|
An Infinity-Norm Based Approach to Design of Collision Avoidance Control System for UAVs |
|
YunA Oh, Myoung Hoon Lee, Jun Moon(Hanyang University, Korea) |
| In this paper, we propose the Infinity-Norm based collision avoidance control system for UAVs. The worst-case collision avoidance is captured via the infinity-norm. With standard two-norm, the actual obstacle may not be covered and this can cause the collision.
We propose the collision detector and design the Model Predictive Control tracker and the modified backstepping controller to generates the optimal trajectory and control the UAVs, respectively.
The hardware-in-the-loop simulations (HILS) and experiments demonstrate the proposed approach. |
|
|
Indoor Path Planning for an Unmanned Aerial Vehicle via Curriculum Learning |
|
Jongmin Park(Yonsei University, Korea), Sooyoung Jang(Electronics and Telecommunications Research Institute, Korea), Younghoon Shin(Yonsei University, Korea) |
| In this study, reinforcement learning was applied to learning two-dimensional path planning including obstacle avoidance by unmanned aerial vehicle (UAV) in an indoor environment. The task assigned to the UAV was to reach the goal position in the shortest amount of time without colliding with any obstacles. Reinforcement learning was performed in a virtual environment created using Gazebo, a virtual environment simulator, to reduce the learning time and cost. Curriculum learning, which consists of two stages was performed for more efficient learning. |
|
|
Collision Avoidance based on CPA algorithm using Relative Distance and Azimuth of Radar for Unmanned Aerial Vehicles |
|
Hyeji Kim, Cheonman Park, Seongbong Lee, Dongjin Lee(Hanseo University, Korea) |
| In this paper, we propose an obstacle states estimation algorithm based on radar and a collision avoidance algorithm based on closest point of approach(CPA) for unmanned aerial vehicles(UAVs), in order to avoid collision with the obstacle that do not provide information. We estimated position of detected objects using radar and selected the object with the highest risk of collision as the obstacle among the detected objects. Based on the CPA calculated using the states of the UAV and the obstacle, we detected collision and generated the velocity command for collision avoidance. The performance of proposed algorithm was verified through the flight tests. |
|
|
Quadcopter Control using the Viscoelastic Control Law |
|
Manas Kumar Sahoo, Jayanta Kumar Dutt, Subir Kumar Saha(Indian Institute of technology Delhi, India) |
| Quadcopter is an unmanned aerial vehicle (UAV) with four rotors and generally, they are placed in a square formation with equal distance from the centre of mass. The thrust produced by each rotor is directed upward and the UAV is controlled by varying the magnitude and direction of the rotor speed with the help of DC motors. In this paper, a novel controller is proposed that imitate the behavior of four-element viscoelastic material. Some unmodelled disturbances was fed to the system to check the controller performance. Then it was compared with other controllers. |
|
|
Viewpoint Selection for DermDrone using Deep Reinforcement Learning |
|
Mojtaba Ahangar Arzati, Siamak Arzanpour(Simon Fraser University, Canada) |
| This paper presents an RL-based method to improve the performance of real-time 3D human pose estimation (HPE) as a positioning feedback for DermDrone which is a micro sized quadrotor designed MetaOptima to capture high resolution full body images for dermatology application. The camera viewpoint is identified as the key parameter in the accuracy of monocular 3D HPE. We present a DRL-based method for determining the best viewpoint given the flight trajectory. Our goal is to present a reliable and accurate positioning feedback for DermDrone using a 3D HPE. DQN and its variants were employed and their performances were investigated by conducting several simulations. |
|
|
A Novel Fixed Twin-Rotor Unmanned Aerial Vehicle with Variable Angle Louver Rudder |
|
Yixin Zhang, Jianfeng Zhou, Shaoping Wang, Mengqi Yang, Shaoshi Li(Beihang University, China) |
| A novel fixed twin-rotor unmanned aerial vehicle (UAV) based on the variable angle louver rudder (VALR) is developed in this work. In view of the mechanical characteristics under the action of the louver rudder, specific kinematics and dynamics models are derived. The aerodynamic characteristics of the UAV are studied through computational fluid dynamics simulation, and a ground measurement system is designed to analyze the aerodynamic characteristics of the prototype. An attitude control model is proposed, and maneuvers such as take-off, landing, hovering, and braking are realized through flight experiments, which verify the feasibility of the prototype design. |
|