Communication Delay Modeling and Prediction for Edge-enabled UAV Control in Cellular Networks

• Main Authors: Wu, Yu-Hsuan, 吳宇軒
• Other Authors: Wang, Kuo-Chen
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/d7ah4s

碩士 === 國立交通大學 === 網路工程研究所 === 107 === For vision based Unmanned Aerial Vehicle (UAV) applications, such as object tracking, the Ground Control Station (GCS) will command UAV movements according to the processing results of video frames sending from the camera on the UAV. However, there is a time delay between the capturing of video frames and receiving of the corresponding control command from the server in the GCS. The time delay will cause the control command outdated and may make the UAV deviate from the scheduled tracks or collide with obstacles. The main goal of this thesis is to shorten the time delay of the control command in LTE wireless cellular networks and predict the range of possible positions where the UAV receives the control command. We have applied the Multi-access Edge Computing (MEC) technology in cellular networks to shorten the time delay of the control command. Specifically, we found the Probability Density Function (PDF) of the time delay of the control command in the MEC network follows a mixture of Erlang distribution. According to the findings, we proposed an analysis model to predict the range of possible positions where the UAV receives the command so that the UAV can be controlled more precisely. In addition, we developed an event-driven simulator to validate the proposed analysis model.