Self-Sufficient

Warehouse
Autonomous system for efficient warehouse product transfer and
organization

Supervisors:
Assoc. Prof. Abdelhady Ghanem
Eng. Ahmed Radwan
Eng. Omar Wahba

2023-2024
Outlines
Introduction 01
Mechanical Design 02
Hardware 03
Autonomous 04
Embedded systems 05
Internet of things 06
Business Case 07
Future work 08
Project Overview
An autonomous robot uses QR code
scanning to streamline warehouse
operations, transporting products to precise
storage locations efficiently and accurately.
Problem Statement
Modern warehouses need help with product sorting and
organizing inefficiencies, largely due to manual labor.

High labor costs

limited operational
hours

Impact on supply
chain
The supply chain is the network involved in producing and
delivering a product from source materials to the end user
Solution
Our project is developing an autonomous warehouse
robot to streamline transferring goods from the product
line to designated warehouse stands.

Key Features

Path Motion Control

Planning

QR Code Autonomous
Integration Operation
The process of products within a warehouse generally
encompasses several important stages

KITTING
KITTING PACKAGING
PACKAGING LABELING
LABELING SHIPPING
SHIPPING
PLC plays a crucial role in warehouse systems by accelerating
product movement throughout various processes

Our project initiates immediately

after products are labeled on the
linestream, ensuring seamless
progression
MECHANICAL
DESIGN
Mechanical Design
Overview
In this project, the design focuses on a smart
Warehouse robot.
Mechanical design is critical for creating efficient
and safe machinery.

Aim: To develop reliable and safe mechanical

systems.
 Concept

Planning
Design Process
We present step by step how
Create
the robot has designed

Testing

Refinement
• Parametric Modelling: Intelligent, adaptable designs
for iterative development.

• Assembly Modelling: Visualizes the integration of

components and mechanisms.

• Simulation and analysis: Tests and refines the

performance of robotic systems.
Movement of the
Robot
Movement of the Robot
Components:
• Two Wheels
• Motors
• Encoder
• Caster Wheels

Specification:
• Robot weight: 25.4 kg
• Lifted weight: up to 10 kg
Lifting Mech
Lifting Mech

•Components:
• Stepper Motor
• Lead Screws
• Gears, Belts, Pulleys
• Design Consideration:
• Load Calculation
• Torque Calculation
• Motor Selection
Lifting Mech

• Simulation of
slider plane
Sliding
Mechanism
Sliding
Mechanism
•Components:
• Motor
• Two Pulleys
• Flat Belt
• Guide Rails
• Design Consideration:
• System Layout
• Belt Installation
• Guide Rails and Sliding
Component
• Control System
HARDWARE
Main
Components

ESP8266
KITTING ENCODER
PACKAGING RASPBERRY
LABELINGPI 4 LIDAR
SHIPPING

MPU-6050
KITTING STM32
PACKAGING ULTRASONIC
LABELING KINECT CAMERA
SHIPPING
 Circuit
Design A B
• Design using Eagle Power Circuit Control Circuit

• Simulation using
rrcc
Proteus

• Protection C D
DC Motor Driver Stepper motor
Components Driver
 A: POWER
CIRCUIT

For Protection, the circuit has:

Fuse for current limit
Diode to protect from reverse
current
Capacitor for smoothing signal
•

•
•

•
Motor Drive works from 5v to 25 v and highlimit current up to 10A and peak
current 30A
 Stepper Motor
Drive

Peak limit Current

3.5A
Selectable micro-stepping (1/1, 1/2, 1/4, 1/8, 1/16 and
1/32 step) isolated
Optically
inputs
Over voltage, under-voltage, over-current, and short-circuit
protection
High starting
speed
High-speed
torque
AUTONOMOUS
The Necessity of Machines
Over Human Labor
Machines excel in tasks that require repetition,
extended working hours so that need to
ensure

Efficiency and High Accuracy

Productivity

Safety Cost-Effectiveness
Core idea of autonomous
robots Robots robots are capable of performing
Autonomous
complex tasks without human intervention

Decision Making Minimal User Advanced

Intervention Feedback
Mechanisms
To enable a robot to operate autonomously, we need to explore
its applications in four main fields

1 2 3 4
Perception SLAM Path planning Control
We used Gazebo simulation for its robust environment-building
and accurate physical action testing
 Perception between humans
and robots
Just as humans perceive their environment by using their
senses to receive external stimuli, robots can achieve the
same through sensors

These sensors allow robots to detect and respond to

external inputs, enabling them to act accordingly
 Sensor
Selection
To effectively operate the perception system for our
autonomous robot, we must meet several key requirements

Detect objects, measure dimensions Camera

and scan QRs

understand the robot's orientation IMU

and motion

Accurate visualising 2D mapping Lidar

For accurate distance measurement Encoder

 Data
Processing
Integrate multi-sensor data for a

Sensor fusion comprehensive environmental

understanding.

Real-Time Ensure the processing can be

done in real-time to maintain the
Processing robot's responsiveness.

Use advanced machine learning

Object Detection algorithms for object detection
and recognition
and Recognition
Kinect
Camera
Our camera provides two key features that enable our
robot to perform essential tasks.

Object QR Detection
Detection
 Object
Detection
To perform object detection we need manily labelled training
datasetand and training model

Automated object Manual Dataset

detection system and model
training
Automated object detection
system
We opted for YOLOv.8 because of its robust dataset and effective training model,but
The issue arose from the model's lack of effectiveness in simulation environments
Manual Dataset and model
training
For applying manual dataset training model we need to follow the leading steps:

1 2 3 4
Collect dataset Label dataset Segmentation Training
We manually gathered a dataset for training purposes from a
simulation environment
We labeled the dataset manually using LabelImg to label each
image with targeted object to be detected
The dataset was divided into training, testing, and validation
segments to ensure the quality of insurance training.
The dataset was ultimately trained using Google Colab for
model development and evaluation
The robot scans QR codes on products to retrieve position
information from the database, guiding it to the correct
placement.
 Decision
Making
When the robot reaches the required place it scans
again the qr to make sure that the robot reaches the
needed position

If the QRs match the robot places the product into

the required floor

If the QRs differ the robot places the product into

the return stock
 PID CONTROLLER
- PID is Proportional + Integral +
Derivative .The PID Controller is an element
used in the industry to adjust magnitudes such
as temperature, level, pressure or fl ow,
automatically and precisely, based on feedback
loop control.
 Importance PID

AUTOMATIC
REGULATION EQUIPMENT PERFORMANCE
OF PROTECTION INCREASE
MAGNITUDES

prevent equipment from acting in

Having different processes in a critical conditions that could cause The PID controls, together with the
production, the PIDs will be in damage (high temperatures, new data analysis systems, make it
charge of keeping the magnitudes pressures or intensity). A correct possible to change its behavior by
controlled around the setpoint adjustment of the PID parameters adjusting the key production
values entered, achieving the allows avoiding extreme measures setpoints in search of greater
desired conditions. at the moment in which the setpoint performance.
value is being reached.
 PATH PLANNING
Path planning is a crucial aspect of robotics and
autonomous systems, where the goal is to find a
suitable path for a robot or vehicle to navigate from
its current position to a target location while
avoiding obstacles. This process is typically divided
into two main components: global path planning and
local path planning.
.
 THE OVERALL PATH-PLANNING PROCESS
5-Step Ordering Process

Map
Global
Representation
Path
Planning

1 2 3 4 5

Perception local Path Execution

Planning
 G-SLAM
SLAM algorithms allow the
vehicle to map out unknown
environments. Engineers use
the map information to carry
out tasks such as path
planning and obstacle
avoidance.
The Slam Toolbox package
incorporates information
from laser scanners in the
form of a LaserScan message
and TF transforms from
odom->base link, and
 WHY G-SLAM
provide
high quality
use to
graph
Localization
and Mappin
easy to
easy to
install
use
44
3
2
1
EMBEDDED
SYSTEMS
AGV’s Problems
in Warehouse

Robot Collision

Feedback

Real-time
AGV’s Problems
in Warehouse

Robot Collision

Feedback

Real-time
Solution
In the automotive industry, Vehicle-to-
Everything (V2X) technology enables real-time
communication between vehicles,
infrastructure, and pedestrians, enhancing
safety and efficiency

Vehicles seamlessly can exchange critical

information

Position
Direction Speed
Direction Direction
Direction
V2X Communication Types

V2V V2I

V2G
V2G V2N
V2N V2P
V2P V2H
V2H
Vehicle to Vehicle
Wireless data exchange between nearby vehicles to
share information like location, speed, and braking status,
enhancing safety and traffic efficiency by anticipating
hazards

The key features of V2V (Vehicle-to-Vehicle) technology

Collision
Collision Sensor
Safety
Safety
Sensor Maneuverability
Maneuverability
avoidance
avoidance Awareness
Awareness
Vehicle to Infrastructure
Wireless communication between vehicles and
transportation infrastructure, allowing data exchange
with traffic signals, toll booths, and road signs for
improved traffic management

The key features of V2I (Vehicle-to-Infrastructure) technology

Data Flow Infrastructure Environmental

Direction
Managment Direction
Managment Direction
ingagement Direction
Monitoring
V2X SCENARIOS
Sudden
Suddenaction
action Parts
Partsfailure
failure Traffic Rerouting

Brake or swerve Failure in motor Traffi c systems

response to fallen cause undefi ned dynamically
object prompts behavior, sudden reroute to avoid
quick reaction. brake, or steering. congestion and
accidents.
Coding process

ESP826 STM32F1
6 and
Read sensors Controlxx
DC motors
decide V2X scenarios. and steppers.
 Reliable connection
DYNAMIC TOPOLOGY STATIC TOPOLOGY

• V2X communication • Wireless technologies in

technologies like DSRC warehouses, like RFID, Wi-Fi,
and C-V2X are tailored for and BLE, enhance inventory
safety-critical management, asset
communication in tracking, and operational
transportation systems. efficiency.
PROJECT STAGES

Containing
system
Ensure the
Suitable HW initialization ,f system performs
components low and its required
selected Including layer software functions
carefully architecture eff ectively and
and module Dynamic design effi ciently

HW Debug and test

Components
Static design
 HW COMPONENTS
STM32F103C8T6 Ultrasonic GPS Module MPU6050 ESP 8266

low power enables enables location- provide data on Aadding Wi-Fi

consumption and proximity based services for vehicle motion connectivity to
high-performance detection for navigation and and orientation for embedded
capabilities collision safety applications. enhanced safety system projects
avoidance. and navigation.
 Static
Design
Layered architecture
Including the required COTS
,modules and components
in the STD architecture.
Dynamic
Design
ESP8266 NodeMCU
flowchart
Dynamic
Design
STM32F103C8T6 MCU
flowchart
 Debug
and Test
Stage 1
“SW debugging”
 Debug
and Test
Stage 2
“Simulation testing”
 Debug
and Test
Stage 3
“HW testing”
INTERNET OF
THINGS (IOT)
What is IOT?

The Internet of Things (IoT) is a

network of interconnected devices
like sensors, gadgets, and
appliances that communicate and
exchange data via the internet.
Why IOT?

It enables machines to complete

tedious tasks without human
intervention

we can automate processes,reduce

labor costs,cut down on waste
and improve service delivery.
APPLICATION
 Main
elments
• Raspberry pi

• MQTT

• Node-Red

• ESP 8266
The brain of the operation is the Raspberry Pi. The Raspberry Pi
hosts Node‑REDsoftware and the MQTT broker
MQTT is a simple, low-bandwidth messaging protocol ideal for
data exchange between IoT devices
The broker stores uploaded data, and each device publishing or
subscribing to a topic that acts as a key to access the storage
WHY MQTT
Familiar with vehicular
environment.

Low latency

scalability

Reliabilty
 Node-
Red
Open-source

Visual programming
tool.

Saves time

scheduling algorithms
Node_Red interface
Node_Red interface
Robot UI
Warehouse UI
BUSINESS CASE
WAREHOUSE AREAS

The area of warehouses in Egypt typically

ranges from 5,000 to 10,000 square meters.

Additionally, there are larger warehouses

spanning from 10,000 to 30,000 square
meters.
 PROBLEMS

Intensifying number of
residence
working hours workers

Employees often work Some employees may The number of

up to 13 hours daily treat the workplace as workers in each
with just two breaks, their primary residence warehouse is very low
and may also be due to work demands, relative to the large
required to work which can lead to volume of work,
overtime or during iss ues like theft of prompting managers
offi cial holidays. elec tronic devices to add additional
stored there. working hours
Labour Force Affected by the Robot
AFFECTED NOT AFFTECTED

Forklift driver Warehouse

manager

Loader and Lead hand

unloader

Warehouse
specialist
 GOALS AND OBJECTIVES

01
Cost
Savings
02
Scalability

03
Adaptability to Technological
Advances
FUTURE WORK
 Future work

We aim to enhance our robot's

efficiency and versatility by focusing
on handling larger products, increasing
transfers per cycle, and optimizing
daily arrangements for cost and time
savings.
 Handling Larger Products

Increasing Product Throughput

How to do it?
Improving our robot can be
Optimizing Arrangement Processes
achieved in many ways.

Expanding Communication and

Coordination

Integration with Warehouse Management

Systems
Handling larger products
Adapt robot for larger products by upgrading lifting
mechanisms, motors, and structures. Refine
gripping and securing for stable transport.

MOTORS Materials Grip

Stronger motors like improved materials Modifying our

diesel motors or ac combination will robot’s grip will
motors can be used to increase our robot allow it to lift
impower the robot to lift strength so it doesn’t larger and heavier
higher weights fracture packages
Modification on planning
techniques
We will implement advanced path-planning
algorithms that enable the robot to navigate
freely throughout the warehouse-like
Dynamic A* algorithm

If a robot encounters a problem like a spill or

obstruction, it will alert the system so all
robots can use alternative routes until the
affected path is clear again.
Expanding
Communication a nd
Coordination
Enh ancing V2V co mmun icatio n
capabilities will be essential as we
scale up the n umber of ro bots
operating within the same
warehou se.

We' ll develop robust co mmun icatio n

proto co ls and centralized co ntrol
systems like “Lora”. Adding Swarm
System to enhance communication
between robots