Jomo kenyatta University of Agriculture and Technology

College of Engineering and Technology

School of Mechanical, Materials, and Manufacturing Engineering
Department of Mechatronic Engineering
————————————————————————————————

Project Proposal

Design and Development of a Cashier-less

Supermarket System

Orengo Lenny Weda (EN292-0688/2015)

Mwaniki Steve Mwenda (EN292-4120/2015)

Supervisors:
Dr. Rehema Ndeda
Dr. Evan Murimi

Technologist:
Mr. Joseph Matunda

Date:December 4th 2019.

————————————————————————————————
 I

Declaration

We hereby declare that the work contained in this report is original; researched and
documented by the undersigned students. It has not been used or presented elsewhere
in any form for award of any academic qualification or otherwise. Any material obtained
from other parties have been duly acknowledged. We have ensured that no violation of
copyright or intellectual property rights have been committed.

1. Orengo lenny Weda

Signature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Date. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2. Mwaniki Steve Mwenda

Signature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Date. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Approved by supervisors:

1. Dr.Rehema Ndeda

Signature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Date. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2. Dr. Evan Murimi

Signature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Date. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
 II

Abstract

The ever-increasing challenges in the retail sector especially increasing overhead costs
has necessitated the advancement of human-free and self-checkout technologies. However,
existing systems are plagued by high upfront and initial costs and inability to scale. We
therefore aim to develop a system that would be inexpensive to purchase and to run.

We intend to achieve this by designing and fabricating a conveyor system that would
facilitate transport of items, an item detection system to scan and identify the items,a
payment system to prompt and validate payment and an access control system to allow
access only after payment confirmation.Lastly, testing of the system functionality will be
done prior to demonstration.

We expect that items purchased by a customer will be accurately scanned and identified,
that the system will automatically display purchased items, validate payment and deliver
a receipt.
 I

Contents
Declaration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Aim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.4 Specific Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Amazon Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.3 Tap and Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.4 Scan and Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.5 Robotic checkouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.5.1 Gaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Mechanical module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2.1 Conveyor system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2.2 Belt Conveyors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2.3 Conveyor guard rails . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2.4 Stand for RFID reader, LCD display and keypad . . . . . . . . . . 11
3.2.5 Mounting base for microcontroller, motor driver,GSM module . . . 11
3.2.6 Access control mechanism . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 Electrical and Electromechanical design . . . . . . . . . . . . . . . . . . . . 11
3.3.1 Proximity Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3.2 RFID reader and RFID tags . . . . . . . . . . . . . . . . . . . . . . 12
CONTENTS II

3.3.3 Microcontroller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.4 PIR sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.5 Motors and motor driver . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.6 Power supply Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3.7 Display Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.3.8 GSM module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.4 Control Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.5 Testing and troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . 14

4 Expected outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
 III

List of Figures
2.1 Amazon-Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Tap and Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Scan and Go . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.4 Robotic Checkout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.1 Belt conveyor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Ultrasonic proximity sensor . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 PIR sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 GSM module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.5 Flow Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.6 Timeplan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.7 Budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Chapter 1. Introduction 1

1 Introduction

1.1 Background

Since the inception of the first supermarket in the 1930, the retail sector has undergone
various changes buoyed by advancement in technologies such as barcode technology and
the use of computers in Point of sale terminals. Currently, due to immense challenges
such as ever-increasing overhead costs and poor inventory management among others,
many technology firms and retail outlets, especially in the West and far-east are coming
up with human-free, self-service and self-checkout concepts. All this is aimed at not only
streamlining the check-out process but also tremendously reducing labour costs given that
on average, a single cashier in Kenya is paid on average Ksh. 25,000 a month. Therefore,
our challenge is to come up with an inexpensive, practical system that can be easily
adopted to serve customers at supermarket check-out points

1.2 Problem Statement

The retail sector continues to leverage technology in a bid to increase profit margins.
This has led to development of various approaches to the cashier-less concept not only to
enhance customer experience but also reduce labour costs that are inevitable when hiring
cashiers. However, adoption of the current available cashier-less technologies has been
limited by extremely high upfront costs and scalability issues. Therefore, there is need to
develop a cashless system that is inexpensive and easily adaptable.

1.3 Aim

Our aim is to develop an automated system that can automatically detect and identify
items put by a customer, compute and display total price, accept and confirm payment,
process a receipt and subsequently avail the items to the customer.
Chapter 1. Introduction 2

1.4 Specific Objectives

• To design the cashier-less supermarket system.

• To fabricate and assemble the cashier-less supermarket system.

• To test and troubleshoot the system.

Chapter 2. Literature Review 3

2 Literature Review

2.1 Overview

The dawn of self-service grocery shopping came in 1916 when the first piggy wiggly opened
in Memphis Tennessee. The customer had to hand the shopping list to a clerk who would
gather and bag the items, that was until the founder Clarence Saunders had the idea to
place items on shelves for shoppers to see and shop.

On August 4,1930, the world’s first supermarket, King Kullen[1], opened in Jamaica,
Queens, New York City. The supermarket was different from other grocery stores as it
offered a wider variety of products. These supermarkets and grocery stores tagged every
single item or kept a ledger on a nearby desk that would have the prices of everything for
sale. When someone brought it up, the cashier would input their prices and a till would
give its grand total. On 26th June 1974,the world’s first barcode was used to determine
the price of the items and update the stock data base[2]. The barcode system requires a
cashier to scan and receive money from the customers.

However, with the aforementioned traditional retail systems, which are still in use today,
the checkout process is manual. Given that the retail industry continually seeks to increase
profit margins, significantly reducing the cost of labour indeed will go a long way to relieve
finances[3]. Therefore, automation is being explored. Below are examples of existing
cashier-less systems:

2.2 Amazon Go

The cashier-less supermarket concept as implemented by Amazon-Go allows customers to

pick the items that they need and simply walk out the door. This completely eliminates
queues and cashiers

The system makes use of a combination of computer vision and different sensors to provide
Chapter 2. Literature Review 4

a learning algorithm that is supposed to recognize which product customers take off the
shelf and perhaps put back again.It relies on a technology similar to facial recognition that
recognizes objects from variety of angles and uses models such as RFID to determine the
respective position of customer and products, sensors are also used to detect the weight
of products on the shelf. Customers check in to the store via a mobile application and
QR code scan and invoicing is also done via the application.

This system is limited by the fact that it’s too expensive to actualize in full-size reg-
ular supermarkets. This is because the computing systems necessary to crunch all the
data required aren’t yet available at a scale and price that makes sense from an eco-
nomic standpoint.[4] Essentially, supercomputer-level power is needed per supermarket.
Additionally, only a certain number of people are allowed in the supermarket.

Figure 2.1: Amazon-Go

2.3 Tap and Go

Customers simply tap a pre-issued card on a shelf tag for items they want and some
minutes later, the customer’s bank account are automatically debited for the amount
they spend. This uses Near Field Communication (NFC) technology. NFC tags contain
chips within them that NFC readers and compatible mobile phones can read. In case the
Chapter 2. Literature Review 5

customer wishes to return the item, they can simply return the product and tap the card
again and the product will automatically be subtracted from the list.[5]

Figure 2.2: Tap and Go

2.4 Scan and Go

The stores are having portable barcode scanners that scans both the customer’s credit card
and products, once a person has all the goods, they want the customer simply chooses
check out on the scanner which automatically computes the price of the products and
subtracts cash from the credit card.

Some scan and go systems make use of a mobile application that allows the customer to
scan and pay via their mobile phones. One simply opens the application and chooses the
shop option then a bar code scanner pops up and allows the customer to scan the products
they buy, the application automatically computes the price of every item and grand total,
once the customer is done shopping they chose the checkout option and transactions are
automatically done via the application which then produces a barcode that one shows at
the exit.

The system has never really caught on because this transferred labour from cashiers to
customers leading to poor feedback and poor customer participation.
Chapter 2. Literature Review 6

Figure 2.3: Scan and Go

2.5 Robotic checkouts

Each item is equipped with an RFID tag. The customer simply picks and places the
items in a special basket.When done,one simply places the basket on the robotic check
out system that encloses the items, automatically reads and computes the price of all
items, once the customer pays, the robot bags the products and allows the customer to
pick them.

The robotic checkout system is limited by the fact that number of items is limited by the
size of the basket.

Figure 2.4: Robotic Checkout

Chapter 2. Literature Review 7

2.5.1 Gaps

• There is a definite need to cut down on labour costs through automation of the
checkout process.

• There is a need to come up with a system that will be significantly cheaper to

employ than high-end solutions like Amazon-Go.
Chapter 3. Methodology 8

3 Methodology

3.1 Background

Having highlighted the need of a cashier less supermarket system, this chapter seeks to
cover the manner in which the system will be crafted. This includes the proposed design
process, fabrication process and testing process of the system to meet the objectives laid
out.

3.2 Mechanical module

This entails the design and fabrication of the following parts:

• Conveyor system

• Items enclosure

• Conveyor guard rails

• RFID, Keypad and LCD stand

• Items enclosure structure

• Assembly stands

• Mounting base for microcontroller and, GSM module and motor driver.

• Access control mechanism

3.2.1 Conveyor system

This is the mechanism that facilitates the transport of items past the RFID scanner.
Conveyor systems are popular because their ability to carry heavy material, their simple
design and ease of maintenance. There are various types of conveyor systems such as:
Chapter 3. Methodology 9

• Roller conveyor systems

• Pneumatic conveyors

• Belt conveyors

• Slat conveyors

• Chain conveyors.

3.2.2 Belt Conveyors

They are the most commonly used powered conveyor systems. They comprise two or
more pulleys with an endless belt which is the carrying medium. The belt, which is made
from PVC or rubber, rotates about the pulleys to move the load.
Features and drawbacks:

• Relatively cost-effective.

• Suitable for batch type production process.

• Gentle product handling.

• Abrasive materials can damage the belt.

• May not be suitable for heavy loads.

Important factors that will be considered when designing the conveyor system include:

• Effective belt width for material

• Volume capacity of belt.

• Mass capacity of belt.

• The power requirements of belt conveyor with and without the load.
Chapter 3. Methodology 10

Figure 3.1: Belt conveyor

Materials that will be required to fabricate the conveyor:

• Bars to support the rollers in place via bolts and nuts and to fabricate stands for
the whole structure.

• Roller bearings.

• rollers.

• conveyor belt material

• Bolts and nuts.

The conveyor system will then be designed in CAD software and afterwards fabricated
and assembled in the JKUAT Engineering Workshops at Machine Shop.

3.2.3 Conveyor guard rails

This will ensure that items don’t fall on either side the conveyor belt. The guard rails
will be designed in CAD software and subsequently fabricated.
Chapter 3. Methodology 11

3.2.4 Stand for RFID reader, LCD display and keypad

This part supports the RFID reader and its antenna in place. On the other hand, the
LCD and keypad will be housed in a casing, attached onto the reader stand.

3.2.5 Mounting base for microcontroller, motor driver,GSM module

This will secure the microcontroller, motor driver and GSM module in place and thus
enhance aesthetics and cable management.

3.2.6 Access control mechanism

This will prevent a customer from exiting without paying for the items. An appropriate
control mechanism will be designed and fabricated to suit this function

3.3 Electrical and Electromechanical design

Entails the design/selection and assembly of the following:

1. proximity sensor

2. RFID reader and tags

3. Microcontroller

4. motors

5. PIR sensor

6. Power supply

7. LCD

8. Keypad
Chapter 3. Methodology 12

9. GSM module

10. Motor driver

3.3.1 Proximity Sensor

Will be used to control the movement of the conveyor, to determine the limit to which
the conveyor should reverse in case of a read failure and determine that there are no more
items on the conveyor.Sensors under consideration:

• Ultrasonic sensor.

• Digital infrared sensor.

Figure 3.2: Ultrasonic proximity sensor

3.3.2 RFID reader and RFID tags

Will be used to determine the product purchased by the customer, each item will be
fitted with a coded RFID tag that will have information about its price and type of
product, the RFID reader will then be used to read the data that will be decoded by the
microcontroller. Its location will have to be optimised to increase chances of read success.
Chapter 3. Methodology 13

3.3.3 Microcontroller

Responsible for processing all the sensor inputs and coordinating response through the
motors. factors for consideration:

• Number of I/O pins.

• Processing speed.

• Amount of flash memory.

3.3.4 PIR sensor

Useful in detecting the presence of a customer for access control.

Figure 3.3: PIR sensor

3.3.5 Motors and motor driver

The conveyor driver roller has to be actuated to transport items on the conveyor.Torque
and speed requirements will be considered. In addition,a motor will be required for the
access control system.
Chapter 3. Methodology 14

3.3.6 Power supply Unit

Will provide electric energy to power electrical and electromechanical components.Selection

consideration will be based on total power consumption of the system.

3.3.7 Display Unit

Will provide the interface between the customer and the system.The customer can there-
fore:

1. View the total price of items to be purchased.

2. Receive acknowledgement of payment.

3.3.8 GSM module

Will enable the customer to pay for the items to be purchased.In addition, it will send
the customer a receipt and a code showing that they have paid for the items.

3.4 Control Design

This entails the control algorithm for the system.

3.5 Testing and troubleshooting

After completing fabrication, the model system will be tested in the following ways:

1. The RFID detection system will be repetitively tested.

2. The conveyor speed will be optimized to a speed that will ensure the RFID tags are
read.
Chapter 3. Methodology 15

Figure 3.4: GSM module

3. The sensors will be tested separately and combined to ensure the system works as
expected.

4. The GSM module will be tested repetitively to ensure communication is consistent

and accurate.

5. The access control system will be tested for consistency of operation.

Chapter 3. Methodology 16

Figure 3.5: Flow Chart

Chapter 3. Methodology 17

Figure 3.6: Timeplan

Chapter 3. Methodology 18

Figure 3.7: Budget

Chapter 4. Expected outcomes 19

4 Expected outcomes

Upon completion of this project we expect that:

1. The RFID tags will be automatically read by the RFID reader

2. The sensors will control the conveyor movements as expected.

3. The GSM module will receive the payment message and send the customer a receipt
message.

4. The access control system will allow access to items only after payment is confirmed.
Chapter 5. Conclusion 20

5 Conclusion

Implementation of this project will effectively replace a cashier. However, it will be limited
to scanning an item at a time. Additionally, it will necessitate major supermarket outlets
to liaise with manufacturers to use RFID tags in place of barcodes.
REFERENCES 21

References

[1] N. geographic, “resource library,first supermarket opens,” 2013.

[2] T. L. Harry E. Burke, Automating Management Information Systems: Barcode Engi-

neering and Implementation, 2004.

[3] C. Orwa, “Kenya supermarket report,” Life in Data points, 2017.

[4] J. Wells, “From smart carts to computer vision, checkout innovations are multiplying,”
2019.

[5] L. Bandoim, “Cashierless shopping with ’tap to go’ technology is coming to more
grocery stores,” 2018.