1. technology and computing

Leader-follower Cart Using Bluetooth and
Infrared Technology with Collision Detection
by
Ronnel Angelo M. Bajon
Billy Fernand P. Macatangay
Darwin Christopher R. Tantuco
John Eldrin M. Tolentino
Bachelor of Science in Computer Engineering
Mapúa Institute of Technology, 2013
A Thesis Report Submitted to the School of EECE
in Partial Fulfillment of the Requirements for the Degree of
Bachelor of Science in Computer Engineering
i
APPROVAL SHEET
This is to certify that we have supervised the preparation of and read the thesis paper
prepared by Ronnel Angelo M. Bajon, Billy Fernand P. Macatangay, Darwin
Christopher R. Tantuco, and John Eldrin M. Tolentino entitled Leader- Follower
Cart Using Bluetooth and Infrared Technology with Collision Detection and that the
said paper has been submitted for final examination by the Oral Examination Committee.
Voltaire B. De Leon
Academe Adviser
Analyn N. Yumang
Thesis Adviser
As members of the Oral Examination Committee, we certify that we have examined this
paper and hereby recommend that it be accepted as fulfillment of the thesis requirement
for the Degree Bachelor of Science in Computer Engineering (major).
Joshua B. Cuesta
Panel Member 1
Jose B. Lazaro
Panel Member 2
Dionis A. Padilla
Committee Chair
This thesis is hereby approved and accepted by the School of Graduate Studies as
fulfillment of the thesis requirement for the Degree Bachelor of Science in Computer
Engineering (major).
Dr. Felicito S. Caluyo
Dean, School of EECE
ii
ACKNOWLEDGEMENT
We are very thankful to all the people that motivated and helped us in
accomplishing our thesis. We would like to thank our adviser, Analyn M. Yumang, for
her assistance, support, patience, and advices. We are also grateful to our course
professor, Voltaire De Leon, for being a dependable professor that guided us through the
prototype and document stage of our thesis.
We would also like to thank Lord God for his continuous guidance and help.
iii
TABLE OF CONTENTS
TITLE PAGE
i
APPROVAL PAGE
ii
ACKNOWLEDGEMENT
iii
TABLE OF CONTENTS
iv
LIST OF TABLES
vi
LIST OF FIGURES
vii
ABSTRACT
ix
Chapter 1:
INTRODUCTION
1
Chapter 2:
REVIEW OF LITERATURE
3
Human-follower Device
Device positioning technique
Radio Frequency Identification System
RFID Tags
RFID Reader
RFID Antennas and Radio
Bluetooth
Comparison between Bluetooth, Zigbee, and Wi-Fi
Infrared LED and Infrared Detector Sensor
Collision Avoidance
Infrared Proximity Sensors
Weight Reading
Load Cell
Lead Acid Battery
3
Chapter 3:
4
7
11
14
16
17
LEADER-FOLLOWER CART USING BLUETOOTH AND
INFRAREDTECHNOLOGY WITH COLLISION DETECTION
18
Abstract
Introduction
Methodology
Block Diagram
System Flowchart
Results and Discussion
18
18
20
22
26
31
iv
Chapter 4:
References
Conclusion
44
45
Chapter 5:
Recommendation
46
APPENDICES
47
Appendix A. Statistical Computation
Appendix B. Schematic Diagram
v
47
56
LIST OF TABLES
Table 3.1: Test for Reliability of RFID
31
Table 3.2: Cart’s Movement following the User’s Movement
34
Table 3.3: Summary of Table 3.2
35
Table 3.4: Tally of the Test for Collision Detection
38
Table 3.5: Data of Theoretical and Actual Weight Reading
41
vi
LIST OF FIGURES
Figure 2.1: A Bluetooth scatternet with two types of relay nodes
9
Figure 2.2: IR sensor and its schematic diagram
12
Figure 2.3: Sample control pulses of an IR LED
13
Figure 2.4: IR Detector Sensor
14
Figure 2.5: Proximity Sensor Response Curves
16
Figure 3.1: System Overall Block Diagram
22
Figure 3.2: Cart’s Components – Top View
24
Figure 3.3: Cart’s Components – Side View
24
Figure 3.4: User Wearable Belt
25
Figure 3.4: User Wearable Belt
25
Figure 3.6: Flowchart of Enabling the Cart
26
Figure 3.7: Flowchart of the Weight Reading and Collision Detection
27
Figure 3.8: Flowchart of the Cart’s Movement
28
Figure 3.9: Sample Screenshot for Unknown Tag
32
Figure 3.10: Sample Screenshot for User Tag
32
vii
Figure 3.11: Sample Screenshot of the Cart following the User
35
Figure 3.12: Testing Collision Detection at the Right
39
Figure 3.13: Testing Collision Detection at the Left
39
Figure 3.14: Testing Collision Detection at the Bumper
40
Figure 3.15: Sample Screenshot for the Obtained Weight using the Cart
42
Figure 3.16: Sample Screenshot for the Theoretical Weight using Weight Balance
43
viii
ABSTRACT
TRANSPORTING OBJECTS USING VEHICLES ARE USED BY EITHER
PUSHING OR DRAGGING. THIS WAY OF TRANSPORTING THINGS
REQUIRES A LOT OF EFFORT AND COULD BE DANGEROUS TO THE
PEOPLE USING THEM. ACCORDING TO THE HEALTH & SAFETY
EXECUTIVE (HSE) MORE THAN ONE THIRD OF ALL ACCIDENT INJURIES
WARRANTING MORE THAN 3 DAYS ABSENCE FROM WORK, INVOLVE
LIFTING OR MANUAL HANDLING. WITH AN INTELLIGENT TRANSPORT
VEHICLE, MOVING AN OBJECT FROM ONE PLACE TO ANOTHER CAN BE
DONE WITH LESS EFFORT. THIS VEHICLE IS CAPABLE OF FOLLOWING
A USER WHILE IN MOTION WITH CONSTANT SPEED. IT
AUTOMATICALLY MOVES ITSELF IN DIFFERENT DIRECTION GUIDED
SENSORS DETECTING A DEVICE CARRIED OR ATTACHED TO A PERSON
IT “FOLLOWS”. MOREOVER, IT CAN DETECT NEARBY OBJECTS THAT
COULD CAUSE COLLISIONS TO THE DEVICE AND IT COULD ALSO
MEASURE THE WEIGHT OF THE OBJECT/S IT CARRIES. THIS DEVICE IS
APPLICABLE TO OFFICES, LIBRARIES, HOSPITALS, AIRPORTS AND IN
CERTAIN BUILDING CONSTRUCTION OPERATIONS.
Keywords: leader-follower, transport, objects
ix
Chapter 1
INTRODUCTION
Manual handling is any transporting or supporting of a load by one or more
workers. It includes the following activities: lifting, holding, putting down, pushing,
pulling, carrying or moving of a load parts (European Agency for Safety and Health at
Work, 2007). Manual transportation of objects is indeed a part of humans’ lives.
Although technology exponentially arises and innovates overtime, manual handling is
still practiced in workplaces: in factories and building sites, shopping malls, offices,
airports and hospitals, from the very beginning until now. As a result, manual handling
injuries are consistently and apparently evident.
Based on the research, more than a third of all over-three-day injuries reported
each year are caused by manual handling - the transporting or supporting of loads by
hand or by bodily force (Op De Beeck, R. and Hermans, V., 2000). Furthermore, about a
fourth of European workers suffer from back pain, which tops the list of all reported
work-related disorders. Loads that are too heavy, too large, difficult to grasp, unbalanced
and unstable and difficult to reach and tasks that are too strenuous and involves awkward
postures and movements are the factors that make manual handling hazardous. As a
result, manual handling can lead to negative health effects such as fatigue, injuries of the
back, neck, shoulders, arms or other body parts (European Agency for Safety and Health
at Work, 2007).
Because of all these negative health effects due to manual handling, prevention
measures were offered such as: having a general risk assessment guidelines and conduct
1
risk assessment of manual handling activities, conduct trainings and learn good handling
techniques for lifting, pushing and pulling (Health and Safety Executive, 2008)
(European Agency for Safety and Health at Work, 2007). However, those accidents are
still inevitable and the present methods of prevention will not guarantee the full safety of
the people.
With this in mind, the researchers proposed a study that aims to develop a
system of a cart capable of following a single person. Specifically, the cart is able to
identify the user and when successfully identified, the cart is able to follow that person in
any direction this user will go. And to complete its functionality, the cart is able to detect
collisions along the way and can determine the weight of the things that it can carry.
This proposed device is used to eliminate the risks of manual handling.
Especially for the accidents that may occur when manually transporting objects from one
location to another (Health and Safety Executive, 2008). Aside from this, it is also
essential to emphasize the huge advantage of automatic transporting of objects over the
manual. The techniques used for collision detection and user identification in the cart
were derived from the previous researches’ techniques but of different approach.
The study covers the development of an automated cart capable of following a
single person. The system will include sensors that have certain maximum and minimum
detection range. Also, the cart could only travel along flat surfaces and inclined or uneven
platforms are outside the scope of this study. The vehicle could carry a maximum weight
load of 40 kilograms. The maximum distance from the cart to the person would be 1
meter.
2
Chapter 2
REVIEW OF RELATED LITERATURE
Human-follower Transporting Device
A human-follower transporting device, typically associated with robots, is a
vehicle capable of carrying a load and travelling on its own and moving along a direction
towards a human, in motion or not. This technology comes from the leader-follower
concept where most of the work in this area focuses on control aspects related to steering
an unmanned follower vehicle so that it follows a lead vehicle (Borenstein, J., et al,
2010). This type of device is applied on the field of robotics and is used for human
assistance and general services such as robotic guides (Kulyukin, V., et al, 2004) and
military mules (Borenstein, J., et al, 2010).One important item a leader-follower device
needs to have is that it requires the knowledge of the position of itself and the followed
human. Researchers and engineers have developed a variety of systems, sensors, and
techniques for mobile device positioning including odometry, magnetic compasses,
active beacons, global positioning systems and landmark navigation (Borenstein J., et al,
1997).
Device positioning techniques
There are two basic ways of measuring position: absolute and relative. Absolute
positioning, also called reference-based system, uses a fixed referential system (e.g.,
maps, GPS or ultrasonic beacons placed at known positions in the operating
3
environment) and requires attaching instruments or mapping the environment while
relative positioning, most commonly known as dead-reckoning, compares positions
expressed in relation to a common reference frame (Fr´ed´eric R., et al, 2008). Previous
researches develop follower robots with relative positioning technique and referencing
the object being followed since its position relative to the robot is used for the algorithm
required for displacing the robot.
Relative positioning uses components such as sensors or cameras in determining
the leader’s location. An example is the study Ultrasonic Relative Positioning for MultiRobot Systems (Fr´ed´eric R., et al, 2008) where the receivers are mounted on the mobile
robot so that the reference frame is the robot itself and the device can determine the
position of a transmitter located on another robot based on distance measurements
returned by the receivers. One does not simply apply this kind of relative positioning for
a situation where multiple robots on a single area are following different leaders for this
may cause errors where a certain robot follows a wrong leader. To solve this problem,
researchers use leader identification techniques such as cameras which are used for face
recognition (Braun T., et al, 2005).
Radio Frequency Identification System
In recent years, radio frequency identification technology has moved from
obscurity into mainstream applications that help speed the handling of manufactured
goods and materials. RFID enables identification from a distance, and unlike earlier barcode technology it does so without requiring a line of sight (Roy W., 2006).
4
RFID tags
The tag is the basic building block of RFID. Each tag consists of an antenna and
a small silicon chip that contains a radio receiver, a radio modulator for sending a
response back to the reader, control logic, some amount of memory, and a power system.
The power system can be completely powered by the incoming RF signal, in which case
the tag is known as a passive tag. Alternatively, the tag’s power system can have a
battery, in which case the tag is known as an active tag (Garfinkel, S. and Henry H.,
2005).
The primary advantages of active tags are their reading range and reliability.
With the proper antenna on the reader and the tag, a 915MHz tag can be read from a
distance of 100 feet or more. The tags also tend to be more reliable because they do not
need a continuous radio signal to power their electronics. Passive tags, on the other hand,
can be much smaller and cheaper than active ones because they don’t have batteries.
Another advantage is their longer shelf life: Whereas an active tag’s batteries may last
only a few years, a passive tag could in principle be read many decades after the chip was
manufactured. Between the active and the passive tags are the semi-passive tags. These
tags have a battery, like active tags, but still use the reader’s power to transmit a message
back to the RFID reader using a technique known as backscatter. These tags thus have the
read reliability of an active tag but the read range of a passive tag. They also have a
longer shelf life than a tag that is fully active (Garfinkel, S. and Henry H., 2005).
Tags come in all shapes and sizes but RFID tags can be promiscuous, in which
case they will communicate with any reader. Alternatively, they can be secure, requiring
that the reader provide a password or other kind of authentication credential before the
5
tags respond. The simplest RFID chips contain only a serial number—think of this as a
64-bit or 96-bit block of read-only storage. Although the serial number can be burned
into the chip by the manufacturer, it is also common for the chips to be programmed in
the field by the end user (Garfinkel, S. and Henry H., 2005).
RFID reader
The RFID reader sends a pulse of radio energy to the tag and listens for the tag’s
response. The tag detects this energy and sends back a response that contains the tag’s
serial number and possibly other information as well. In simple RFID systems, the
reader’s pulse of energy functioned as an on-off switch; in more sophisticated systems,
the reader’s RF signal can contain commands to the tag, instructions to read or write
memory that the tag contains, and even passwords. Historically, RFID readers were
designed to read only a particular kind of tag, but so-called multimode readers that can
read many different kinds of tags are becoming increasingly popular. RFID readers are
usually on, continually transmitting radio energy and awaiting any tags that enter their
field of operation. However, for some applications, this is unnecessary and could be
undesirable in battery-powered devices that need to conserve energy (Garfinkel, S. and
Henry H., 2005).
Like the tags themselves, RFID readers come in many sizes. The largest readers
might consist of a desktop personal computer with a special card and multiple antennas
connected to the card through shielded cable. Such a reader would typically have a
network connection as well so that it could report tags that it reads to other computers.
6
The smallest readers are the size of a postage stamp and are designed to be embedded in
mobile telephones (Garfinkel, S. and Henry H., 2005).
RFID Antennas and Radio
The RFID physical layer consists of the actual radios and antennas used to
couple the reader to the tag so that information can be transferred between the two.
Radio energy is measured by two fundamental characteristics: the frequencies at which it
oscillates and the strength or power of those oscillations. As with most radio systems, the
larger the antenna on the reader and the tag, the better an RFID system will work because
large antennas are generally more efficient at transmitting and receiving radio power than
are small antennas. Thus, a large antenna on the reader means that more power can be
sent to the RFID tag and more of the tag’s emitted energy can be collected and analyzed.
A large antenna on the tag means that more of the power can be collected and used to
power the chip. Likewise, a large antenna on the chip means that more power can be
transmitted back to the reader (Garfinkel, S. and Henry H., 2005).
Bluetooth
Bluetooth is emerging as an important standard for short range, low-power
wireless communication. It provides a decentralized communication substrate that
standardizes the link-layer medium access (MAC) and physical layer functionalities of
the traditional networking protocol stack. It operates in the 2.4 GHz frequency band
employing a pseudo-random frequency hopping scheme (Godfrey T., et al, 2001).
7
The Bluetooth MAC protocol is designed to facilitate the construction of ad hoc
networks without the need for manual configuration, cables, or wired infrastructure. It is
based not on distributed contention resolution, as in traditional wireless LANs, but on a
master-slave mechanism. A Bluetooth piconet consists of one master and up to seven
slaves. The master allocates transmission slots 1 (and therefore, channel bandwidth) to
the slaves in the piconet. The basic idea is for the master and slaves to use alternate
transmission slots, with each slave slot (an odd-numbered slot, by convention) being used
only by the slave to which the master sent a frame in the previous (even-numbered)
transmission slot. This MAC protocol is an example of a time division duplex (TDD)
scheme. Frequency hopping allows multiple concurrent Bluetooth communications
within radio range of each other,without adverse effects due to interference. This
facilitates high densities of communicating devices, making it possible for dozens of
piconets to co-exist and independently communicate in close proximity without
significant performance degradation. This raises the possibility of internetworking
multiple piconets. The Bluetooth specification alludes to this possibility, calling it a
scatternet, but does not specify how it is to be done (Godfrey T., et al, 2001).
8
Figure 2.1: A Bluetooth Scatternet with Two Types of Relay Nodes:
Node 1 is a “Slave Relay”, Node 2 is a “Master Relay”.
An obvious starting point is to judiciously choose nodes, such as nodes 1 and 2
in Figure 1, to participate as relays in multiple piconets, forwarding data between
piconets. Since two slave nodes cannot be linked together directly, the path of a packet
must alternate between master and slave nodes, until it reaches its final destination. While
the basic idea is simple enough, a number of challenging problems need to be solved
before this can become a reality (Godfrey T., et al, 2001).
The link formation process specified in the Bluetooth baseband specification
consists of two processes: inquiry and page. The goal of the inquiry process is for a
master node to discover the existence of neighboring devices and to collect enough
information about the low-level state of those neighbors (primarily related to their native
clocks) to allow it to establish a frequency hopping connection with a subset of those
neighbors. The goal of the page process is to use the information gathered during the
9
inquiry process to establish a bi-directional frequency hopping communication channel
(Godfrey T., et al, 2001).
During the inquiry process, a device enters either the INQUIRY or the
INQUIRY SCAN state. A device in the INQUIRY state repeatedly alternates between
transmitting short ID packets containing an Inquiry Access Code (IAC) and listening for
responses. A device in the INQUIRY SCAN state constantly listens for packets from
devices in the INQUIRY state and responds when appropriate. The Bluetooth
specification states that a node in the INQUIRY state devotes sufficient amount of time
transmitting and listening whereas a node periodically enters the INQUIRY SCAN state
to scan continuously over a short window (Godfrey T., et al, 2001).
The Bluetooth specification assumes that each node knows whether it is to be a
master or a slave. The need for manual configuration of master or slave roles is
unattractive when more than a few nodes are attempting to form a connected scatternet in
an ad hoc fashion. To deal with this problem, the Bluetooth specification provides a Host
Controller Interface (HCI) specification that provides a standardized method of accessing
the Bluetooth baseband capabilities. This interface can be used to implement various
topology formation schemes (Godfrey T., et al, 2001).
Comparison between Bluetooth, ZigBee, and Wi-Fi
Bluetooth, also known as the IEEE 802.15.1 standard is based on a wireless
radio system designed for short-range and cheap devices to replace cables for computer
peripherals, such as mice, keyboards, joysticks, and printers. This range of applications is
10
known as wireless personal area network (WPAN). Two connectivity topologies are
defined in Bluetooth: the piconet and scatternet (Jin-Shyan L., et al, 2007).
ZigBee over IEEE 802.15.4 defines specifications for lowrate WPAN (LRWPAN) for supporting simple devices that consume minimal power and typically operate
in the personal operating space (POS) of 10m. ZigBee provides self-organized, multihop, and reliable mesh networking with long battery lifetime [8-9]. Two different device
types can participate in an LR-WPAN network: a full-function device (FFD) and a
reduced-function device (RFD) (Jin-Shyan L., et al, 2007).
Wireless fidelity (Wi-Fi) includes IEEE 802.11a/b/g standards for wireless local
area networks (WLAN). It allows users to surf the Internet at broadband speeds when
connected to an access point (AP) or in ad hoc mode. The IEEE 802.11 architecture
consists of several components that interact to provide a wireless LAN that supports
station mobility transparently to upper layers (Jin-Shyan L., et al, 2007).
Infrared LED and Infrared Detector Sensor
Infrared (IR) sensors are extensively used for measuring distances. Therefore,
they can be used in robotics for obstacle detection or avoidance. They are cheaper in cost
and faster in response time than ultrasonic (US) sensors. However, they have non-linear
characteristics and they depend on the reflectance properties of the object surfaces. So
knowledge of the surface properties must be known prior. In other words, the nature in
which a surface scatters, reflects, and absorbs infrared energy is needed to interpret the
sensor output as distance measure. IR sensors using reflected light intensity to estimate
the distance from an object are reported in the bibliography. Their inherently fast
11
response is attractive for enhancing the real-time response of a mobile robot. Some IR
sensors described in the bibliography are based on the measurement of the phase shift,
and offer medium resolution from 5 cm to 10 m, but these are very expensive
(Mohammad T., 2009).
Figure 2.2: IR Sensor and Its Schematic Diagram
An infrared emitter, or IR emitter, is a source of light energy in the infrared
spectrum. It is a light emitting diode (LED) that is used in order to transmit infrared
signals from a remote control. In general, the more they are in quantity and the better the
emitters are, the stronger and wider the resulting signal is. A remote with strong emitters
can often be used without directly pointing at the desired device. Infrared emitters are
also partly responsible for limits on the range of frequencies that can be controlled. An IR
emitter generates infrared light that transmits information and commands from one
device to another. Typically one device receives the signal then passes the infrared (IR)
signal through the emitter to another device (Mohammad T., 2009).
The IR signal emitted from a remote control caries the information needed to
control the appliance. This signal consists of pulses that code 0 and 1 bits, instructing the
12
appliance to do a certain operation. One of the most common protocols used to code the
IR signal is Philips - RC5 protocol. The signal consists of two parts, the control pulses
and the carrier wave as seen in the image below (Mohammad T., 2009).
A common frequency used for the carrier is 38 KHz and control pulses
frequency is in the range of 1-3 KHz. The carrier signal is modulated by the control
pulses and the resulting signal is emitted by remote in IR band of electromagnetic
spectrum. IR band is invisible to human eye. You can see if an IR led is emitting light or
not using a camera. Point the camera to the led and you will see that light comes off
(Mohammad T., 2009).
Figure 2.3: Sample Control Pulses of an IR LED
IR detectors are little microchips with a photocell that are tuned to listen to
infrared light. They are almost always used for remote control detection. Every TV and
DVD player has one of these in the front to listen for the IR signal from the clicker.
Inside the remote control is a matching IR LED, which emits IR pulses to tell the TV to
turn on, off or change channels. IR light is not visible to the human eye, which means it
13
takes a little more work to test a setup. Infrared detectors are also LEDs. They are wired
differently so that they convert incoming infrared light to an electric current. The current
is sent to a device that reads the current to determine the strength of incoming light, or to
interpret signals meant to control a television, for example. IR detectors can be made
more sensitive through electronics. Amplifiers make IR detectors extremely sensitive so
that even very faint signals are recorded. Sensitivity can be adjusted so a detector is
suitable for its anticipated purpose (Mohammad T., 2009).
Figure 2.4: IR Detector Sensor
Collision Avoidance
Collision avoidance is a technology used for the prevention of any contact from
a device to another. Real-time obstacle avoidance presents the problem of navigating
around unknown objects in a dynamic environment. The virtual force field (VFF) method
described by Borenstein uses an occupancy grid to determine the position of obstacles in
14
the environment. A VFF is then calculated from these obstacles and the robot is ―pushed‖
away from them (Chris G. and Mohan T., 1994).
Infrared proximity sensors
The infrared proximity sensors used consist of one or more infrared light
emitting diodes (emitters) and a single silicon phototransistor (receiver). The electrical
signal given by the receiver depends on how much of the incident ray of the emitter(s) is
reflected back and detected by the receiver. The signal from the receiver is first preamplified before it is low-pass filtered to eliminate high frequency noise. The sensor's
signal output, when no object is near the sensor, is offset to zero. The signal is then
amplified to a 0-10 volt full scale range (D.J. Balek, 1985).
Based on the research, infrared-proximity sensors were used to determine the
approximate range and bearings of nearby objects; while the system was effective for the
experiment presented in the work, it would be very susceptible to interference from any
other obstacles and not appropriate for general experiments (Jim P., et al, 2008).
A typical signal strength response of the proximity sensor as a function of the
normal distance from a work piece is shown in Fig. 1. The solid curve is the response of
the sensor to a work piece with a smooth surface. The dotted curve corresponds to the
response of the same proximity sensor to a work piece with a rougher surface quality.
The proximity sensor's response versus the angle of a work piece’s surface normal is also
shown in Fig. 1 (D.J. Balek, 1985).
15
Figure 2.5: Proximity Sensor Response Curves
Weight Reading
This is one additional feature of this device: that it’s able to measure the weight
of the objects that it will carry.
Load Cell
A load cell is a transducer which converts force into a measurable electrical
output. Load cells are designed to sense force or weight under a wide range of adverse
conditions; they are not only the most essential part of an electronic weighing system, but
also the most vulnerable. In order to get the most benefit from the load cell, the user must
have a thorough understanding of the technology, construction and operation of this
unique device. In addition, it is imperative that the user selects the correct load cell for
the application and provide the necessary care for the load cell during its lifetime (Revere
Transducers, 2001).
16
Lead Acid Battery
This serves as the source of power supply of the device. Lead acid charging uses a
voltage-based algorithm that is similar to lithium-ion. The charge time of a sealed lead
acid battery is 12–16 hours, up to 36–48 hours for large stationary batteries. With higher
charge currents and multi-stage charge methods, the charge time can be reduced to 10
hours or less; however, the topping charge may not be complete. The disadvantage of a
lead acid battery is that it is sluggish and cannot be charged as quickly as other battery
systems.
17
Chapter 3
LEADER-FOLLOWER CART USING BLUETOOTH AND INFRARED
TECHNOLOGY WITH COLLISION DETECTION
Abstract
Transporting objects using vehicles are used by either pushing or dragging. This
way of transporting things requires a lot of effort and could be dangerous to the
people using them. According to the Health & Safety Executive (HSE) more than
one third of all accident injuries warranting more than 3 days absence from work,
involve lifting or manual handling. With an intelligent transport vehicle, moving an
object from one place to another can be done with less effort. This vehicle is capable
of following a user while in motion with constant speed. It automatically moves itself
in different direction guided sensors detecting a device carried or attached to a
person it “follows”. Moreover, it can detect nearby objects that could cause
collisions to the device and it could also measure the weight of the object/s it carries.
This device is applicable to offices, libraries, hospitals, airports and in certain
building construction operations.
Keywords: automatic, transport, objects
Introduction
Manual handling is any transporting or supporting of a load by one or more
workers. It includes the following activities: lifting, holding, putting down, pushing,
pulling, carrying or moving of a load parts (European Agency for Safety and Health at
Work, 2007). Manual transportation of objects is indeed a part of humans’ lives.
Although technology exponentially arises and innovates overtime, manual handling is
still practiced in workplaces: in factories and building sites, shopping malls, offices,
airports and hospitals, from the very beginning until now. As a result, manual handling
injuries are consistently and apparently evident.
18
Based on the research, more than a third of all over-three-day injuries reported
each year are caused by manual handling - the transporting or supporting of loads by
hand or by bodily force (Op De Beeck, R. and Hermans, V., 2000). Furthermore, about a
fourth of European workers suffer from back pain, which tops the list of all reported
work-related disorders. Loads that are too heavy, too large, difficult to grasp, unbalanced
and unstable and difficult to reach and tasks that are too strenuous and involves awkward
postures and movements are the factors that make manual handling hazardous. As a
result, manual handling can lead to negative health effects such as fatigue, injuries of the
back, neck, shoulders, arms or other body parts (European Agency for Safety and Health
at Work, 2007).
Because of all these negative health effects due to manual handling, prevention
measures were offered such as: having a general risk assessment guidelines and conduct
risk assessment of manual handling activities, conduct trainings and learn good handling
techniques for lifting, pushing and pulling (Health and Safety Executive, 2008)
(European Agency for Safety and Health at Work, 2007). However, those accidents are
still inevitable and the present methods of prevention will not guarantee the full safety of
the people.
With this in mind, the researchers proposed a study that aims to develop a
system of a cart capable of following a single person. Specifically, the cart is able to
identify the user and when successfully identified, the cart is able to follow that person in
any direction this user will go. And to complete its functionality, the cart is able to detect
collisions along the way and can determine the weight of the things that it can carry.
19
This proposed device is used to eliminate the risks of manual handling.
Especially for the accidents that may occur when manually transporting objects from one
location to another (Health and Safety Executive, 2008). Aside from this, it is also
essential to emphasize the huge advantage of automatic transporting of objects over the
manual. The techniques used for collision detection and user identification in the cart
were derived from the previous researches’ techniques but of different approach.
The study covers the development of an automated cart capable of following a
single person. The system will include sensors that have certain maximum and minimum
detection range. Also, the cart could only travel along flat surfaces and inclined or uneven
platforms are outside the scope of this study. The vehicle could carry a maximum weight
load of 40 kilograms. The maximum distance from the cart to the person would be 1
meter.
Methodology
The study started with the observation and acquiring information from the past
researches necessary for completing our objective. After reviewing the acquired related
literatures, the researchers planned and designed a leader-follower cart equipped with
RFID Technology, Bluetooth Device, Infrared Sensors, Infrared LED, Proximity Sensors,
and Load Cell. The system components are attached on either a belt, which the user will
be wearing, or on the cart itself. The RFID technology is used for the user’s
identification. A passive RFID tag is carried by the user to be used for activating the cart
by tapping it on the RFID reader on the cart. Several technologies are used such as IR
20
LED, IR sensors and proximity sensors to detect the user’s position, which direction the
user is facing and can calculate the distance between them. The IR LED is located in
front of the cart and five IR sensors are attached to the belt. Bluetooth technology is
responsible for continuous transmission of data between the cart and the belt. The cart’s
wheels are to be operated by DC geared motors, and to complete the cart’s functionality,
it is integrated with proximity sensors to prevent the cart from collisions while it is
moving specially when turning. For additional features, the cart is to be equipped with
load cell sensor to measure the weight of objects to be carried by the cart, and to ensure
that the weight will not exceed to 40kg. The batteries to be used are two series connected
12volts Rechargeable Lead Acid Batteries.
21
Block Diagram
INPUT
PROCESS
OUTPUT
CART
Proximity Sensor
Microswitch
LCD
Infrared LED
Microcontroller
Load Cell
Buzzer
Wiper Motor
RFID Reader
Master Bluetooth
BELT
Slave Bluetooth
Left Inrared Sensor
Middle Infrared
Sensor
Microcontroller
Right Infrared Sensor
Figure 3.1: System Overall Block Diagram
22
LCD
The block diagram presented in Figure 3.1 shows the different input, process,
and output components for both the belt and the cart. All input devices are being process
by the microcontroller unit and each input corresponds to an output. The IR Sensor
located in the belt guides the cart on which direction to go. Bluetooth Technology is used
for continuous communication between the cart and the belt. Aside from the DC-powered
wheels, the output components of the system are used mainly to indicate and show any
important information about the current process with text-based display and buzzer.
Both the belt and cart have microcontroller that process data from other
components. The overall hardware components are shown in Figure 3.2, Figure 3.3 and
Figure 3.4.
23
Figure 3.2: Cart’s Components – Top View
Figure 3.3: Cart’s Components – Side View
24
Figure 3.4: User Wearable Belt
Figure 3.5: The Basic Scheme
Figure 3.4 shows the basic scheme of the system; a cart following a human who
wears a belt. The IR LED located in front of the cart can be adjusted with respect to the
user’s height. For proper functionality, IR LED and IR Sensor must be on the same level
25
with respect to the ground. Proximity sensor prevents collision between the cart and the
user.
System Flowchart
Start
Connected to
power?
No
Stop
Yes
Initialize All
Variables
Tap RFID
No
RFID
Recognize?
Yes
Enable Cart
A
Figure 3.6: Flowchart of Enabling the Cart
26
A
Output the
weight of
load
C
Check
reading of
proximity
sensor
Is there
collision?
Yes
Stop Cart
No
B
Figure 3.7: Flowchart of the Weight Reading and Collision Detection
27
Figure 3.8: Flowchart of the Cart’s Movement
When powered on, the cart will initiate the RFID reader and waits for RFID tag
to be tapped to start its leader-following function as shown in Figure 3.6. Once the RFID
28
tag is confirmed by the cart, the system will initialize the infrared signal transmission and
the Bluetooth master and will search for the Bluetooth slave on the belt to send and
receive.
Collision detection of the system is done using four components. Two proximity
sensors are attached on both sides of the cart to determine any collision that may happen
when the cart is turning to a different direction, one proximity sensor is attached at the
front side to retain distance from the user and the cart, and a microswitch is attached on
the front bumper that is switched on when an object collided on the lower front side of
the cart. A separated toggle switch is used for activating the weight reading using load
cell. This is necessary since the load cell uses a separate battery when it is used. When
the load cell reading is turned on, the cart will show a real time value of the current
weight loaded on the cart. Since only one microcontroller is used on the cart, the weight
reading and collision detection operations are done sequentially as seen on Figure 3.7.
The cart’s DC powered wheels will not rotate until an infrared reader among the
five attached on the belt receives the continuously transmitted signal from the infrared
transmitter. The pair of wheels rotates in different pattern depending on which IR reader
received the signal.
29
Materials Used
-
RFID Tag and Reader for the user’s identification and powering up of the cart
-
Bluetooth Master and Slave for continuous transmission of signal from the cart to
the belt, and vice versa.
-
Infrared Sensors that are used as a switch of the Bluetooth’s continuous signal
transmission
-
Proximity Sensors used to avoid collisions when the cart is turning left or right or
even moving towards the user that moves forward
-
Load Cell sensor mainly used for weight reading and LCD for display
-
12V Lead Acid Battery served as the power supply of the cart
30
Results and Discussion
Problem 1: Can the cart be able to identify the user whom it will follow?
Null Hypothesis H0: The cart will not be able to identify the user it will follow
The researchers tested the reliability of the RFID embedded on the cart. To
obtain valid results, the researchers used different RFID tags and the cart’s role is that it
should still be able to identify the user’s RFID tag and reject those tags that are not of the
user’s. Hence, we have used 12 different RFID tags and one of them, the Tag # 7, was the
user’s. The researchers used the binary numbers to indicate YES (1) if the tag matched
the reader on the cart, and the NO (0) if otherwise.
Trial
Tag
1
Tag
2
Tag
3
Tag
4
Tag
5
Tag
6
Tag
7
Tag
8
Tag
9
Tag
10
Tag
11
Tag
12
1
0
0
0
0
0
0
1
0
0
0
0
0
2
0
0
0
0
0
0
1
0
0
0
0
0
3
0
0
0
0
0
0
1
0
0
0
0
0
4
0
0
0
0
0
0
1
0
0
0
0
0
5
0
0
0
0
0
0
1
0
0
0
0
0
6
0
0
0
0
0
0
1
0
0
0
0
0
7
0
0
0
0
0
0
1
0
0
0
0
0
8
0
0
0
0
0
0
1
0
0
0
0
0
9
0
0
0
0
0
0
1
0
0
0
0
0
10
0
0
0
0
0
0
1
0
0
0
0
0
Total
0
0
0
0
0
0
10
0
0
0
0
0
Table 3.1: Test for Reliability of RFID
31
Using distribution in the Univariate Analysis which analyzes the summary of the
frequency of individual values or ranges of values for a variable, the researchers have
identified that the RFID tag of the user was the Tag # 7, and the researchers used the
formula for the percentage error (see Appendix for computation) which led the test of
RFID to have a hundred percent reliability of the data gathered. Hence, the cart was able
to identify the user that it will follow: the null hypothesis was rejected.
Figure 3.9: Sample Screenshot for Unknown Tag
Figure 3.10: Sample Screenshot for User Tag
32
Problem 2: Is there any significant difference on the cart’s movement and the user’s
movement?
Null Hypothesis H0: There is no significant difference on the cart’s movement and the
user’s movement.
The second hypothesis stated that there is a significant difference on the cart’s
movement and the user’s movement, which means that there is a significant error on the
following of the cart. To prove the validity of the hypothesis, the researchers set-up series
of tests wherein after the cart had already identified its corresponding user, and then the
user moved to his/her directions, combinations of whether going left, going right, going
forward or stop, the cart was expected to move to that user’s direction. In Table 3.2, there
were 4 columns that indicate the 4 movement directions, and each column was divided
into two: the actual column which was the movement of the cart, and the expected
column which was the movement of the user.
33
Forward
Left
Stop
Right
Trial
Actua
l
Expecte
d
Actua
l
Expecte
d
Actua
l
Expecte
d
Actua
l
Expecte
d
1
1
1
1
1
1
1
1
1
2
3
5
3
3
1
1
1
1
3
3
4
2
2
1
1
1
1
4
2
2
3
3
1
1
2
2
5
1
1
2
2
1
1
1
1
6
1
1
1
1
1
1
2
2
7
3
3
2
2
1
1
1
1
8
4
4
1
1
1
1
0
2
9
2
2
2
2
1
1
2
2
10
3
3
1
2
1
1
2
2
Tota
l
23
26
18
19
10
10
13
15
Table 3.2: Cart’s Movement following the User’s Movement
The data in Table 3.2 shows the combination of movements of the user and the
cart’s behavior relating to the user’s movement. The researchers devised this table by
counting the number of movements the user did per trial. For example, in trial 1, the user
decided to move one forward, one right, one left and one stop, in whatever sequence the
user decided to go. The sequence the user decided to make might be: Forward  Left 
Stop  Right, etc. However, the researchers did not specify the movement patterns or the
sequence of movements, for they are only interested in the magnitude response of the
user’s and the cart’s movement and the same went on to the different trials. The expected
column shows the tally of the user’s movement while the actual column shows the tally
of the cart’s.
34
Forward
Left
Stop
Right
Total
Success
23
18
10
13
64
Failure
3
1
0
2
6
Total
26
19
10
15
Table 3.3: Summary of Table 3.2
70
Table 3.3 shows the summary of the movement of the cart following the user. It
shows the number of successes and failures of the cart, which simply pertains to its
performance of following the user which then shows the differences between the cart’s
and the user’s movement. To prove the hypothesis whether it is acceptable or not, the
researchers used the test called the Chi-Square (X2) as shown in Appendix A.
Figure 3.11: Sample Screenshot of the Cart following the User
So the researchers obtained the Chi-Square (X2) Value of 1.9209 and the degree
of freedom (df3) based on the number of rows and columns at 0.01 level of significance is
35
11.345. To be significant at .01 level with 3 degrees of freedom (df3), the X2should be
equal to or greater than 11.345. Hence, there is no significant difference in the movement
of the cart and the movement of the user, which indicates consistency of the cart’s
performance in any direction it goes as led by the user, and that consistency is proven
correct by just analyzing the ratio of the number of successes against the number of
failures, the lowest percentage per movement is 86.67% (right) and the highest is 100%
(stop), and accumulates the overall performance of 64/70 or 91.43%. Thus the null
hypothesis is accepted.
Problem 3: Can the cart be able to detect collisions during the following process?
Null Hypothesis H0: The cart will not be able to detect collisions during the following
process.
The third hypothesis states that the cart won’t be able to detect collisions during
the following process tested the capability of the cart to observe and detect collisions
along the way and when detected, the cart will simply turn to halt, and when the signal
between the user and the cart is disconnected (e.g. the user continued to walk while the
cart stopped due to detected collision) the user’s belt will create a noise indicating that
the cart’s not anymore following the user. So to make the cart be able to follow the user
again, it’s the user’s duty to take away the collision from the cart.
36
To prove that the hypothesis is true and can be accepted, the researchers tested
the functionality of the three sensors on the cart that detects the collision. The first two
sensors tested were located on the two sides of the cart, positioned higher than the back
wheels of the cart and were approximately one foot. from the ground. These sensors were
used to detect left and right collisions especially when turning left or right. Examples are
the street corners that could inflict damages to the cart and to the items inside the cart.
The other sensor was called the bumper sensor located on the lower front part of the cart.
This was used to detect collisions that were unseen by the user when moving forward.
Examples of these obstacles were small objects with the height less than one foot, for
they could not be neglected for they too can cause serious damages and delays to the cart
and to the items inside the cart. Table 3.4 shows the results of the testing conducted. For
the researchers to have a numerical data for this testing, they again used the binary
numbers to indicate YES (1) if the cart was able to detect collisions, and the NO (0) if
otherwise.
37
Trial
Left Side
Right Side
Bumper
1
1
1
1
2
1
1
1
3
1
1
1
4
1
1
1
5
1
1
1
6
1
1
1
7
1
1
1
8
1
1
1
9
1
1
1
10
1
1
1
Total
10/10 = 100%
10/10 = 100%
10/10 = 100%
Table 3.4: Tally of the Test for Collision Detection
Based on the results shown in Table 3.4, the cart is 100% able to detect the
collisions using the proximity sensors on the left and on the right, and also the bumper
sensor on the front. Hence the null hypothesis was proven false.
38
Figure 3.12: Testing Collision Detection at the Right
Figure 3.13: Testing Collision Detection at the Left
39
Figure 3.14: Testing Collision Detection at the Bumper
Hypothesis Problem 4: Is there any significant difference between the weight reading
using the weighing scale from the weight reading using the cart?
Null Hypothesis H0: There is no significant difference between the weight reading using
the weighing scale from the weight reading using the cart.
The hypothesis above states that no significant difference exists between the
theoretical weight reading or reading using the weighing scale and the obtained weight
from the cart. Again to prove if the hypothesis is valid, the researchers conducted a test
called the t-test. First, the researchers gathered the data of theoretical and actual weight
reading of 10 different sample items listed in Table 3.5, and from that table we can now
apply the t-test, see Appendix A.
40
Item
Theoretical
Weight using
Weighing scale
(kg)
Obtained
Weight from
the Cart (kg)
Weight
Difference
%
Difference
Pair of Shoes
1.56
1.57
0.01
0.64
Bundled Clothes
2.25
2.28
0.03
1.33
Canned Goods
2.82
2.85
0.03
1.06
Rice Grain
3.06
3.10
0.04
1.31
Drinking Water
3.32
3.36
0.04
1.20
Gas Tank
4.29
4.33
0.04
0.93
Medicine Ball
5.01
5.08
0.07
1.39
Plant Pot
5.56
5.64
0.08
1.44
Television
6.64
6.72
0.08
1.20
Books
7.54
7.66
0.08
1.06
Table 3.5: Data of Theoretical and Actual Weight Reading
41
Table 3.5 shows the weight readings of two different variables. It was also
observed that the testing was limited to only 10 trials, and the weight testing was limited
to maximum weight of 7.66 kgs, for the researchers assumed that the succeeding weights
will also yield the same results. So the researchers obtained the T-Test Magnitude Value
of 0.058143 and the degree of freedom (df18) that was based on the number of cases of
Figure 3.15: Sample Screenshot for the Obtained Weight using the Cart
the first and the second variable minus two at 0.01 level of significance. Thus, df18is
2.878, and to be significant at .01 with 18 degrees of freedom (df3), the t should be equal
to or greater than 2.878. However the value of t obtained is less than 2.878, so there is no
significant difference between the weight reading using the weighing scale and the
weight reading using the cart, which also indicates the accuracy of the weight reading
using the cart. Thus the null hypothesis is accepted.
42
Figure 3.16: Sample Screenshot for the Theoretical Weight using
Weight Balance
43
REFERENCES
Bluman, A. (2009). Elementary Statistics: A Step by Step Approach, 7th Edition,
McGraw-Hill International Edition, New York, USA.
Borenstein J., et al, (1997). Mobile Robot Positioning - Sensors and Techniques. Journal
of Robotic Systems, Special Issue on Mobile Robots. Volume 14 (4), 231-249.
Braun, T., et al, (2005). Detecting and Following Humans with a Mobile Robot.
University Kaiserslautern, Robotic Systems Group, Kaiserslautern, 67653, Germany
Calmorin, L., et al, (2007). Research Methods and Thesis Writing, 2nd Edition, Rex
Bookstore, Inc., Manila, Philippines.
European Agency for Safety and Health at Work, (2007). Hazards and risks associated
with manual handling in the workplace
Fr´ed´eric R., (2008). Ultrasonic Relative Positioning for Multi-Robot Systems
Garfinkel, S. and Henry H., (2005). Understanding RFID Technology. Garfinkel Book,
15-35
Goel P., Roumeliotis, S. I. and G. S. Sukhatme. (2000). Robot Localization Using
Relative and Absolute Position Estimates. Department of Compuer Science, Institute for
Robotics and Intelligent Systems, University of Southern California, Los Angeles, CA
Health and Safety Executive. (2008). Getting to grips with manual handling,
Kulyukin, V., et al., (2004). RFID in Robot-Assisted Indoor Navigation for the Visually
Impaired.
Lee, J., et al., (2007). A Comparative Study of Wireless Protocols: Bluetooth, UWB,
ZigBee, and Wi-Fi. Information & Communications Research Labs Industrial Technology
Research Institute (ITRI), Hsinchu, Taiwan
Muhammad, Tarek. (2009). Using Ultrasonic and Infrared Sensors for Distance
Measurement. World Academy of Science, Engineering and Technology
Op De Beeck, R. and Hermans, V. (2000). European Agency for Safety and Health at
Work, Research on work-related low back disorders. Luxembourg, Office for Official
Publications of the European Communities.
Pugh J., et al, (2008). A Fast On-Board Relative Positioning Module for Multi-Robot
Systems.
Roy Want, (2006). An Introduction to RFID technology
Revere Transducers, (2001). Load Cell Technology in Practice.
Tan, G., et al., (2001). Forming Scatternets from Bluetooth Personal Area Networks.
MIT Laboratory for Computer Science
44
Chapter 4
CONCLUSION
A leader-follower cart using Bluetooth and infrared technology with collision
detection was designed and developed. The researchers tested the functionality of the cart
including its capability to identify the user, follow the leader to any direction and detect
collision and the accuracy of its weight reading. After testing and statistical analysis, the
researchers conclude that the designed cart was able to satisfy the objectives of this study.
Through Univariate analysis, the first test yields a weighted arithmetic mean of
100% proving that the cart is able to identify the user whom it will follow. The card is
only able to identify the right user. Using Chi-Square analysis, the second test yields an
overall performance of 64/70 or 91.43%. and showed that the cart can be able to
consistently follow the identified user. After conducting the third test and yielding a
100% of arithmetic mean, the cart is able to detect collisions during the following
process. Possible collisions from the front and both sides of the cart are successfully
detected. Lastly, the final test using t-test analysis and T-Test magnitude value of
0.058143 proved that the cart is able to measure the weight of its load and the weight
measurement from the load cell is properly displayed in an LCD panel.
45
Chapter 5
Recommendation
Although the basic features of a leader-following cart with collision detection are
achieved in this study, improvements can be done to enhance the capability of the present
system. The usage of additional sensors can make the system detect more colliding object
along the cart’s way during travel. user identification and activation can be done
wirelessly using different technology such as Wi-Fi or ZigBee. The cart is limited to a
maximum weight load of forty kilograms and so designing a similar system with more
load capacity is appropriate to more industry application. Also, one of the limitations the
researched have identified is placing a buzzer to notify the user that the cart carries
weight exceeding to 40kgs. The belt can be improved by using smaller components for
the electronic device connected to it; thus eliminating the need for the user to carry the
device by hand. Lastly, the future researchers can make the speed of the cart vary
according to the speed of the user.
46
APPENDIX A
Statistical Computations
47
Hypothesis Problem 1: Can the cart be able to identify the user whom it will follow?
Null Hypothesis H0: The cart will not be able to identify the user it will follow.
Trial
Tag
1
Tag
2
Tag
3
Tag
4
Tag
5
Tag
6
Tag
7
Tag
8
Tag
9
Tag
10
Tag
11
Tag
12
1
0
0
0
0
0
0
1
0
0
0
0
0
2
0
0
0
0
0
0
1
0
0
0
0
0
3
0
0
0
0
0
0
1
0
0
0
0
0
4
0
0
0
0
0
0
1
0
0
0
0
0
5
0
0
0
0
0
0
1
0
0
0
0
0
6
0
0
0
0
0
0
1
0
0
0
0
0
7
0
0
0
0
0
0
1
0
0
0
0
0
8
0
0
0
0
0
0
1
0
0
0
0
0
9
0
0
0
0
0
0
1
0
0
0
0
0
10
0
0
0
0
0
0
1
0
0
0
0
0
Total
0
0
0
0
0
0
10
0
0
0
0
0
Correct =
Error =
Error=
%Error =
Hypothesis Problem 2: Is there any significant difference on the cart’s movement and the
user’s movement?
48
Null Hypothesis H0: There is a significant difference on the cart’s movement and the
user’s movement.
Forward
Left
Stop
Right
Total
Success
23
18
10
13
64
Failure
3
1
0
2
6
Total
26
19
10
15
70
Using the Chi-square Test,
∑
Where:
is the Chi-square
O is the Observed Frequency
E is the Expected Frequency
Expected Frequency Computation
49
O
E
O-E
(O-E)2
23
23.7714
-0.7714
0.5951
0.0250
18
17.3714
0.6286
0.3951
0.0227
10
9.1429
0.8571
0.7346
0.0803
13
13.7143
-0.7143
0.5102
0.0372
3
2.2286
0.7714
0.5951
0.2670
1
1.6826
-0.6826
0.3951
0.2348
0
0.8571
-0.8571
0.7346
0.8571
2
1.2857
0.7143
0.5102
0.3968
Total 70
70.0540
0.0000
To find the degree of freedom (df) at
level of significance ( ),
(refer to figure 1)
50
1.9209
51
Hypothesis Problem 4: Is there any significant difference between the weight reading
using the weighing scale from the weight reading using the cart?
Null Hypothesis H0: There is no significant difference between the weight reading using
the weighing scale from the weight reading using the cart.
Item
Pair of Shoes
Bundled Clothes
Canned Goods
Rice Grain
Drinking Water
Gas Tank
Medicine Ball
Plant Pot
Television
Books
Theoretical
Weight using
Weighing scale
(kg)
1.56
2.25
2.82
3.06
3.32
4.29
5.01
5.56
6.64
7.54
Obtained Weight
from the Cart
(kg)
Weight
Difference
% Difference
1.57
2.28
2.85
3.10
3.36
4.33
5.08
5.64
6.72
7.66
0.01
0.03
0.03
0.04
0.04
0.04
0.07
0.08
0.08
0.08
0.64
1.33
1.06
1.31
1.20
0.93
1.39
1.44
1.20
1.06
Using the t-test,
̅̅̅
√
Where:
t
=
t-test
̅̅̅
=
Mean of the First Variable
̅̅̅
=
Mean of the Second Variable
=
Variance of the First Variable
=
Variance of the Second Variable
52
̅̅̅
=
Number of cases of the First Variable
=
Number of cases of the Second Variable
∑
∑
Computation of the Arithmetic Mean ( ̅ )
̅̅̅
̅̅̅
∑
∑
Theoretical Weight (Weighing
Scale)
̅̅̅
Obtained Weight (Cart)
̅̅̅
̅̅̅
̅̅̅
1.56
-2.64
6.996025
1.57
-2.689
7.230721
2.25
-1.955
3.822025
2.28
-1.979
3.916441
2.82
-1.385
1.918225
2.85
-1.409
1.985281
3.06
-1.145
1.311025
3.10
-1.159
1.343281
3.32
-0.885
0.783225
3.36
-0.899
0.808201
4.29
0.085
0.007225
4.33
0.071
0.005041
5.01
0.805
0.648025
5.08
0.821
0.674041
5.56
1.355
1.836025
5.64
1.381
1.907161
6.64
2.435
5.929225
6.72
2.461
6.056521
7.54
3.335
11.122225
7.66
3.401
11.566801
53
Total
42.05
0
Computation of Variance (
∑
̅̅̅
∑
̅̅̅
34.37325
42.59
0
35.49349
)
Computation of T-test
̅̅̅
̅̅̅
√
√
√
**
Computation of the degree of freedom (df) at
(refer to figure 2)
54
level of significance ( ),
**Sign does not affect the obtained magnitude of t because the test is only focused in the
difference between the means
55
APPENDIX B
Schematic Diagrams
56
Schematic Diagram of Cart
57
Schematic Diagram of Belt
58