the adjustable impact time of the missile swarm was realized. First, the impact-time and angle control guidance law was designed for the leader missile to realize its terminal impact time and angle control. Next, in the design of the follower missile guidance law, the leader-follower tracking error term and the follower local cooperative error term were introduced into the guidance command, and a time-cooperative guidance law suitable for directed communication topologies among follower missiles under angle constraints was proposed. Then, the auxiliary function and the additional control term guidance command were designed to solve the singularity problem of the guidance law. The failure phenomenon of the impact time control command caused by singularity was analyzed; the criterion for eliminating the complete singularity phenomenon of the guidance law was given, and the prescribed-time stability of the proposed guidance law was proved. Finally, numerical simulations were conducted to verify the effectiveness of the proposed method. The research results indicate that the proposed guidance law enables the impact time of the missile swarm to converge to the desired value within the prescribed time, achieves the effect of adjustable impact time, realizes the spatiotemporal cooperative engagement of the target, and has no singularity phenomenon in the guidance commands.
Ma Baoquan , Fan Wenlong , Yang Yuchen , Han Yuzhe , Hao Bowen , Wu Xiaowen , Wang Tianzhu
2026, 58(8):154-163. DOI: 10.11918/202511105
Abstract:To solve the problem of insufficient small object detection performance of existing YOLO series single-stage detection algorithms caused by severe object scale variation, complex backgrounds, and a high proportion of small objects in unmanned aerial vehicle aerial images, this paper proposed an unmanned aerial vehicle object detection method fusing hybrid pooling and local attention mechanism, named HP-LS-YOLO. Based on the YOLOv11 framework, this method first designed a spatial pyramid hybrid pooling fast (SPHPF) strategy to introduce multi-receptive field contextual information at a single scale by parallelly fusing max pooling and average pooling features and enhanced the feature representation capability under complex backgrounds through convolution-based adaptive filtering. Secondly, a C2PLSA module integrating wavelet convolution and local attention mechanism was constructed, in which the wavelet transform attention effectively enlarged the receptive field while keeping the model lightweight and introduced large-scale scene semantics around small objects. The local self-attention block adopted a Top-K sparse attention mechanism to highlight the local edge and texture features of small objects, while suppressing redundant background interference. Finally, comparative validation experiments were carried out on multiple public datasets to comprehensively verify the enhancement effect of the proposed improvement strategies on dense small object detection in unmanned aerial vehicle aerial images by comparing with the baseline model and similar lightweight detection algorithms. The results indicate that compared with YOLOv11, the proposed method improves APval50∶95, APval50, APvalS on the validation set of the MSCOCO dataset by 0.9%, 0.9%, 1.4%, respectively. Additional experiments on VisDrone and UAVDT datasets further verify the robustness of this method in complex unmanned aerial vehicle scenarios. The proposed HP-LS-YOLO can effectively improve the detection performance of dense small objects in unmanned aerial vehicle aerial images while keeping the model lightweight.
Ma Baoquan , Wang Yunhan , Yu Yinhu , Hao Bowen , Li Jiawei , Yang Yang
2026, 58(8):164-172. DOI: 10.11918/202511097
Abstract:To address the loss of geometric information of distant objects caused by the sparsity of light detection and ranging (LiDAR) point clouds and the problems of projection noise interference and non-standardized feature interaction in existing modal fusion methods, this paper proposed an end-to-end multi-modal three-dimensional detection network, VAF-Net. First, this paper constructed a noise-resistant virtual point enhancement framework, which utilized images to generate virtual points to densify the point clouds and employed VirConvNet as the backbone network to extract global geometric features while effectively filtering the noise introduced by the virtual points. Second, this paper designed a color semantic feature extraction branch to precisely extract local textures and contextual semantics of images aligned with the three-dimensional space through a semantic lifting module. Finally, this paper proposed a bidirectional cross-attention fusion module based on standard Transformer principles. Through a symmetric dual-stream architecture and a gating mechanism, it achieved deep, stable, and bidirectional interaction between LiDAR geometric features and image semantic features. Experiments on the KITTI three-dimensional object detection benchmark indicate that VAF-Net achieves state-of-the-art performance. Particularly in the detection of hard samples, its 3D average precision reaches 82.74%, significantly outperforming mainstream methods such as Voxel-RCNN. The research results demonstrate that VAF-Net effectively realizes the complementary enhancement of multi-modal information through the standardized bidirectional fusion architecture and significantly improves the detection robustness in complex scenarios.
Gao Zhiyun , Tang Yichen , Jin Shuwei , Liang Yuling
2026, 58(8):173-182. DOI: 10.11918/202510031
Abstract:To address the cumbersome reward design of existing decision-making methods in complex scenarios of multi-unmanned aerial vehicle game confrontation and to improve corresponding training stability and sample efficiency, this paper proposed a Critic and Meta enhanced multi-agent deep deterministic policy gradient (CM-MADDPG) algorithm by combining multi-agent reinforcement learning with Meta-reinforcement learning. First, this paper introduced appropriate structural and learning enhancements to the centralized Critic side to improve the value estimation accuracy of cross-unmanned aerial vehicle interaction and suppress overestimation, thereby making training more stable and achieving higher steady-state performance. Second, a Meta-strategy was adopted during the training period for adaptive scheduling (such as dynamically adjusting loss weights and exploration intensity) to improve sample efficiency and convergence speed without increasing inference-time overhead. Finally, under identical experimental conditions, red unmanned aerial vehicles respectively adopting multi-agent deep deterministic policy gradient (MADDPG), Critic multi-agent deep deterministic policy gradient (Critic-MADDPG), Meta enhanced multi-agent deep deterministic policy gradient (Meta-MADDPG), and CM-MADDPG conducted a game confrontation against blue unmanned aerial vehicles with random maneuvering. Simulation results show that CM-MADDPG achieves significant improvements in final win rate, rewards of each unmanned aerial vehicle, and convergence speed. This proves that CM-MADDPG successfully alleviates sparse reward and training instability issues without increasing inference computational overhead and improves convergence speed and sample efficiency. The research results of this paper provide a new idea and approach for theories of multi-unmanned aerial vehicle game confrontation.
Zheng Zhengang , Meng Yue , Li Xinkai , Zhang Hongli , Wu Yinfei
2026, 58(8):183-202. DOI: 10.11918/202604004
Abstract:To address the problems of delayed obstacle avoidance, frequent unnecessary maneuvers, and reduced traffic efficiency of unmanned aerial vehicle swarms under dynamic obstacle interference in virtual tubes environments, a predictive obstacle avoidance method for unmanned aerial vehicle swarms in three-dimensional virtual tubes scenarios was proposed to improve safe navigation capability and swarm cooperative efficiency in complex dynamic environments. First, a dynamic obstacle avoidance model based on virtual location prediction (VLP) was constructed. By predicting the future position changes of dynamic obstacles within a given time horizon, temporary non-passable regions were established to achieve early identification and proactive avoidance of potential collision risks. Second, considering that the traditional dynamic window approach (DWA) and velocity obstacle (VO) algorithms are mainly applicable to two-dimensional motion scenarios, the DWA and VO algorithms were extended to three-dimensional space. Combined with the spatial constraint characteristics of elastic regions in virtual tubes, a fusion obstacle avoidance framework for three-dimensional virtual tubes environments was developed. Among them, the DWA algorithm was employed to generate candidate velocity sets satisfying the kinematic constraints of unmanned aerial vehicles, and the VO algorithm was utilized to eliminate dangerous velocity regions with collision risks, thereby improving the local dynamic obstacle avoidance capability of unmanned aerial vehicles in three-dimensional space. Finally, a dynamic obstacle avoidance strategy based on risk state classification was designed. By comprehensively considering relative distance, prediction time, relative velocity, and motion direction, the collision risk was dynamically evaluated, enabling unmanned aerial vehicles to execute avoidance maneuvers only when actual collision risks existed, thus reducing ineffective detours and frequent deceleration. The proposed method was validated through multiple types of dynamic obstacle scenarios, a virtual tubes swarm traversal scenario involving 15 unmanned aerial vehicles, and physical test. The results demonstrate that the proposed method can achieve safe and orderly navigation of unmanned aerial vehicle swarms under unknown dynamic obstacle interference conditions and has good trajectory stability and dynamic obstacle avoidance capability. At the same time, it can effectively reduce unnecessary avoidance behaviors and improve the traffic efficiency of swarms in complex environments, which verifies the engineering application potential of the method in scenarios such as low-altitude logistics, cooperative inspection, and complex low-altitude airspace management.
Dai Pei , Liu Xiaogang , Xu Yuansi , Jiang Yinghao , Bi Meng , Tian Yanhao , Huo Mingying , Deng Jianing , Qi Naiming
2026, 58(8):203-214. DOI: 10.11918/202509046
Abstract:To enhance the safety and efficiency of the wireless charging hangar-unmanned aerial vehicle system in urban scenarios and facilitate the development of the low-altitude economy, this paper investigated the spatiotemporal path planning for this collaborative system. Firstly, considering the 120 m flight altitude restriction for unmanned aerial vehicles in China, this paper established an urban gridded environment model containing four layers of 0,0, 90, and 120 m. Secondly, this paper proposed a hierarchical path planning algorithm of “global coarse planning and local detailed planning”: At the global level, the improved A* algorithm was applied to complete path optimization between hangar nodes; at the local level, the improved rapidly-exploring random tree (IRRT) algorithm was adopted to achieve precise obstacle avoidance in densely built urban areas. Finally, this paper integrated a spatiotemporal planning mechanism: Spatially, height layers were allocated according to path characteristics to avoid airspace conflicts; temporally, the charging time was dynamically adjusted based on the multi-stage constant current (MCC) strategy. Simulation results indicate that the proposed dynamic charging scheme reduces the total charging time by 55% compared with the full-charging strategy; the hierarchical path planning algorithm realizes the efficient adaptation between the connection of global hangar nodes and local building obstacle avoidance, and the altitude layering model can conform to the urban building distribution and regulatory requirements for low-altitude flight. This paper validates the effectiveness of the proposed model, algorithm, and mechanism, addresses the problems that traditional single planning algorithms struggle to balance global and local aspects and that traditional charging management strategies have low efficiency, and provides support for the practical application of the wireless charging hangar-unmanned aerial vehicle system.
Zhang Luyao , Zhang Zhihong , Wang Qingqi
2026, 58(8):215-226. DOI: 10.11918/202511016
Abstract:To address the conflict management pressure induced by large-scale urban low-altitude operations of electric vertical take-off and landing (eVTOL) aircraft and to enhance system safety and efficiency under varying operational densities, a density-adaptive hybrid conflict resolution strategy was proposed. First, based on the traffic density variation characteristics, a “strategy applicable density interval” was defined to enable adaptive switching and cooperative operation between centralized and decentralized resolution mechanisms under different density conditions. Second, an air-ground collaborative evaluation system integrating the conflict resolution failure rate and ground risk was constructed to comprehensively quantify the overall performance of different strategies in safety and efficiency. To address the strong conservativeness of the traditional velocity obstacle method in high-density scenarios, a local density parameter was introduced, and a density-weighted velocity obstacle model was proposed, which adaptively improved the environmental adaptability of decentralized strategies by adjusting the conflict detection range. Finally, based on a three-dimensional urban simulation platform, same-priority and multi-priority operation modes were constructed, and a systematic comparison was conducted among the centralized, traditional decentralized, and improved decentralized strategies. The results show that in low-density scenarios, the centralized strategy combined with trajectory replanning achieves better performance; in high-density scenarios, the improved decentralized strategy combined with fixed routes exhibits the highest stability; in intermediate density intervals, the two types of strategies can achieve dynamic switching through a cooperative buffering mechanism. The proposed hybrid strategy reduces the conflict resolution failure rate by approximately 24.1% and 18.3% and reduces the ground risk by approximately 39.1% and 8.4%, respectively.
Guo Yulong , Fan Xing , Shi Lingxia , Wang Jialong , Cheng Jidong , Lei Xu
2026, 58(8):227-239. DOI: 10.11918/202507044
Abstract:To address the problems of insufficient accuracy, poor generalization ability, and reliance on complex manual feature engineering in existing state-of-health (SOH) prediction methods of lithium-ion batteries when lacking massive labeled data, this paper proposed a new method for predicting SOH of lithium-ion batteries based on improved large language model (LLM), named T-LLM. First, this paper specifically designed a textualization strategy to convert numerical battery health indicator data into language sequences understandable by LLM. This strategy retained the internal structure and relative relationships of numerical values, enabling the LLM to utilize its powerful autoregressive structure and context understanding ability to capture the complex degradation dynamics of batteries. Second, experimental verification on the NASA and University of Oxford battery datasets demonstrates that T-LLM exhibits excellent performance and generalization ability. On the NASA dataset, the R2 of this method stabilizes above 96%, significantly outperforming models such as ARIMA and CNN-LSTM. Finally, it achieves a prediction accuracy of over 94% with only 30% of the training data, effectively overcoming the dependence of traditional methods on large-scale labeled data. This method also successfully realizes zero-shot cross-dataset prediction on the Oxford dataset; the R2 reaches over 88% without fine-tuning, solving the problem of the poor generalization ability of traditional models. The research results demonstrate that the T-LLM method not only has high prediction accuracy but also quickly adapts to different battery types, providing a novel solution for achieving precise SOH prediction of batteries with high data efficiency and strong generalization ability.
Yang Xiaojian , Cui Peng , Cui Qiming , Li Haosong , Ren Boda , Zhang Qi , Zhang Yifan , Li Mingyu , Zhu Wenbo
2026, 58(8):240-246. DOI: 10.11918/202502018
Abstract:To solve the magnetic performance failure of conical magnetic flux conductors caused by eddy current heat accumulation under high-frequency conditions and to break through the technical bottleneck that thermal and magnetic properties are difficult to co-optimize in traditional designs, this paper established a quantitative eddy current heating model and proposed an optimized segmented cutting structure scheme. First, based on the magnetic flux density distribution function and Joule heating theory, this paper established a mathematical model to reveal the quantitative relationship between eddy current loss power and geometric/electromagnetic parameters. Second, given the asymmetric characteristics of conical magnetic flux conductors, this paper innovatively designed asymmetric petal-shaped cuts to reconstruct the eddy current path. Finally, this paper verified the effectiveness of the scheme through temperature field tests, magnetic flux density distribution experiments, and power density simulations. The results show that the optimized magnetic flux conductor reduces the temperature of the maximum temperature rise region by 28% on the premise of maintaining a solder joint temperature drop of only 1.8%. Moreover, the magnetic flux distribution test indicates that no magnetic field distortion is caused, and the error between simulation and actual measurement is less than 5%. The research results provide a thermo-magnetic co-optimization method with both theoretical accuracy and engineering practicability for the design of high-frequency magnetic flux conductors.
You are the visitor 
Address:Box 136, 92 Xidazhi Street, Nangang District, Harbin
Postcode:150001 E-mail:hitxuebao@hit.edu.cn
ICP:京ICP备09084417号-23
Copyright:Journal of Harbin Institute of Technology ® 2026 All Rights Reserved
