A5080295555- Air Traffic Management and Optimization
- Aerospace and Aviation Technology
- Evacuation and Crowd Dynamics
- Natural Language Processing Techniques
- Topic Modeling
- Target Tracking and Data Fusion in Sensor Networks
- Traffic control and management
- Transportation Planning and Optimization
- Handwritten Text Recognition Techniques
- Robotic Path Planning Algorithms
- Robotics and Sensor-Based Localization
- Domain Adaptation and Few-Shot Learning
- Opportunistic and Delay-Tolerant Networks
- Satellite Communication Systems
- Vehicular Ad Hoc Networks (VANETs)
- Cloud Computing and Resource Management
- Advanced Algorithms and Applications
- Autonomous Vehicle Technology and Safety
- Multimodal Machine Learning Applications
- Human-Automation Interaction and Safety
- Software-Defined Networks and 5G
- Human Mobility and Location-Based Analysis
- Indoor and Outdoor Localization Technologies
- Vehicle License Plate Recognition
- Energy Efficient Wireless Sensor Networks
Wuhan University of Technology
2016-2025
Sanya University
2021-2025
Beihang University
2025
Chinese Academy of Civil Aviation Science and Technology
2025
Chinese Aeronautical Establishment
2025
China National Petroleum Corporation (China)
2025
Inner Mongolia Electric Power (China)
2023-2024
Beijing University of Posts and Telecommunications
2021
China Shenhua Energy (China)
2021
Anhui Polytechnic University
2021
For autonomous robots to navigate a complex environment, it is crucial understand the surrounding scene both geometrically and semantically. Modern employ multiple sets of sensors, including lidars, radars, cameras. Managing different reference frames characteristics merging their observations into single representation complicates perception. Choosing unified for all sensors simplifies task perception fusion. In this work, we present an end-to-end pipeline that performs semantic...
Accurate path planning is essential for effective regional avoidance in multiple unmanned aerial vehicle (multi-UAV) systems. Existing static path-planning techniques often fail to integrate information sources, resulting diminished performance information-rich and dynamic environments. This paper proposes a distributed collaborative algorithm dynamically changing targets complex environments with multisource information. More specifically, multi-UAV collaboration method based on...
In current semi-supervised object detection techniques, one-stage detectors typically realize only limited improvements when compared to two-stage detectors. The limitation is largely due the assignment of incorrect labels samples during label allocation phase, as process influenced by inaccurate pseudo-labels. To tackle problem, we propose a Cropping-Classification Semi-supervised Learning (CCSL) for After detection, employ module (CCM) designed refine pseudo-labels and bolster precision...
Traffic light is regarded as one of the most effective ways to alleviate traffic congestion and carbon emission problems. However, traditional cannot meet challenges in regulation posed by fast growing number vehicles increasing complexity road conditions. In this paper, we propose a dynamic method based on virtual (VTL) for Vehicle Ad Hoc Network (VANET). our framework, each vehicle can express its "will" - desire moving forward share among another value related information at controlled...
Currently, the growth of mobile technologies, lead to a necessity develop handwritten recognition applications. While Latin and Chinese has been extensively investigated using various techniques, so little works have done on Arabic recognition, none existing techniques is accurate enough for practical application. Over past few years, deeper convolutional neural networks (CNNs) widely employed improving performance. In this paper, we enhance popular AlexNet Handwritten Words Recognition...
Under the current Internet environment, middlebox management has become a significant challenge for network operators. Schemes in prior works tried to simplify it with Software-Defined Networking (SDN) technologies, and they provided reliable flexible approaches configure middlebox-related flow entries. However, these schemes are inefficient resource utilization dynamically changing traffic requirements, as mainly focus on stationary hardware middleboxes. Further-more, latencies of packets...
This paper proposes an improvement to the PageRank algorithm. Most existing algorithms expect a strong correlation among consecutively accessed webpages, which in reality should be fuzzy relationship when user accesses pages on arbitrarily basis. We mine data from search-behavior logs by analyzing chronological sequential patterns, and cluster all webpages using C clustering. The weight of each is identified with information entropy, then used adjust average weight. A sample 1 million for...
The combination of pre-trained language models (LM) and knowledge graphs (KG) can enhance the reasoning ability for Question Answering.However, previous methods typically fuse two modalities in a shallow or knowledgedraining manner, not taking full advantage representation both.How to effectively different representations is still problem current research.In our work, novel model proposed that fuses LM modal graph neural network (GNN) deeply over multiple layers modality interaction...
This paper introduces a concept of integrating commercial space flight operations with conventional aviation in the Next Generation Air Transportation System (NextGen). reviews current surveillance systems and conducts feasibility study on use Automatic Dependent Surveillance-Broadcast (ADS-B) to aid integration. The scope this is focused services future suborbital vehicles (e.g., Reusable Launch Vehicles (RLVs)), inside National Airspace (NAS), their impacts or potential hazards NAS....
This paper presents a new aircraft traffic tracking algorithm for surveillance applications that integrates sensory information from multiple avionics sensors in an efficient manner and assesses the sensor consistency integrity purposes. Measurements various ownship are used to form relative baseline vector. Once these vectors formed, they integrated estimated using Interacting Multiple Model (IMM) filter, which has Kalman filters with different dynamics models running parallel interacting...
The lack of aircraft state awareness has been one the leading causal and contributing factors in aviation accidents. Many these accidents were due to flight crew's inability understand automation modes properly monitor energy attitude state. capability providing crew with improved airplane (ASA) is essential ensuring safety. This paper focusses on predictive alerting methods achieve ASA describes used predict (a) stall overspeed conditions, (b) high-and-fast (c) low-and-slow (d) unstable...
The extraction of vein traits from venation networks is great significance to the development a variety research fields, such as evolutionary biology. However, traditional studies normally target reticulate structure (ReSTs), which not sufficient enough distinguish difference between orders. For hierarchical (HiSTs), only few tools have made attempts with human assistance, and obviously are practical for large-scale extraction. Thus, there necessity develop method automated hierarchy...
VCPS (Vehicular Cyber Physical Systems) is a special kind of networked cyber physical system in which each vehicle regarded as communication unit. Vehicle's movement restricted by road and environment VCPS, while traditional random mobility model waypoint cannot reflect the realistic traces. In with high speed vehicles, network topology undergoing tremendous changes all time, greatly undermines stability between vehicles. The diversity complexity traffic scenarios have also increased...