• Corpus ID: 237563194

Autonomous Vision-based UAV Landing with Collision Avoidance using Deep Learning

  title={Autonomous Vision-based UAV Landing with Collision Avoidance using Deep Learning},
  author={Tianpei Liao and Amaldev Haridevan and Yibo Liu and Jinjun Shan},
The autonomous vision-based Unmanned Aerial Vehicles (UAVs) landing is an adaptive way to land in special environments such as the global positioning system denied. There is a risk of collision when multiple UAVs land simultaneously without communication on the same platform. This work accomplishes vision-based autonomous landing and uses a deep-learning-based method to realize collision avoidance during the landing process. Specifically, the landing UAVs are categorized into level I and II… 

Figures from this paper


Finding every car: a traffic surveillance multi-scale vehicle object detection method
The inverted residuals technique is used to improve the convolutional layer of YOLOv3 to reduce the missing of predicted boxes due to vehicle overlaps, and soft non maximum suppression is used in order to cope with the overlapping of vehicles in traffic videos.
Prototype conflict alerting system for free flight
The development of a prototype alerting system for a conceptual free conceptual free environment is discussed, based on a probabilistic model of aircraft sensor and trajectory uncertainties that need not be Gaussian distributions.
YOLOv4: Optimal Speed and Accuracy of Object Detection
This work uses new features: WRC, CSP, CmBN, SAT, Mish activation, Mosaic data augmentation, C mBN, DropBlock regularization, and CIoU loss, and combine some of them to achieve state-of-the-art results: 43.5% AP for the MS COCO dataset at a realtime speed of ~65 FPS on Tesla V100.
ARTag, a fiducial marker system using digital techniques
  • M. Fiala
  • Computer Science
    2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05)
  • 2005
This proposed new marker system, ARTag has very low and numerically quantifiable error rates, does not require a grey scale threshold as does other marker systems, and can encode up to 2002 different unique ID's with no need to store patterns.
Surveillance, Privacy and the Ethics of Vehicle Safety Communication Technologies
  • M. Zimmer
  • Computer Science
    Ethics and Information Technology
  • 2006
Recent advances in wireless technologies have led to the development of intelligent, in-vehicle safety applications designed to share information about the actions of nearby vehicles, potential road
AprilTag 2: Efficient and robust fiducial detection
  • John Wang, E. Olson
  • Computer Science
    2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
  • 2016
This work describes AprilTag 2, a completely redesigned tag detector that improves robustness and efficiency compared to the original AprilTag system, and improves performance with higher detection rates, fewer false positives, and lower computational time.
Determination of thermal properties of composting bulking materials.
First order regression models developed in this paper represent the relationship between degree of saturation and volumetric heat capacity and thermal conductivity well and can contribute to improved thermodynamic modeling and heat management of composting processes.
ASAP-NMS: Accelerating Non-Maximum Suppression Using Spatially Aware Priors
An Accelerated NMS algorithm which leverages Spatially Aware Priors, or ASAP-NMS, and improves the latency of the NMS step from 13.6ms to 1.2 ms on a CPU without sacrificing the accuracy of a state-of-the-art two-stage detector on COCO and VOC datasets.
Alibaba tests drone deliveries after Amazon push
  • [Online]. Available: https://www.cnbc.com/2015/02/04/ alibaba-tests-drone-deliveries-after-amazon-push.html
  • 2015
Google colab