- Advanced MIMO Systems Optimization
- Cognitive Radio Networks and Spectrum Sensing
- Energy Efficient Wireless Sensor Networks
- Distributed and Parallel Computing Systems
- Cloud Computing and Resource Management
- Cooperative Communication and Network Coding
- Advanced Computational Techniques and Applications
- Advanced Algorithms and Applications
- Satellite Communication Systems
- Indoor and Outdoor Localization Technologies
- Parallel Computing and Optimization Techniques
- Telecommunications and Broadcasting Technologies
- Embedded Systems Design Techniques
- Advanced Sensor and Control Systems
- Photonic and Optical Devices
- Cloud Data Security Solutions
- Distributed Control Multi-Agent Systems
- Caching and Content Delivery
- Embedded Systems and FPGA Design
- Wireless Communication Networks Research
- IoT and Edge/Fog Computing
- Mobile Ad Hoc Networks
- Microwave Engineering and Waveguides
- Advanced Decision-Making Techniques
- Power Systems and Technologies
Nanjing University of Aeronautics and Astronautics
2018-2025
State Radio Regulation Of China
2015-2024
North China Electric Power University
2021-2024
University of Electronic Science and Technology of China
2013-2024
Xinjiang University
2024
University of Victoria
2007-2023
China Academy of Information and Communications Technology
2023
NARI Group (China)
2023
Nanjing University of Science and Technology
2013-2022
China Electric Power Research Institute
2021
Graph Neural Networks (GNNs) have achieved promising performance on a wide range of graph-based tasks. Despite their success, one severe limitation GNNs is the over-smoothing issue (indistinguishable representations nodes in different classes). In this work, we present systematic and quantitative study GNNs. First, introduce two metrics, MAD MADGap, to measure smoothness over-smoothness graph representations, respectively. Then, verify that smoothing nature critical factor leading low...
We present a novel method for single image depth estimation using surface normal constraints. Existing methods either suffer from the lack of geometric constraints, or are limited to difficulty reliably capturing context, which leads bottleneck quality. therefore introduce simple yet effective method, named Adaptive Surface Normal (ASN) constraint, effectively correlate with consistency. Our key idea is adaptively determine reliable local geometry set randomly sampled candidates derive we...
This paper investigates the stability and boundary of a swarm with general directed weighted topology. The is generally considered as cohesiveness. We construct symmetric eigenmatrix define an orthogonal eigenparameter which reflects degree orthogonality between left eigenvector coupling matrix corresponding to its zero eigenvalue eigenvectors nonzero eigenvalues. prove that, if topology underlying strongly connected, then stable in sense that all agents will globally exponentially converge...
Integrated satellite and terrestrial networks (ISTN) are rapidly evolving to meet the ever-increasing demands of higher throughput, lower latency, wider coverage for future communications. Meanwhile, ultra-dense deployment low earth orbit (LEO) satellites has emerged as a promising architecture ISTN, which will enable seamless broadband while posing huge challenges wireless resource allocation. Artificial intelligence can be applied ISTN based on LEO constellations scenarios, endowing...
In this article, we consider a stochastic model of wireless sensor networks (WSNs) in which each node randomly and alternatively stays an active mode or sleep mode. The consists two phases, called the full-active phase semi-active phase. When referenced is mode, it may sense data packets, transmit sensed receive relay received packets. However, when switches from full to phase, only able transmit/relay data. does not interact with external world. first, develop for WSN, then derive explicit...
Appropriate candidate frequency bands are extremely important for the development of future 5G systems. In this work, researches on spectrum around world summarized. Then potential systems investigated based practical utilization in China. For below 6GHz, feasibility possible system analyzed, which mainly come from 2G/3G/4G re-farming, identified by footnotes IMT Regulations Radio Frequency Allocation China, and WRC-15 Agenda Item 1.1. Moreover, propagation characteristics proposed China...
The integration of caching and ultra-dense network (UDN) can not only improve the efficiency content retrieval by reducing duplicate transmissions but also throughput system energy (EE) UDN. In this paper, we focus on consumption aspects cache-aided UDN (CUDN) develop a novel strategy to EE. Different from existing researches, consider more realistic scenario where popularity cache is dynamic unknown. proposed deep reinforcement learning (DRL)-based approach that uses Q-network approximate Q...
With the evolution of sixth generation (6G) mobile communication technology, ample attention has gone to integrated terrestrial-satellite networks. This paper notes that four typical application scenarios networks are into ultra dense satellite-enabled 6G architecture. Then subchannel and power allocation schemes for downlink heterogeneous introduced. Satellite edge computing (SMEC) with caching in three-layer serves reduce link traffic Furthermore, a scheme interference management is...
Benefiting from the flexibility and low operational cost, dispatching unmanned aerial vehicles (UAVs) to collect measurements is promising in spectrum cartography (SC). The main goal optimize trajectory of an UAV seek most informative measurement under environment with dynamic emitters. In this letter, we formulate a Markov Decision Process find optimal flight that maximizes accuracy SC minimizes energy consumption. However, due unavailable instantaneous feedback about SC, existing methods...
The Radio Environment Map (REM) provides an effective approach to Dynamic Spectrum Access (DSA) in Cognitive Networks (CRNs). Previous results on REM construction show that there exists a tradeoff between the number of measurements (sensors) and accuracy. In this paper, we analyze determine error is decreasing convex function (sensors). concept geographic entropy introduced quantify relationship. And influence sensor deployment accuracy examined using information theory techniques. obtained...
A RFID-based Kindergarten Intelligence Security System is proposed in this paper. The RF-KISS helps to achieve better children security control, as well improve of school environment. To purpose, it automates the supervision and provides integration with current management system for Kindergarten. In system, RFID tags are embedded uniforms. addition, kindergartens equipped intelligent terminal reader antenna support automatic data scanning storage location checking. Moreover, a middle layer...
Considering that coverage evaluation and control is one of the critical areas in surveillance, this study we use deployment entropy to assess quality a swarm micro air vehicles (MAVs). Depending on result entropy, also develop two-level redeployment algorithm cope with MAVs' problem for maximizing coverage. With goal distributing MAVs evenly over interested region, divide process into two levels: global level local level. We propose strategy MAV at both levels. From numerical simulation,...
Based on polynomial interpolation and approximation theory, a special feed-forward neural network using power activation functions is constructed in this paper. The model employs three-layer structure with the hidden-layer neurons activated by group of order-increasing (while other layers' use linear functions). In addition, weights-updating formula for such could be derived from standard BP training method. A pseudoinverse-based method (or termed, weights-direct-determination /...
A coupled plasmon waveguide resonance (CPWR)-based optical sensor takes advantage of the evanescent wave at a boundary to excite surface polaritons, which can link electromagnetic energy enhance propagating resonance. Herein, novel having BK7 prism/Ag/SiO2/SiO2:Ag structure and that couples localised Ag nanoparticles in SiO2 film was designed fabricated. The results showed proposed exhibited higher sensitivity (~1396.85 nm RIU−1) figure merit (~11.36 values than traditional CPWR (i.e.,...
The emergence of coverage holes is unavoidable in wireless sensor networks. In this letter, a novel method for describing graphically proposed. Our proposed graphic hole description divided into two phases, namely, detecting and describing. Simulation results show that the indicates not only locations configurations but also vulnerable parts holes. Besides, can serve as useful tool healing
The strength of the GPS signal on earth's surface averages-160 dBw. While many receivers leave large space for dynamics, enough power is left signals to be spoofed. Spoofing completely different from jamming. objective jamming simply interrupt availability in at receiver. effect cause receiver corrupted so that no valid can decoded by goal spoofing, other hand, provide with a misleading signal, fooling use fake positioning calculations. will produce position solution. P-code heavily...
Summary Due to the scarcity of spectrum resources, non‐geostationary orbit (NGSO) satellite constellations, which have achieved rapid development in recent years, deal with challenge spectral coexistence geostationary (GSO) systems. OneWeb has put forward “progressive pitch” strategy for GSO networks protection; however, technical details are not exposed. Herein, we propose a method obtain exact progressive pitch parameters suitable system terms route this strategy. Constrained by system's...
There are many Distributed Denial of Service (DDoS) attack accidents in the world, which use Internet Things (IoT) devices to launch attacks and make network unavailable such as Mirai. IoT have reached more than 8 billion units by 2018 world. However, there exist some problems environment, simple protocols, fixed bandwidth, a lot devices, easily controlled. Therefore, vulnerable intrusion become puppet for DDoS attacks. The shortcomings traditional methods include slow learning speed low...