In the past few years, countless public and private research developers have invested considerable amount of resources for contriving human-friendly unmanned aerial vehicles or drones. Within the platform, our aim is to tie a knot between technological development (e.g. robotics, Artific
Autonomus drone inspection
In the past few years, countless public and private research developers have
invested considerable amount of resources for contriving human-friendly
unmanned aerial vehicles or drones. Within the platform, our aim is to tie
a knot between technological development (e.g. robotics, Artificial
Intelligence, image processing) and application in spheres of society (e.g.
forestry, agriculture, rescue operation & others) in order to make them
informed about each other in a collaborative learning environment
The aim is to design an autonomous surveillance drone with capabilities
like system autonomy and data analysis.
To minimize the manual control of drone.
? To develop an automated technology which can be used in different
areas of its applications.
? To develop a drone capable of data collection and data analysis
Our project consists a drone kit, flight controller, logic unit and a camera.
Here our logic unit is the brain of our system which will compute logics
based on the images captured by our camera, after which commands will
be sent to our flight controller for further functionalities.
As per our plan first phase of our project consists our simulation on AirSim,
As AirSim uses Unreal plugin we will be building our environment on
Unreal Engine and then we will cascade AirSim with it. AirSim provides
us multiple built-in environments with Unreal and in order to develop our
drone within time we will be utilize those environments for the training of
our copter.
Though AirSim supports C++, C# and Java Api’s, the programming dialect
we choose for our computation scripts is the AirSim supported PythonApi
Page | 8
because of having good commands over the language. After setting up
everything and getting familiar with the environment our objective is to
achieve good commands over the controlling part of the drone for the
surveillance, once controlling part has been successfully achieved. Our
next objective will be modifying the script for image processing and
computer vision. This sums up our simulation part.
Once simulation part has been completed successfully, we will move
towards the second phase of hardware implementation. Hardware is a
critical phase as things work differently in real world. At initial stages we
will decide assembling our drone kit with all the components. Next step
will be the controlling of ‘BLDC (Brushless DC)’ motor using ‘ESC’s
(Electronic Speed Controller) ‘as ESC controls the volts provided to the
motor which ultimately controls the rpm. In general, a 1V change can
increase or decrease the rpm up to 1000 rpm. Next step will be the
connection of our RPi (Raspberry Pi) and FC (Flight Controller) through
MavLink, once this connection is established, we will be establishing
connection between RPi and our computer over Wi-Fi using SSH. At last
we will deploy our computation script on our RPi.
After conducting surveys and visiting different societal field we figured out
that most of them are still deprived of such advanced technology.
Developing an autonomous drone with surveillance and data collection
capabilities can help sectors like civil services, agriculture, rescue
operation teams, survey of remote or underdeveloped areas and others.
Furthermore, surveillance drones can be equipped with payloads (e.g.
Optical camera payload which provides images and videos, often in high
definition, thermal cameras can also be attached for night vision.
The main characteristic feature of our project is to work on the algorithm
in such a way that it will have full autonomous control with no need of
human interaction of any sort in the controlling sector. Here are some
features of our project.
i. Using machine learning algorithms in the best practical approach.
ii. Implementing autonomous flight control logic hence no manual
control is required.
iii. Drone is able to capture the images/video at any point which can be
shown to our ground control station.
iv. Implementing the concepts of computer vision and image processing
to detect the specified objects on which the drone is trained.
v. Capturing of live images/video will be used for surveillance of a
particular area.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| pixhawk | Equipment | 1 | 20000 | 20000 |
| Drone frame | Equipment | 1 | 5000 | 5000 |
| Electronic Speed Controllers | Equipment | 4 | 1000 | 4000 |
| Brushless DC Motors | Equipment | 4 | 1000 | 4000 |
| Raspberry Pi | Equipment | 1 | 7000 | 7000 |
| Battery | Equipment | 1 | 6500 | 6500 |
| Propellers | Equipment | 4 | 500 | 2000 |
| Miscellaneous | Miscellaneous | 1 | 5000 | 5000 |
| Total in (Rs) | 53500 |
Now a days, the demand of electronic motor control is increasing rapidly. Everyone need co...
Video Forgery Detection using Natural Scene Statistics All video forgery detection techniq...
my project is E-doctor, which is software based (android and IOS). patients use that softw...
Stabalization, command and control of two wheels robot is a challenging problem in automat...
Human civilization advance on a very basic principle that ?Necessity is the mother of inve...