Abstract:As “low altitude economy” is included in the government reports， it become the hot topic in 2024. Due to the advantages of high efficiency， flexibility， low cost， and multi payload， unmanned aerial vehicles （UAVs） are regarded as the main form of “low altitude economy”. As the key factor to guarantee flight safety and communication security， the spectrum management of UAV communication is an indispensable factor in promoting the vigorous development of the “low altitude economy”. This paper starts with the changes of UAV spectrum management policies from 2015 to 2023， and then deeply explores the regulations， standards and technologies of UAV communication， including the operating frequency bands and flight supervision， as well as standard specifications represented by international organizations， such as International Telecommunication Union （ITU）， Institute of Electrical and Electronics Engineers （IEEE） and The 3rd Generation Partnership Project （3GPP）. The subjects of channel models and interference mitigation strategies that closely related to UAV communication spectrum management are also discussed. Finally， current challenges and future research directions of UAV communication spectrum management are presented.

Abstract:The low-altitude intelligent network， as a new type of productivity， has facilitated the rapid development of the low-altitude economy. However， the widespread application of unmanned aerial vehicles （UAVs） has posed significant challenges for airspace regulation. This paper mainly focuses on the performance analysis of two potential UAV flight regulation technologies applied to the low-altitude intelligent network： Automatic dependent surveillance-broadcast （ADS-B） and remote identification （Remote ID）. Firstly， we systematically introduce the basic mechanisms of ADS-B and Remote ID. Then， based on current technical standards， theoretical transmission distances of these two technologies are analyzed， and methods for evaluating positioning accuracy are defined. We build ADS-B and Remote ID experimental systems that meet performance requirements， estimate the actual transmission distances through measured signal strength， and measure the positioning accuracy of latitude， longitude， and altitude， as well as the packet loss rate. Through the analysis of the measured data， this paper comprehensively evaluates practical application effects of ADS-B and Remote ID in low-altitude intelligent network for the first time. Results show that ADS-B outperforms Remote ID in terms of transmission range and positioning accuracy， while Remote ID performs better in altitude positioning. In terms of communication stability， ADS-B provides stable reception over long distances， while Remote ID performs well in short-range scenarios. Finally， the future development directions of UAV regulation technology are discussed， and solutions for optimizing transmission distance， coverage range， positioning accuracy， and packet loss rate are proposed.

Abstract:The low-altitude intelligent network （LAIN） serves as a fundamental infrastructure for the development of the low-altitude economy， and spectrum management and control （SMC） is one of the key technologies to address the problems of illegal spectrum use and malicious attacks in LAIN， as well as to improve the utilization rate of spectrum resources. Embodied artificial intelligence （EAI）， as a key research direction in the new generation of artificial intelligence， offers new possibilities for the development of SMC for LAIN due to its characteristics of physical embodiment， environmental interaction， and intelligent growth. Firstly， this paper introduces the requirements of SMC for LAIN from the aspects of technical framework， research status， and main challenges. Secondly， by sorting out the connotation and advantages of EAI， the concept and significance of EAI-enabled SMC for LAIN are analyzed. Furthermore， based on the closed-loop structure of “perception-decision-action-feedback”， a technology of SMC for LAIN is proposed， which includes low-altitude embodied spectrum sensing， inferring and decision-making， and action and feedback. This provides a possible technical pathway for achieving efficient and secure SMC for LAIN.

Abstract:Aiming at the problems of beam training and target localization and tracking in millimeter-wave （mmWave） low-altitude unmanned aerial vehicle （UAV） scenarios， inspired by information theory， this paper proposes a hierarchical beam training algorithm based on the channel coding principle and a UAV target localization and tracking algorithm based on mmWave radar sensing， respectively. The proposed algorithms have high generalization and robustness， and are applicable not only to static and dynamic scenarios， but also to far-field， near-field， reconfigurable intelligent surface （RIS） assisted communication， and distributed cellular-free network scenarios， as well as illegal UAV intrusion sensing， etc. The algorithms are also validated through simulations and hardware platform tests. Specifically， the channel coding beam training algorithm can significantly improve the training accuracy by using coding gain and error correction mechanism. The mmWave radar algorithm combines Capon beam formation， constant false alarm rate （CFAR） detection and density-based spatial clustering of applications with noise （DBScan） to achieve UAV detection and tracking. Both simulation and hardware test results show that these algorithms can effectively improve the efficiency of beam training and the accuracy of sensing and localization in mmWave low-altitude UAV scenarios， providing technical support for the further prosperous development of low-altitude economy.

Abstract:Unmanned aerial vehicle （UAV）-mounted base stations possess characteristics， such as rapid deployment and flexible coverage， making them an effective solution for emergency communication in railways. However， the low-altitude communication network formed by UAVs faces the constraint of the limited energy storage and the risk of data eavesdropping or tampering. This paper introduces blockchain technology into a UAV-assisted railway wireless communication system to ensure data security. Considering the constraints on transmission delay and data queue stability， this paper proposes a joint optimization problem aimed at minimizing the energy consumption of the UAV-assisted communication system and the latency of the blockchain. To solve this non-convex， mixed-integer， and time-varying stochastic optimization problem， a Lyapunov-based drift-plus-penalty method is proposed to transform the long-term stochastic optimization problem into sub-problems of multiple time slots. Deep reinforcement learning based on D3QN-TD3 is designed to solve these sub-problems， and then the optimal association strategies and power control for each time slot are obtained. Experimental results demonstrate the significant effectiveness of the proposed method in reducing the energy consumption and delay.

Abstract:The rise of the low-altitude economy has led to the widespread adoption of small unmanned aerial vehicles （UAVs） in logistics， surveying， and entertainment， yet the associated security risks have grown increasingly prominent. Detecting low-slow-small （LSS） targets is therefore crucial in domains such as national security， airspace regulation， and UAV defense， as it effectively mitigates potential threats posed by small and low-altitude flying objects. In response to limitations of current sensors with regard to cost-effectiveness， operation under complex lighting conditions， and broad field-of-view coverage， this paper proposes an LSS target detection system that leverages both an event camera and an RGB camera. First， the high-speed imaging and wide dynamic range of event camera are employed for an initial “sweep”， and an event-based detection algorithm provides preliminary target localization. Next， an information collaboration module fuses the two-modal data to enhance detection accuracy. Finally， the high-resolution and dynamic zoom features of RGB camera enable a “gaze” mode， combined with a dedicated image recognition algorithm for fine-grained target identification and tracking. Under complex lighting and wide field-of-view scenarios， this system achieves both low cost and high performance， offering an effective new approach to LSS target detection.

Abstract:Image is one of the important ways to obtain information. With the increasing demand of image transmission and storage， especially in bandwidth limited or cloud storage situations， compressing images at extremely low bitrates is of great significance for improving transmission efficiency and saving storage space. Based on this， this paper presents a systematic review of very low bitrate compression techniques for lossy images. Firstly， on the basis of problems of image compression derivative algorithms based on generative adversarial network （GAN） in terms of high-resolution image compression， generating image blur， and neglecting semantic and texture information， the latest very low bitrate image compression methods are introduced. Then， this paper elaborates image compression methods that achieve very low bitrate using non-GAN models such as layered compression， object based， and region of interest. After that， the commonly used datasets and image quality evaluation methods under lossy compression conditions are described. Finally， a summary of very low bitrate lossy image compression techniques are made， and an outlook on their subsequent development is given.

Abstract:Accuracy and efficiency are the key metrics for evaluating the performance of feature selection algorithms. They correspond to the attribute dependence and reduction scale of neighborhood rough sets respectively. Conventional feature selection algorithms often optimize solely based on maximum attribute dependence reduction， overlooking the significance of reduction scale. However， as data feature dimensions increase and category hierarchies emerge， category information becomes complex and structural relationships become chaotic. Traditional attribute dependency calculations fail to effectively utilize category hierarchy information， leading to suboptimal classification performance. In response to this， a mixed hierarchical dependency that considers the relationship between attribute importance and category hierarchy structure is constructed. This treats mixed hierarchical dependency and reduction scale as two independent optimization objectives， and introduces a multi-objective evolutionary algorithm to optimize them independently. This approach improves attribute reduction performance from both the attribute dependency and attribute scale perspectives， resulting in reduction results that meet target constraints. Experimental results demonstrate that the proposed algorithm achieves higher-quality reduction results within target constraints， leading to the improvement of classification accuracy.

Abstract:Graph convolutional neural networks obtain the node representation by aggregating the neighbor node information with high similarity，and selecting the appropriate neighborhood for the node and conducting effective aggregation are the keys to the graph convolutional networks. Most of the existing graph convolutional neural networks directly aggregate the node information in the multi-hop neighborhood，without considering the difference of the aggregation weights of different hop neighborhoods on different nodes in the network. Aiming at this，a path connectivity based neighbor-awareness node classification algorithm （PCNA） is proposed. The node neighborhood is determined by the path connectivity information in the network，and the influence weight of different length paths on the similarity calculation between nodes is adaptively perceived to guide the neighborhood aggregation process of graph convolutional neural network. Specifically，PCNA is composed of a neighborhood perceptron and a node classifier. The neighborhood perceptron adaptively obtains the aggregated neighborhood of each node and the influence weights of paths with different lengths based on the reinforcement learning mechanism，and then uses the path connectivity information between nodes to obtain the similarity matrix. The node classifier uses the obtained similarity matrix to perform neighborhood aggregation to obtain node representation and classify nodes. The comparison experiments with 10 classical algorithms on eight real datasets show that the proposed algorithm has better performance in node classification tasks.

Abstract:Different from ground communications， space communications usually involve signal vehicles that travel over long-distance at a high speed. In these scenarios， the signal transmission faces two difficulties： A low signal-to-noise-ratio （SNR） caused by the long-distance path-loss and a dynamic Doppler-shift caused by the high-speed movement. For Doppler-shift acquisition， the low SNR requires a long-time accumulation to accumulate a large number of signals. However， during this period， the dynamic Doppler-shift disperses all signals’ total energy over a wide frequency range. To address the energy dispersion problem， this paper proposes a local-clustering-acquisition （LCA） algorithm. The LCA algorithm uses the largest elements from the global-ranges to construct a local-range， then selects some large elements from this local-range for clustering， and finally searches the largest cluster from the clustering results to obtain the acquisition result. Theoretical analysis and simulation validation results demonstrate the LCA algorithm’s significant advantages in increasing acquisition probability， as compared with the existing algorithms.

Abstract:Wireless sensor network （WSN） is constrained by limited battery energy and insufficient computing power， and the limited battery life hinders its widespread deployment. In this paper， wireless power transmission （WPT） and multi-access edge computing （MEC） technologies are used to solve the problem of limited energy consumption of sensor nodes. By jointly optimizing the decision of the node offloading， wireless power supply duration and bandwidth resource allocation， the average task completion delay of sensor nodes is minimized to the greatest extent possible. The optimization problem is modeled as a mixed integer programming problem. In order to adapt to the complex and dynamic channel environment， a deep reinforcement learning delay minimization （DrlDM） algorithm based on soft actor critic （SAC） is proposed. The original optimization problem is modeled as a Markov decision process （MDP）. Simulation results show that compared with three baseline experiments， the average delay of the DrlDM algorithm proposed in this paper is reduced by 62.11 %， significantly shortening the average task completion time of nodes.

Abstract:For the physical layer security communication scenario in double-layer satellite networks consisting of geostationary earth orbit （GEO） satellite and low earth orbit （LEO） satellite cluster， we propose a robust secure beamforming （BF） algorithm based on Lagrange multiplier method under the condition that only the imperfect channel state information （CSI） of the eavesdropper is known. LEO satellite clusters， acted as relays， adopt amplification and forwarding protocols to assist GEO satellite in serving the ground legitimate user. An eavesdropper on the ground tries to steal the satellite signal. Firstly， a joint optimization problem is formulated to maximize the signal-to-noise ratio of the ground legitimate user while satisfying the intercept probability （IP） constraint of the eavesdropper and the maximum transmit power constraint of the GEO satellite. Secondly， we use the matrix transformation method and the cumulative density function of the exponential distribution to transform the complex nonconvex IP constraint into convex one. Thirdly， using Lagrange multiplier method， we attain the transmit power of GEO satellite and the closed-form solution to the distributed BF weight vector of LEO satellite clusters in an efficient way. The simulation results show that the proposed algorithm can guarantee the secure communication of the considered double-layer satellite network under different channel errors， verifying the effectiveness of the algorithm in secure transmission and its robustness to channel errors.

Abstract:The near-field steering vector contains the angle and range parameters. They are coupled with each other and difficult to separate. Most existing methods adopt the approximate decoupling model to estimate the angle and range parameters step by step. However， such an approximate decoupling model will inevitably introduce a systematic model error， which could lead to severe localization performance degradation. To address the above challenges，this paper proposes a near-field sources localization method for sparse representation via a non-uniform grid. It directly models the complex near-field sources localization as a lower-dimensional sparse signal recovery problem and adopts sparse Bayesian learning （SBL） to adaptively recover the sparse signal， avoiding the approximate error and improving the parameters estimation accuracy.In the proposed method， the non-uniform grid only contains a few points， reducing the computational complexity greatly. The nearby points neither share the same direction of arrival （DOA） nor the range value， effectively overcoming the high correlation basis. And the grid refinement trick is additionally introduced to further solve the mismatch problem caused by the coarse grid. The numerical simulation results confirm the superiority of the proposed method.

Abstract:In order to improve the accuracy of radar system parameter estimation， this paper proposes a data fusion and parameter fusion method for parameter estimation based on Bayesian principle. We derive the distance information， entropy error and mean square error （MSE） for a multi-radar system under additive complex Gaussian noise conditions， and derive an upper bound on the distance information. The theoretical derivation shows that the maximum a posteriori estimate （MAP） of the position estimation is consistent with the maximum ratio of the position information. The equivalent signal-to-noise ratio of the multi-radar system is equal to the sum of the signal-to-noise ratios of radars in the system. Experimental results indicate that， in general， the performance of data fusion is always superior to that of parameter fusion. However， data fusion relies on the assumption of uniform distribution and requires distortion-free acquisition of the received signals from all nodes， representing an idealized scenario. In contrast， parameter fusion is more aligned with real-world scenarios， and its estimation accuracy is not significantly inferior to that of data fusion. The findings of this study provide valuable guidance for improving the accuracy of target parameter estimation in practical environments.

Abstract:Recent unsupervised person re-identification studies have used clustering and memory dictionaries for pseudo labels to train models. However， these studies ignore that the datasets of person re-identification are collected by different cameras， that is， the distribution difference between cameras is large， and a larger camera variance will lead to decrease in model accuracy. Therefore， camera cluster contrast learning is proposed， which includes cluster contrast loss and camera contrast loss. The cluster contrast loss can realize the consistent update of memory dictionary and reduce the influence of noise labels on the model. Camera contrast loss reduces camera variance by building camera cluster center for each cluster in each camera， narrowing the camera cluster center distance of the same cluster， and making different camera cluster centers farther apart. By camera cluster contrast learning， the impact of camera variance and noise labels on the model is reduced， and the performance of person re-identification is improved. On the four public datasets， camera cluster contrast learning has shown excellent results， effectively alleviating the impact of camera variance on the model.

Abstract:Although the pedestrian re-identification task has made significant progress， the occlusion problem caused by different obstacles is still a challenge in practical application scenes. In order to extract more effective features from occluded pedestrians， a learnable mask and position encoding （LMPE） method is proposed. Firstly， a learnable dual attention mask generator （LDAMG） is introduced to adapt to different occlusion patterns， significantly improving the re-identification accuracy of occluded pedestrians. It makes the network more flexible and better adapts to diverse occlusion situations. At the same time， the network learns contextual information through the mask， which further improves the understanding of the scenes. In addition， we introduce the occlusion aware position encoding fusion （OAPEF） module to solve the problem of losing position information in Transformer. This method helps to perform the fusion of different regional position encoding and allows the network to gain stronger expressive ability. The integration of position encoding in all directions enables the network to understand the spatial correlation between pedestrians more accurately， and improves the ability to adapt to the occlusion situation. Finally， simulation experiments are conducted， and results demonstrate that LMPE performs well on Occluded-Duke and Occluded-ReID occluded datasets and Market-1501 and DukeMTMC-ReID unoccluded datasets， which confirms the effectiveness and superiority of the proposed method.

Abstract:In order to solve the lack of reasonable arrangement of multi-modal fish disease knowledge， and at the same time reduce the redundant data in the knowledge distillation process， so as to deploy a recognition model with low storage， small samples， and high accuracy， this paper proposes a new method， named as FSFDAI-TMRD. In terms of multi-feature collaborative prediction， this paper focuses on improving the original multi-feature collaborative multi-feature prediction architecture of multi-tasks. Firstly， the finer-grained begin-middle-end-single （BMES） method is used instead of the rough labeling of the begin-inside-outside （BIO） method in the original work. Secondly， the formula for calculating the joint probability distribution of the original architecture is modified， so that the model can better recognize the nested noun entities. In terms of cross-modal multi-head distillation， this paper proposes to employ a cross-modal attention mechanism. Firstly， it calculates the multi-head relationship matrix after merging， splitting， and dot product， and secondly， it utilizes the relative entropy for knowledge distillation， so that the model can better align the intermediate features between heterogeneous teachers and students. Meanwhile， this paper also applies the biaffine attention and adversarial weight perturbation function to enhance the learning of multi-feature knowledge such as semantic phonology and word form word meaning. Compared with the mainstream model， the precision， recall and F₁ value of the FSFDAI-TMRD method are improved by 0.45%， 3.96% and 2.28%， respectively. The storage optimization ratio is improved by 3.01% and the model parameter size is reduced by 94.86%.

Abstract:The classification of pancreatic cystic neoplasms into benign and malignant categories is crucial for medical decision-making. This paper is dedicated to enhancing the accuracy of pancreatic cystic neoplasms classification to assist physicians in formulating more precise diagnostic and therapeutic plans. Utilizing radiomics technology and the ResNet50 neural network， a novel classification method for pancreatic cystic neoplasms is proposed， integrating multi-kernel learning and multi-source feature fusion. The key steps of this method include feature selection， kernel matrix fusion， and the construction of the classification model. Feature selection is performed using the least absolute shrinkage and selection operator （LASSO） to reduce redundant features and improve the model’s generalization ability. Subsequently， multi-source features， screened through feature selection， are mapped in basic kernel functions to construct basic kernel matrices for multi-source features. The weights of these kernel matrices are then optimized and summed up to form a fused kernel matrix. Finally， a support vector machine （SVM） classifier is utilized to categorize pancreatic serous and mucinous cystic tumors. The significance of this process lies in SVM’s ability to use the kernel matrix for inner product operations in high-dimensional spaces， thereby finding a hyperplane to classify data in such spaces. The fused kernel matrix， containing multi-source information after feature mapping， provides higher-dimensional and more complex feature representations. Experimental results demonstrate significant performance improvements in the classification task of pancreatic cystic neoplasms， offering more reliable auxiliary information to physicians and holding substantial clinical application potential.

Abstract:The construction of the measurement matrix is a crucial factor influencing the reconstruction performance of compressive sensing techniques. To address the high storage cost of random measurement matrices and the difficulty in satisfying the restricted isometric property （RIP） with deterministic matrices， an improved method for constructing measurement matrices based on chaotic mapping is proposed. This method combines the random Gaussian matrix with the deterministic matrix and chaotic sequences， taking full the advantages of a small number of measurements from random Gaussian matrices and the lower correlation provided by chaotic mappings. Simultaneously， an analysis is conducted on the phase space characteristics of chaotic sequences， the RIP properties of measurement matrices， and the computational complexity involved in constructing optimized measurement matrices. Finally， simulation experiments compare random Gaussian matrices， Toeplitz matrices， and existing composite matrices. The results show that the proposed optimized measurement matrices outperform the other three types of matrices in terms of relative error， success reconstruction probability， and signal-to-noise ratio for one-dimensional random signals. Additionally， these optimized measurement matrices also exhibit improvements in the reconstruction time complexity， peak signal-to-noise ratio， structural similarity index， and mean structural similarity index for two-dimensional images， indicating better reconstruction performance and significant practical value.