Agile Quadrotor

Abstract

Due to their several applications, unmanned aerial vehicles (UAVs) have been a focus of research for the past few years. In Pakistan as well, some academic groups have shown interest in this area. However, we believe that the universities in Pakistan have not been able to utilize their full potential in UAV research. The research of UAVs can be boosted in Pakistan if a cheap, customizable and programmable UAV is available. The aim of this project is to develop such a Quadrotor, which is the most popular type of UAV in the micro to mini scale. There have been several other projects on Quadrotors in the past. However, our project has some significant differentiating features. First of all, most existing designs use processors with low computational power that are not suitable for research. Our design will be based on the high speed (and surprisingly inexpensive) ARM processor. The second main difference would be that our design will be customizable for different applications. The project poses some challenges in terms of weight reduction and controllability. For more controllability and stability, the weight should be as minimum as possible. The project includes optimized selection of all the components of Quadrotor. The essential components include controller, body frame, Motors, Propellers, Electric Speed Controllers (ESCs), Inertial Measurement Unit (IMU) and Battery etc. The control signals are sent by controller to the ESCs which in turn generate PWMs to run Motors. The IMU comprises of 3DOF Accelerometer to provide linear and gravitational acceleration and 3DOF Gyroscope to provide rotational velocity. This is used for sensing tilt in Quadrotor. The measurements of IMU are sent to the controller where they are filtered to get reasonable readings. The control system of Quadrotor requires synchronization of four individual motors which is done using closed loop PID control techniques and there is no harm in saying that it is quite a task to do. In the future, we need to develop a comprehensive mathematical model of Quadrotor using System Identification techniques.