Article
Chemistry, Analytical
Li-Yu Lo, Chi Hao Yiu, Yu Tang, An-Shik Yang, Boyang Li, Chih-Yung Wen
Summary: This paper presents a learning-based UAV system for autonomous surveillance, utilizing the YOLOv4-Tiny algorithm and integrating 3D object pose estimation and Kalman filter to enhance perception performance. The fully autonomous system includes UAV path planning and is validated through flight experiments, demonstrating robustness, effectiveness, and reliability in performing surveillance tasks.
Article
Engineering, Aerospace
Tagir Z. Muslimov, Rustem A. Munasypov
Summary: The paper introduces a novel approach for controlling a swarm of fixed-wing UAVs to fly in parallel formation, using cooperative control laws and backstepping techniques to ensure stability and precision in flight.
AEROSPACE SCIENCE AND TECHNOLOGY
(2021)
Article
Cell Biology
Livia Autore, James D. O'Leary, Clara Ortega-de San Luis, Tomas J. Ryan
Summary: This study investigates the impact of forgetting on neuronal encoding in mice. The results show that forgetting leads to decreased reactivation of brain cells, but optogenetic stimulation can induce memory retrieval. Forgotten neuronal ensembles can be reinstated through the presentation of similar or related environmental information. These findings suggest that forgetting is a reversible and updatable process.
Article
Engineering, Geological
Filippo Catani
Summary: Recent advancements in mobile surveying platforms and crowdsourced geoinformation have led to a vast amount of unvalidated data available for research. This has prompted the development of convolutional neural networks specifically designed for automated landslide recognition in non-standard images, showing promise in improving efficiency in landslide hazard studies.
Article
Remote Sensing
Yuzhuang Wan, Yi Zhong, Yan Huang, Yi Han, Yongqiang Cui, Qi Yang, Zhuo Li, Zhenhui Yuan, Qing Li
Summary: This paper proposes a two-stage Adaptive Region Selection Detection framework for object detection in high-resolution Unmanned Aerial Vehicles images. The framework utilizes coarse localization and target clustering algorithms to select object-dense sub-regions for detection. Experimental results demonstrate the effectiveness and adaptiveness of the proposed framework.
Article
Computer Science, Information Systems
Yuzhao Liu, Wan Li, Li Tan, Xiaokai Huang, Hongtao Zhang, Xujie Jiang
Summary: Unmanned aerial vehicle (UAV) object detection technology is widely used for real-time collection and analysis of image data to determine the category and location of targets. However, detecting small-scale targets can be challenging and compromise security surveillance effectiveness. In this study, a novel dual-backbone network detection method (DB-YOLOv5) is proposed to enhance the extraction capability of small-scale target features and improve accuracy. Experimental results on the VisDrone-DET dataset demonstrate a 3% improvement over the benchmark model, highlighting the effectiveness of the proposed method. This approach enhances security surveillance in UAV object detection and provides a valuable tool for protecting critical infrastructure.
Article
Automation & Control Systems
Yen-Chen Liu, Tsung-Wei Ou
Summary: This paper presents the design, analysis, and implementation of an adaptive backstepping controller for underactuated quadrotors to track time-varying trajectories with parameter uncertainties. By decoupling uncertain mass and inertia from lifting force and moment torque, stability and convergence of tracking errors are ensured through adaptive laws. The control scheme is extended to quadrotors with velocity motor input, addressing trajectory generation and tracking performance without the need for acceleration information. Simulation and experimental results demonstrate the efficacy of the proposed controller for object transportation.
IET CONTROL THEORY AND APPLICATIONS
(2021)
Article
Forestry
Xiongwei Lou, Yanxiao Huang, Luming Fang, Siqi Huang, Haili Gao, Laibang Yang, Yuhui Weng, I. -K Uai Hung
Summary: This study demonstrated the successful identification of tree crowns and widths in two loblolly pine plantations using machine learning algorithms combined with aerial imagery captured by UAV. The application of deep learning object detection methods, especially the SSD model, proved to be effective in estimating crown sizes, offering a cost-effective alternative to traditional ground measurement for forest inventory purposes.
JOURNAL OF FORESTRY RESEARCH
(2022)
Article
Computer Science, Information Systems
Himanshu Gupta, Om Prakash Verma
Summary: This article proposes novel aerial image traffic monitoring and surveillance algorithms based on advanced DL object detection models, with YOLOv4 demonstrating superior efficiency and real-time practical implementation compared to other developed models.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Remote Sensing
Pavol Kurdel, Marek Ceskovic, Natalia Gecejova, Frantisek Adamcik, Maria Gamcova
Summary: This article focuses on the analysis of UAV flight control and the quality of prediction and elimination of errors in maneuvers. The aim is to highlight the solvability of the complexities in such flight procedures and assess the readiness for the descent phase of autonomous UAVs.
Article
Computer Science, Artificial Intelligence
Xiaoyu Yu, Fuchao Li, Pengfei Bai, Yan Liu, Yinglu Chen
Summary: Data augmentation helps diversify the information in the dataset, and copy-paste augmentation generates new class information to mitigate class imbalance. The authors propose a self-adaptive data augmentation algorithm called CPA, which addresses the issues of over-fitting and under-fitting. CPA generates class weights based on model evaluation results and class imbalance information, replenishes different amounts of class information accordingly, and incorporates the generated images into the training dataset. Experimental results show that CPA can alleviate class imbalance.
IET COMPUTER VISION
(2023)
Article
Computer Science, Artificial Intelligence
Ruiqian Zhang, Zhenfeng Shao, Xiao Huang, Jiaming Wang, Yufeng Wang, Deren Li
Summary: Object detection in UAV imagery is crucial in various fields, but it faces challenges due to the complex characteristics of the images. To address this, a novel Adaptive Dense Pyramid Network (ADPN) is proposed, which incorporates object distribution information and density prediction to improve detection accuracy.
Article
Engineering, Aerospace
Yuanliang Xue, Guodong Jin, Tao Shen, Lining Tan, Lianfeng Wang
Summary: This paper addresses the problem of visual object tracking for Unmanned Aerial Vehicles (UAVs). A simple yet efficient tracker following the basic architecture of the Siamese neural network is proposed, which improves the classification ability from three stages. Experimental results demonstrate that the proposed tracker achieves favorable tracking performances in aerial benchmarks beyond 41 frames/s.
CHINESE JOURNAL OF AERONAUTICS
(2023)
Article
Automation & Control Systems
Simone Baldi, Spandan Roy, Kang Yang, Di Liu
Summary: This article presents a new adaptive autopilot design for fixed-wing UAVs based on uncertain Euler-Lagrange dynamics. The design takes into account the under-actuation, reduced structural knowledge of cross-couplings and trimming points in the dynamics. The effectiveness of the control design is validated through hardware-in-the-loop tests and comparisons with other autopilots.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Automation & Control Systems
Simone Baldi, Spandan Roy, Kang Yang, Di Liu
Summary: The study introduces a new adaptive autopilot design to effectively address unmodeled effects and uncertainties of fixed-wing UAVs during flight, with hardware-in-the-loop tests validating the effectiveness of the control design.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Engineering, Electrical & Electronic
Soohyun Park, Haemin Lee, Chanyoung Park, Soyi Jung, Minseok Choi, Joongheon Kim
Summary: This paper surveys recent efforts in multiagent reinforcement learning and neural Myerson auction deep learning to improve mobility control and resource management in autonomous vehicles. The findings suggest that integrating MARL CommNet and Myerson techniques is essential for improved efficiency and trustworthiness.
Article
Computer Science, Software Engineering
Soohyun Park, Hao Feng, Chanyoung Park, Youn Kyu Lee, Soyi Jung, Joongheon Kim
Summary: This article introduces an efficient quantum train engine (EQuaTE) as a development tool for quantum neural network (QNN) autonomous driving software. It proposes the use of gradient variances to confirm the presence of local minima situations (barren plateaus) in the QNN and tests the stability and feasibility of QNN-based software during runtime operations. The engine also provides visual feedback by identifying barren plateaus in local autonomous driving platforms and visualizing corresponding information in the cloud, enabling automatic reorganization and retraining of the QNN for eliminating barren plateaus.
IEEE INTERNET COMPUTING
(2023)