Localization of an unmanned aerial vehicle in an indoor environment

• Main Author: Chen, Xing Yu
• Other Authors: Er Meng Joo
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/67850
• Institution: Nanyang Technological University

Unmanned aerial vehicle (UAV) is also called a drone. In contradiction to other types of aircrafts, it does not have a pilot on board however it can be controlled by several other modes instead, for instance, Remote Control mode is flown by human on the ground station or Autonomously flown by preprogrammed plan. For both of these control types, knowing the location of the aircraft is so important that it offers operator the control flexibility and information so as to make any corrections of the control command in real time. Meanwhile, constantly feedback the locations or the waypoints to the control units laid the foundation of Autonomous piloting activities and played an even critical role in Autonomous mode since the decision making is taken over by computer processers instead of human senses. In order to achieve fully autonomous, obstacle avoiding and self-localization needed to be constantly improve on as well. The most common UAV self-localization method is through GPS communication, it works perfect in open air condition where the communication signal is good. However it losses its capability significantly when GPS data is not available, indoor environment could be one of the scenarios which it faces such challenges. In this report, we will take a look at another auxiliary approach based on computer vision to serve as an alternative. It aims to computer the correlation of different picture in real time hence to get the amount of transform carried out from current position to previous waypoint and determine the current position if the starting point is known. The algorithm of the method will be discussed and details computer vision technic will be studies as well.