School of Electrical and Computer Engineering: Addis Ababa Institute of Technology

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 11

Addis Ababa institute of Technology

School of Electrical and Computer Engineering

Stream Computer
Title Self Driving RC Toy Car
Semester/Year 1st Semester / 5th Year

Contact Person Mr. Kinde Mekuria


Advisor Mr. Biruk Tilahun
Signiture

Submission Date 07/11/2018

Students’ Name ID No
Feben Tesfaye ATR/1366/07
Terbinos Getachew ATR/9045/07

1|Page
Table of Content

Content Page Number


Abstract ……………………………………………………………3

Introduction ……………………………………………………….4

Problem Statement ………………………………………………...5

Objective ……………………………………………......................6

Methodology ………………………………………………………7

System Design ……………………………………………………..9

Timelining …………………………………………………………10

Reference …………………………………………………………..11

2|Page
Abstract
Our nation is progressing politically, socially and economically. Urban areas are soon to be
expanding and our cities are soon to be the center of attention from whole over the world
therefore we need to find methods to avoid road transportation and car accident problems. We
plan to build this prototype because we believe it will be a solution to the ever growing accidents
and further more encourage the concept of artificial intelligence in our nation.

Introduction

3|Page
Technology is progressing. From Data storage capacity, that enrich the reservoir of available
data as well as that of geographical and traffic data, to analysis capacity with processes such as
data mining and predictive algorithms, technology is making it possible to go beyond simple data
collection. Following that, the rapid growth of automation and artificial intelligence has impacted
the technological revolution we are in.Hence the combination of these innovative technologies is
gradually paving the way for the autonomous vehicle.

The benefits of autonomous vehicles can go way beyond removing the need of a human driver.
It can provide solutions to our basic social problems—the high cost of traffic crashes and
transportation infrastructure, the millions of hours wasted in traffic jams, and the wasted urban
space given over to parking lots. And most of all brings solution to the consistent car accidents
caused in our little nation.

4|Page
Problem statement
Around 2.6 million people are injured in vehicles every year. This is billions of dollars of car
repairs - in deductibles alone. If we could cut this by 25%, that’d be better than a typical tax
break. The major causes of these accidents are distraction, speeding, drunk driving etc … Self-
driving cars are dedicated to driving and can notice more, from all angles, and react more
quickly. No amount of text messages will have any effect on the car’s ability to stay focused on
the road.

Now imagine a world where the computers are driving all the cars. The only accidents will be
from car failures (if the wheel falls off it isn’t going to matter much who is driving), sensor
failures (the computer can’t see the road or other cars), or software bugs. The latter will only
happen once because it gets added to a regression test and all the cars running that software
automatically learn how to avoid that mistake. This project will create a proof-of-concept model
for a next generation car that will avoid crashes perform driving tasks semi-autonomously. Even
though this technology is completely new to Ethiopia why not try it here. It’s a more enjoyable
experience in addition to being safer.

5|Page
Objective
The purpose of this project is to create a small prototype of a self driving car which should be
able to drive itself without human intervention in the urban areas. According to the latest WHO
data published in 2017 road traffic accidents deaths in Ethiopia reached 27,140 or 4.27% of total
deaths. These accidents are majorly caused due to human error and a computer as a driver will
never make any error. Additionally this prototype can illustrate how problems like traffic jams
can be avoided as well as give hope for disabled people to enable them travel longer distance.

Not only so, artificial intelligence is a relatively new field to our team and we want to familiarize
with this important concept and somehow exercise using this project as a stepping stone.

6|Page
Methodology
This project involves high level of machine learning and python. Since the toy car will be driving
by itself , it has to trace a some specific path and follow that path along the way, that is Pattern
recognition.That highly involves Computer Vision and Artificial Intellegence. Hence, this
concept requires some machine learning skills, specifically Convolutional Neaural Network
Deep Learning, and ofcourse the project will be done using 100% python, no other programming
language will be used .

General flow of the projecct model:

Start the DC motor => Detect Path => Avoid Obstacle => Follow the path

Start the DC motor : the motor has to start first inorder for the toycar to move

Detect path: the toycar has to detect a path which in this case is a line to follow.

Avoid Obstacle: the toycar will be introduced to an obstacle along the way and it has to avoid the
obstacle and change direction and regain its previous path.

Finally when the line doesn’t exist anymore the toycar has to stop.

This project also involves several pieces of hardwares, softwares and ofcourse differerent
conceptual methodologies which are considered to be very helpful.

Softwares
PythonIDE and Jupyter: simple text editors which are very suitable for writing a pytho code.

OpenCV (Open Computer Vision): is an open source computer vision and also image processing
library. It converts images/videos into numeric matrices. There a lot of sub-modules in it and one
of the sub-modules, i.e haar cascades, can be used for object detection . So generally, if anyone
want to do anything with machine vision with python this library is the most used through out
the world and that is why we have choosen this library.

 Haar Cascades
1. Sub-module in OpenCV
2. Efficiently scan entire image , multiple sizes
3. Very fast

7|Page
Teensor Flow: is google’s open source machine learning library and has various flavors of
neural networks. The reason why this library will be used is that it is very fast, easy-to-use,
scalable and incredibly popular.

Hardwares
Here are some of the basics hardwares that we’ve planned to use. But these materials may be
subjected to change.

 Toy Car
 Raspberry Pi3
 Camera (Raspberry Pi camera)
 Micro SD card
 Wire
 External Battery
 Soldering Equipment

Flowchart for the self driving RC toy car

Toy car visual


Start Motor Starts processing

Direction Neural
Change Network
Direction Process

Stop

Stop Moving

8|Page
General System Design

Computer

Training Process

Training Output (Forward, Left, Right )

Control System

Feeds Pictures Raspberry Pi

Sensor Camera

Toy Car Feeds Direction

Raspberry

DC Motors

9|Page
Project Timeline(Semester-1)
Version – 1

Activity November Novembe December December January


13 – r 29 – 8– 17 – 11 –
November December December December January
28 7 16 24 25
Literature Review
Model Design
Model Train (Implementation)
Image Processing and Object
Detection Design
Image Processing Implementation

10 | P a g e
Reference
Documentation on Autonomous Driving Cars from Ponticifia University
web.mit.edu/6.111/www/f2010/projects

Center for Automotive Research on Self-driving cars: The next revolution

11 | P a g e

You might also like