English-version Introduction of My Graduation Project

w10d4

 

Autonomous Landing of a Quadrotor UAV

On an Intended Target

Keywords:  Computer Vision ; Path-planning; Quadrotor UAV;

SHU-WPI International Joint Project

 

A quadrotor UAV (unmanned aerial vehicle), also called quadrotor, is a multi-rotor helicopter that is lifted and propelled by four rotors. By controlling the rotational speed of the rotors, it is convenient to control its attitude in the space. In these last years, a growing interest has been shown in this kind of aerial robot platform. Researchers from prestigious universities and institutes like MIT and Stanford have made outstanding contributions on various areas including quadrotor structure, control algorithms for stability and path-planning. Now, this project is aiming at setting up a quadrotor platform and is committed to do some researches on autonomous landing skills of quadrotor.

 

 

Goal:  A circle will be drawn as a mark in outdoor environment. The quadrotor is expected to find it autonomously. After some calculations it will fly towards the mark, and eventually, landing on the mark precisely, safe and sound.

To reach this goal, there are two problems to solve as follows:

 

• How to build a quadrotor that is able to fly in a stable state?
• How to make the quadrotor recognize the target, flying towards it precisely, and landing safe and sound?

 

Technical Solutions:

About the first problem, thanks to the dedicated efforts in the early stage, ShanghaiUniversity has succeeded in installing the structure of a quadrotor, and configuring architecture of its electronics part. This project use FPGA as the main controller unit, and GY-86 module as the inertial measurement unit (IMU).The following are details of this structure.

 

Quadrotor Parameters:

Frame（450cm）

Brushless DC motor(7100r/min）

Lithium Battery(11.1v,2200mA) 

Propellers （1047inch）

Electronic Speed Controller  (30A) 

IMU sensors:  GY-86 

Sonar

Camera

FPGA Core Board:Altera FPGA

2.4G Remote Controller

Zigbee   module

IDE：NIOS II，QuartusII 11.0 (C is the main programming language)

[picture of the drone]

 

About the solution to the second problem，which is also the main focus of this SHU-WPI International Joint Project, camera is planned to use to locate target. As such, it needs to cover some technologies in the area of computer vision, such as how to dig useful information from pictures. If every thing is OK, we will finally make quadrotor find the target and finish the landing task quickly, precisely and stably.    The following is the flow chart, and we will design the specific algorithm during the period of this International Joint Project.

 

Discussion: When the camera is fixed to quadrotor，the quality of image will be truly influenced due to vibration. Here are two probable solutions: The first one is to correct the influences by digital filtering algorithms. The second one is to use a structure, which adds two servo motors to the camera (similar to PTZ camera) and let camera gain two degree of freedoms to balance vibration to meet the needs of the lowest precision requirements.

 

Conclusion:

       This project is mainly dedicated to explore a way of 3D-location and path-planning by camera. Thanks to the early stage, we also cover the modeling of quadrotor, stabilization of flight, and wireless signal transmission, which, of course, can be developed if any change could increase the whole performance of our quadrotor.

 

 

Reference:

[1]http://en.wikipedia.org/wiki/Quadcopter   wiki encyclopedia

[2] www.multiwii.com

[3]datasheets of  GY-86  module  (10DOF MS5611 HMC5883L MPU6050)

[4]Modelling,Identification and Control of a Qaudrotor Helicopter  Tommaso Bresciani

[5]Christian Schlaile, Oliver Meister, Natalie Frietsch, Christoph Kebler, Jan Wendel, Gert F. Trommer. Using natural features for vision based navigation of an indoor-VOTL MAV

[6]Tarek Hamela, Robert Mahony. Image based visual servo control for a class of arial robotic systems