Autonomous quadcopter cartographer

Abstract

Research into autonomous navigation of Micro-Aerial Vehicles (MAVs) has become increasingly popular in recent years. Outdoor autonomous features rely mainly on GPS technology to estimate its position, orientation and velocity. Due to the weak GPS signals in indoor areas, the same method cannot be used for indoor environments. This is a combined project among three Renaissance Engineering Programme (REP) students aiming to design and create a quadcopter capable of performing autonomous SLAM and exploration algorithms. All three members will be focusing on three different aspects of the quadcopter project. At the end, the different contributions made in those three aspects should culminate together as much as possible into an integrated product which aims to meet the objectives of this project. This report covers details of essential hardware that is mounted on the quadcopter. These include a LIDAR laser scanner, a pcDuino development board for high-level computations, a Pixhawk flight controller module for low-level processing, and a power distribution board. In addition, this report also covers the development process and implementation of a quadcopter platform that uses a LIDAR sensor for navigation. The two main control frameworks of a quadcopter that will be focused on in depth is the altitude and position control. As there are various methods that can achieve successful implementation of the two control frameworks, this report aims to find the most suitable one for application to this project. Test flights are also carried out to ensure that theories on paper are proven to be true in real-world conditions as well. Other aspects of the project will also be mentioned briefly in this report. One of them includes an investigation of the possibility of using a ducted frame design to improve thrust performance, while another presents an overview of SLAM and selfexploration methods to be used in indoor GPS-denied environments.