SubramaniamSenthilKumar_FYP

SIM UNIVERSITYSCHOOL OF SCIENCE AND TECHNOLOGY

WIRELESS NETWORK FOR MOBILE ROBOTS

STUDENT : SENTHIL KUMAR S/O SUBRAMANIAM (Q0704622)SUPERVISOR : DR CARLOS ACOSTAPROJECT CODE : JUL2010/ENG/067

A project report submitted to SIM UniversityIn partial fulfillment of the requirements for the degree of

Bachelor of Engineering (or Bachelor of Electronics)

May 2011

Student Name: Senthil Kumar s/o Subramaniam (Q0704622)Project: JUL201/ENG/067

ABSTRACT

Wireless technology is increasingly used worldwide. From radio waves for listeners to listen to

their favourite radio stations, to things like Bluetooth earpieces for cell phones communications

while driving. And with potential to do more, people are finding more usage for this technology.

One such usage could be using this technology to help with disaster aiding exercises.

This project would be focusing on how to use a type of wireless technology, known as Zigbee, to

help in such situations. In scenarios where it is required to search for other potential survivors in

a disaster hit area, like in a half collapsed building, it would be risky to send in rescue workers

for the fear of the building collapsing again. Instead a team of robots could help in such a

situation without putting more people at risk.

These robots would work as ‘eyes’ to the rescue workers to see if there are signs of life in the

debris. ZigBee chips would be mounted on the robots to enable the robots to communicate with

one another and to the rescue workers as well. If there is indeed a survivor in the rubble, the

robot could either send a signal back to the workers or enhancements can be made to the robot

for it to help by itself.

If there is no survivor or anyone trapped, at least the rescue workers would not have to take the

extra risks. And hence these robots would prove to be useful in such situations.

ZigBee was chosen because its technology is simple to use and less expensive than other

WPANs (wireless personal area network) such as Bluetooth. It also has a long battery life.

The experiments done show the capabilities of ZigBee and the chosen Digimesh network.

Possible future developments would also be mentioned.

2ENG 499 Capstone Final Report


ACKNOWLEDGEMENT

The author would like to express his most sincere gratitude to his project supervisor, Dr Carlos

Acosta for his patience and guidance throughout the project. He has been approachable and

helpful at all times when he needed his input and suggestions. He had made time and

arrangements to meet up with the author despite both their busy schedules and helped to ensure

the project went smoothly. He would also like to thank Dr Carlos’ students for their help in

conducting the experiments, because without them, it would be impossible to carry out the

experiments.

He would also like to sincerely thank his employer; ESPNSTAR Sports and his manager Miss

Tan Wee Sze and supervisors for allowing him to make flexible arrangements to focus on his

project during the critical period.

Lastly he would like to thank his family and friends for their support and understanding

throughout the project especially when he was feeling lost and demoralized, due to personal

reasons. And being understanding for not being able to meet and keep in touch with them

especially during the last few months due to time constraints. He would not have been able to

succeed without their help and understanding. A big Thank you to everyone involved.



TABLE OF CONTENTS Page

ABSTRACT 2

ACKNOWLEDGEMENT 3

TABLE OF CONTENTS 4

LISTS OF FIGURES 6

LIST OF TABLES 7

CHAPTER ONE: INTRODUCTION 8

1.1 BACKGROUND AND MOTIVATION 8

1.2 OBJECTIVES 9

1.2.1 PROJECT OBJECTIVES 9

1.2.2 OVERALL OBJECTIVES 9

1.3 PROJECT SCOPE 10

1.4 EQUIPMENT LIST 11

1.4.1 ABOUT ARDUINO 12

1.5 LAYOUT OF THE PROJECT REPORT 13

CHAPTER TWO: REVIEW OF THEORY AND PREVIOUS WORK 14

2.1 WIRELESS TECHNOLOGY 14

2.2 WIRELSS SENSOR NETWORK 15

2.3 TOPOLOGIES 16

2.3.1 BUS TOPOLOGY 16

2.3.2 RING TOPOLOGY 16

2.3.3 TREE TOPOLOGY 17

2.3.4 STAR TOPOLOGY 17

2.3.5 MESH TOPOLOGY 17

2.3.6 HYBRID TOPOLOGY 18



Page

CHAPTER THREE: ZIGBEE 19

3.1 INTRODUCTION TO ZIGBEE 19

4.1.1 IEEE 802.15.4 19

4.1.2 ZIGBEE VS OSI MODEL 20

3.2 ZIGBEE SERVICES 20

3.3 HOW DOES ZIGBEE WORK? 20

3.4 WHY ZIGBEE? 22

3.5 ZIGBEE SERIES 1 VS SERIES 2 23

3.6 DIGIMESH 24

3.7 DIGIMESH VS ZIGBEE SERIES 2 25

CHAPTER FOUR:

EXPERIMENTS AND METHOD OF DEPLOYING ROBOTS 28

4.1 EXPERIMENTS WITHOUT ROBOTS 28

4.1.1 ZIGBEE SERIES 1 EXPERIMENTS 28

4.1.2 ZIGBEE SERIES 2 EXPERIMENTS 29

4.1.3 DIGIMESH EXPERIMENTS 29

4.2 EXPERIMENTS WITH ROBOTS 31

CHAPTER FIVE: PROJECT MANAGEMENT, PROBLEMS ENCOUNTERED

AND REFLECTIONS 37

5.1 PROJECT PLANNING 37

5.2 PROJECT MANAGEMENT 39

5.3 TECHNICAL PROBLEMS 41

5.4 PERSONAL PROBLEMS 41

5.4 REFLECTIONS 42

CHAPTER SIX: CONCLUSION AND FUTURE DEVELOPMENTS 42

6.1 CONCLUSION 42

6.2 FUTURE DEVELOPMENTS 43

REFERENCES 44

APPENDIX A: THE INITIAL GANTT CHART 45



LIST OF FIGURES

Page

Figure 1.1: The radiation detecting robot in Japan 9

Figure 1.2: The four-wheeled Omni directional Robot 9

Figure 1.3: ZigBee chips 11

Figure 1.4: Arduino Duemilanove board and Arduino Xbee shield board 11

Figure 2.1: A typical wireless sensor network 15

Figure 2.2: The various types of topology 16

Figure 2.3: Partially connected mesh network vs a fully connected mesh network 17

Figure 3.1: OSI model vs ZigBee 20

Figure 3.2: The node types 21

Figure 3.3: The ZigBee network 21

Figure 3.4: The Digimesh network vs ZigBee series 2 network 24

Figure 3.5: Interchangeable nodes 25

Figure 4.1: Series 1 peer to peer connection 28

Figure 4.2: Series 1 failed broadcast 28

Figure 4.3: Series 2 network connection 29

Figure 4.4: Series 2 network failure 29

Figure 4.5: Layout of the Robotics Laboratory in SP 30

Figure 4.6: Final layout of the nodes, experiment without robots, using digimesh 31

Figure 4.7: Robot 1 33



Page

Figure 4.8: Robot 2 33

Figure 4.9: Robot 3, co-space robot 34

Figure 4.10: Final layout of experiment with robots, using digimesh 34

Figure 4.11: The ZigBee connectivity flowchart 35

Figure 4.12: The ZigBee signal flowchart 36

Figure 5.1: Flowchart of proposed approach and method 37

Figure 5.2: The final Gantt chart 40

LIST OF TABLES

Table 3.1: Zigbee series 2 vs Digimesh 27

Table 5.1: The summary of project tasks list 39



1. INTRODUCTION

1.1 Background and motivation

There are certain times when human lives are put at risk. Like searching for survivors in a

building which has collapsed, or trying to rescue survivors from other forms of disasters like

floods. And as there is an increasing need for search and rescue operations due to the number of

either man- made or natural disasters occurring, more human lives are put at risk, to save their

fellow humans.

In the present day robots are already being used mainly recreational activities (being used in

movies for example) and for some research as well. Robots are also already being used in certain

scenarios whereby humans’ lives are at risk, for example, bomb disposal. Another scenario is the

current nuclear disaster at Japan, whereby the Japanese Engineers deployed robots to check the

radiation levels before any human can go in. These robots are equipped with radiation detectors,

infra red and cameras (Fig 1). It is very likely that robots would have more functions in the near

future and perhaps one of the functions could be helping in search and rescue missions. And

thereby increasing the possibilities of saving more lives and at the same time reduce the risk

involved for the search and rescue teams.

These robots would also have to work in ‘teams’ and communicate with one another and

therefore speeding up the search process, much like how their human counterpart do. The robots

would also have to maintain communication with the base station, await instructions and return

feedback. And be able to forward messages to one another. In other words, these robots would be

able to work with little supervision and just respond/react when needed to. And of course

function as a group.

This project would help bringing this scenario one step closer to reality.



Fig 1.1(left): The radiation detecting robot in Japan, The New Paper 28/04/11

Fig 1.2(right): The four-wheeled Omni directional robot

1.2 Objectives

Initially, four-wheeled Omni directional robots (Fig 1.2) were supposed to be used in this

project. However as long as any robot is embedded with the Zigbee chip, the robot would be able

to work as well as documented in the experiments in the later chapters.

1.2.1 Project Objectives

The main goal of the project is enabling the moving robots to communicate with the other robots

in the group as well as maintaining communications with the base station using the Zigbee

technology. The robots would form a network and pass signals or messages through one another,

informing about possible survivors or dangers ahead or any other information. The main focus

would be designing the method of deploying the robot to maximize the coverage of an area. This



would be done by analyzing the behavior of the ZigBee signal by comparing and contrasting the

various topologies and choosing the best topology. The topology chosen, digimesh, would be

explained in the chapter 4. Another objective is to derive the algorithm for the robots, which is

also shown in the flowcharts at chapter 5.

The academic goal of this project is to develop the various skills of him like research, conducting

various tests and evaluation. It has also given him the opportunity to enhance his report writing,

project management, and presentation and communication skills, the last two points being

critical for the project presentation.

1.2.2 Overall Objectives

As mentioned in the introduction, robots and wireless technology are increasingly being used in

our societies. Instead of recreational usage now, using robots instead of humans for dangerous

assignments like search and rescue would reduce risking more people for such assignments. And

most importantly a project like this could lead to future search and rescue robot teams; it is

definitely a good investment especially considering it helps to save human lives.

1.3 Project Scope

This project requires knowledge in various fields/areas in order to develop the system and

complete it successfully.

The wireless technology used is Zigbee, which offers the following characteristics: a worldwide

public standard, highly secure, low power networks, different network topologies.

Familiarization with Arduino software and hardware was also required as the Arduino software

and boards are used to carry out the experiments. An Arduino Duemilanove board and Arduino

Xbee shield board would be used. Some Xbee chips would also be used, the Xbee chips

representing the Zigbee chips, while the 2 boards would be required to connect to the computers

to enable tests and experiments to be conducted on. All these would be explained in next

paragraph.



1.4 Equipment List

The following equipments were essential and used in order to carry out the project.

Hardware:

Zigbee series 1 chips, MaxStream Xbee chips were the 1st generation chips before MaxStream

was bought over by Digi [1].

Fig 1.3: Xbee series 1 chips

Arduino Duemilanove board and Arduino Xbee shield board:

Fig 1.4: Arduino Duemilanove board and Arduino Xbee shield board



Other hardware includes a PC installed with Windows XP software and PCs in SP using Linux

OS. And the robots used would be shown in the experiments section.

Software:

Arduino software was used to download the programmes into the Arduino boards. Arduino is

compatible with ZigBee and is user friendly as well. Windows and Linux OS, Microsoft Office

and Internet explorer were the other software used.

1.4.1 About Arduino

Arduino is an open-source physical computing platform based on a simple I/O board, and a

development environment for writing software. The Arduino programming language is an

implementation of Wiring / processing language. Arduino can be used to develop stand-alone

interactive objects or can be communicate with software running on your computer (e.g. Flash,

Processing, and MaxMSP.)

"Duemilanove" means 2009 in Italian and is named after the year of its release. The

Duemilanove is the latest in a series of USB Arduino boards with auto reset function (from

Arduino0009 IDE), extended power connector, built in LED, USB over current protection and

more.

Duemilanove can be powered via the USB connection or with an external power supply. The

power source is selected automatically. This newly upgraded version is RoHS compliant (lead-

free), fully assembled and tested in Italy. The on-board ATMega328 microcontroller with 32K

bytes of program memory is pre-loaded with boot loader which allows user to download program

without additional hardware. Note: The production and sales of Duemilanove had ceased at

the point of writing.

The Xbee shield allows an Arduino board to communicate wirelessly using Zigbee. The module

can communicate up to 100 feet indoors or 300 feet outdoors (with line-of-sight). It can be used

as a serial/usb replacement or you can put it into a command mode and configure it for a variety

of broadcast and mesh networking options.



1.5 Layout of the Project Report

This section would explain and summarize the layout of the report.

Chapter 1, Introduction: This chapter is the introduction to the project as well as the ‘why’ and

‘what’ of this project. The equipments used for the project are also mentioned.

Chapter 2, Review of previous work and theory: This chapter would go through the basics of

wireless technology, wireless sensor networks and topologies which would help to introduce

ZigBee.

Chapter 3, Zigbee: This chapter focuses on ZigBee, which is the wireless technology used for

this project. ZigBee consists on series 1, series 2 and the most importantly Digimesh network,

which was the main network used for this project.

Chapter 4, Experiments and method of deploying robots: This chapter would show the

experiments conducted which helped to determine the flowcharts and the algorithms to how the

robots would function. Experiments were conducted with and without the robots as well.

Chapter 5, Project management, Problems encountered and Reflections: This chapter would

summarize the how the project was managed and planned initially, the problems encountered

throughout the project and amendments made and the reflections.

Chapter 6, Conclusion and future development: In this last chapter, the conclusion and possible

future developments for this project is discussed.



2. REVIEW OF THEORY AND PREVIOUS WORK

This chapter would focus on the various aspects relevant to this project which are wireless

technology in general, the wireless sensor networks and the various topologies.

2.1 Wireless Technology

In the past and presently, communications between, for example a modem and a computer, were

done through wired technology, i.e. both the modem and the computer were connected

physically by a cable wire. While the wire does provide a direct and strong link to one another,

there are limitations to such technology. For example the two equipments can only be spread out

at the maximum of the wire length. Another limitation is that you would require one wire for

every connection, so for example if you were to connect a printer to the computer, you would

need another wire. Connections to a mouse and keyboard would require wires too and as there

would be many wires, it would look unsightly and people may trip on them as well. And for

places like radio and TV stations, to use wires to communicate to listeners and viewers would be

impractical. For example, if we are taking a 'live' feed like soccer for England, it would be near

impossible to use wires connected from England all the way to Singapore to transmit the

programme.

An alternative to this scenario is the wireless technology which is rapidly being used all over the

world. Wireless communication is the transfer of information over a distance without the use of

enhanced electrical conductors or wires [2]. For the scenario mentioned earlier, the feed from

England would be transmitted via Satellite to our screens at home. This is one form of wireless

technology. Other forms include converting the information and transmitting it via infra red

(remote controls being a good example and even wireless mouse and keyboards), radio waves,

blue tooth and ZigBee among others.

While wireless technology does have its drawbacks like interference and range, it has more

possibilities and it has more flexibility than its wired counterpart and hence the heavy usage.

This project would be focusing more on the form of wireless technology known as Zigbee and

the wireless sensor network.



2.2 Wireless sensor network

‘A wireless sensor network (WSN) consists of spatially distributed autonomous sensors to

cooperatively monitor physical or environmental conditions, such as temperature, sound,

vibration, pressure, motion or pollutants’ [3].

It is used mainly in the military for surveillance and increasingly used in other fields such as

medical/healthcare, industrial and even at home.

A sensor node consists of a radio transceiver or some other wireless communication device, a

small microcontroller and an energy device, which is usually a battery.

Figure 2.1: A typical wireless sensor network [3]

Unlike most computers which work with data created by human, wireless sensor networks work

with whatever details provided by the environment around it. It works in a network and has the

ability to read the information around it and then transmit the information to the user as seen in

the picture above. However, only the important news would be transmitted, such as when it

detects a possible survivor in a collapsed building site. For cases like earth quake early warning

devices, it detects abnormalities like vibrations and if it is deemed dangerous enough, it would

transmit a warning back to the base station.

It is also worth noting that each node has a relatively short range and hence they have to work

collectively as a group to convey the information [5]. The advantage of this is that each sensor

node would be able to have a better understanding the immediate environment around it and

hence the information would be more precise. For this project, the robots would be working in a

wireless sensor network and each robot would act like a 'sensor node'.



2.3 Topologies

Figure 2.2: various types of topology

Network topology is defined as the interconnection of the various elements (links, nodes, etc.)

of a computer network [6]. It can exist either physically or logically. Physical topology refers to

connections ‘physically’ like using wires, while logical refers to how the data is actually

transferred in a network as opposed to the physical design.

Topology can be considered as a virtual shape or structure of a network. This shape does not

correspond to the actual physical design of the devices on the computer network. The computers

on a home network can be arranged in a circle but it does not necessarily mean that it represents

a ring topology.

There are 6 types of topologies: Bus topology, Ring topology, Tree topology, Star topology,

Mesh topology and Hybrid topology. All 6 can either exist physically or logically.

I shall focus mainly on the Bus, Star and Mesh topologies and briefly touch on the rest. Star and

Mesh network are essential for the ZigBee network while the Bus network is harder to explain

and hence touched on more.

2.3.1 Bus topology: All the nodes are connected by a cable each and then connected to a main

bus cable as seen in figure 2.2. A terminator is used to prevent the signal from bouncing back

and forth on the bus cable. A signal from the source travels in both directions to all machines

connected on the bus cable until it finds the MAC address or IP address on the network that is

the intended recipient. If the machine address does not match the intended address for the data,

the machine ignores the data

2.3.2 Ring topology: As the name says, all the machines are connected in a closed loop or ring.

However if one of the machine in the ‘ring’ fails, the whole network fails to work too.



2.3.3 Tree Topology: Referring to figure 2.2, the machines are connected in a level like manner.

Machines on level one branches out to machines in level two and level two to level three and so

on. This means if there is a message to be transmitted to a level 3 machine, the message is sent to

a level 3 machine via the level 2 and level 1 machine connected to it.

2.3.4 Star Topology: A central hub is crucial to ensure the running of a star topology. All the

nodes are connected to this central hub and hence all the messages go through the central hub

before going to the individual nodes. The hub acts as a booster or repeater and hence ensures the

signal remains strong enough for the nodes to read. Star topologies are easy to design and

implement. A main advantage would be that it is easy to add new nodes/machines into the

existing network. For example, a 5 node star topology network can easily add a sixth member as

long as it is connected to the node. However the main disadvantage of Star topology is that once

the central hub fails, the whole network would fail as well as the information would not be able

to flow through. Zigbee network tend to use this topology often.

2.3.5 Mesh Topology: For this topology, all nodes/machines are directly connected to each other

whether physically or logically. The number of connections follow Reed’s Law which is, the

number of connections in a full mesh = n (n - 1) / 2. However Mesh technology is usually used in

a smaller network of nodes as it would prove to be costly and impractical for a fully physical

network connection. For this instance, a partially connected mesh technology would be

implemented as it would reduce cost and redundancies, which also makes the network less

complex.

Figure 2.3: Partially connected mesh network vs a fully connected mesh network



As seen in figure 2.3, a partially mesh network does look less complicated but yet they are all

connected. It is important to note that in partial connected network, the information travels in the

shortest path possible and if a node fails somewhere in the network there would usually be an

alternate route to use to transmit the message and hence this makes the mesh system practical.

2.3.6 Hybrid Topology: Lastly the hybrid topology usually refers to a combination of 2 or more

topologies and the end result is a topology which does not really resemble the original

topologies. A tree topology network which combines with another tree topology network would

not be a hybrid as it would still resemble a tree topology but a star topology network combining

with another star topology network would be considered a hybrid topology as it would resemble

something else.

Note: For Zigbee networks, mesh and star topologies are preferred as mesh topology ensures

connectivity at all times while star topology is easier to implement



3. ZIGBEE

This chapter would focus on the wireless technology used for this project, namely ZigBee.

3.1 Introduction to ZigBee

ZigBee is a wireless network technology that allows you to build medium to large wireless

networks of sensors and controllers. ZigBee offers the following characteristics: a worldwide

public standard, highly secure, low power networks, different network topologies. This projects

aims to develop a group of robots with communication via ZigBee. The introduction below just

illustrates how IEEE 802.15.4 and ZigBee work on a typical OSI model [7].

3.1.1 IEEE 802.15.4

This standard defines a communication layer at level 2 in the OSI (Open System

Interconnection) model. Its main purpose is to let the communication between two devices. It

was created by the Institute of Electrical and Electronics Engineers (IEEE), entity which main

task is to set standards so that technological developments can count with a common platform of

rules to be set over.

3.1.2 ZigBee vs OSI Model

This standard defines a communication layer at level 3 and uppers in the OSI model. Its main

purpose is to create a network topology (hierarchy) to let a number of devices communicate

among them and to set extra communication features like those mentioned earlier.



Figure 3.1: OSI model vs ZigBee [7]

3.2 ZigBee services

Zigbee offers mainly four different types of services which are:

1) Extra Encryption services for security purposes,

2) Association and authentication in which only valid nodes can join the network

3) Routing protocol which basically routes and forwards information to any node in the network

4) Application services where the nodes can be separated into clusters and each cluster can be

assigned a different task.

3.3 How does ZigBee work?

ZigBee organizes the network. So before anything, in order for a node to join the network, it has

to ask the coordinator for a network address (16bits) as part of the association process. Once the

node has successfully joined the network, it can send information to the others in the network

through the routers which are always 'awake' waiting for packets (information). When the router

gets the packet and the destination is in its radio of signal, the router checks to see if the

destination end device (node) is awake or has slept. If the end device is 'a sleep', the router would

bufferize the packet till the end device 'wakes up' and asks for news to the router.



The pictures below show how ZigBee works in network:

Figure 3.2: The node types

Figure 3.3: The ZigBee network

Figure 3.3 shows the 3 type of nodes in a ZigBee network,

The Coordinator is like the master device which controls the network.

The Router routes the information which is sent by the end devices

The End Device (the motes) is the sensor nodes which take the information from the

environment.



The motes (node) are usually battery powered while the routers and coordinator are not. Zigbee

creates star topology which goes by the following guidelines:

The end devices connect to a router or a coordinator

The routers can connect among end devices and the coordinator

The router and coordinators cannot sleep and have to save the packets of information in

heir buffer which goes to the end devices

The end devices can sleep

ZigBee also works in a 'Mesh Network' also known as peer to peer (P2P). What this means is

that all the devices are able to communicate with each other directly unlike the star network

where call communication goes through the central hub.

3.4 Why ZigBee?

ZigBee operates in the industrial, scientific and medical (ISM) radio bands; 868 MHz in Europe,

915 MHz in the USA and Australia, and 2.4 GHz in most jurisdictions worldwide. The

technology is intended to be simpler and less expensive than other WPANs (wireless personal

area network) such as Bluetooth [8].

ZigBee can activate (go from sleep to active mode) in 15 msec or less, the latency can be very

low and devices can be very responsive — particularly compared to Bluetooth wake-up delays,

which are typically around three seconds. ZigBees sleep most of the time, so the average power

consumption can be very low, resulting in long battery life.

So in summary the low cost of technology, the long battery life and the higher reliability and

longer distance of mesh networks make Zigbee a smart choice for people intending to use

wireless technology.



3.5 ZigBee series 1 vs Zigbee series 2

The XBee-PRO 802.15.4 (formerly Series 1) OEM RF module is an IEEE 802.15.4 compliant

solution that satisfies the unique needs of low-cost, low-power wireless sensor networks. The

module is easy-to-use, requires minimal power and provides reliable delivery of critical data

between devices. Innovations stamped in the XBee-PRO design enable it to yield two- to three-

times the range of standard ZigBee modules. [9]

The XBee-PRO ZNet 2.5 (formerly Series 2) OEM RF module satisfies the unique needs of low-

cost, low-power wireless mesh sensor networks. The Designed for ZigBee module is easy to use,

delivers exceptional range and provides reliable delivery of critical data between devices. Its

small form factor saves valuable board space. The XBee-PRO ZNet 2.5 module operates within

the ISM 2.4 GHz frequency band and is pin-for-pin compatible with XBee 802.15.4 (formerly

Series 1) modules. [10]

ZigBee series 1 makes a very good and reliable connection for only 2 ZigBees in a peer-to-peer

mode. This is mainly due to the fact that Zigbee series 1 does not have broadcasting abilities.

Meanwhile, ZigBee series 2 allows connection between many Zigbees in a typical network

mode, with the necessary assignments, as mentioned in the earlier sub-chapter.

Experiments which were conducted, as explained in the next chapter, would conclude why

Zigbee series 1 and series 2 were not suitable for this project. Instead the Digimesh network was

used.



3.6 Digimesh

Zigbee series one can only work in peer to peer mode, while series 2 can work in a network, but

the nodes require fixed assignments like a coordinator or router. However, it is not practical for a

group of mobile robots to have a fixed assignment as they would always be moving and their

positions could change about. So how do we resolve this problem? In order to solve this

problem, a Digimesh network is used instead. What is a Digimesh network? A digimesh network

looks similar to a series 2 network as seen in the figure below.

Fig 3.4: The Digimesh network vs ZigBee series 2 network

Comparing the two figures, one can see the differences between the 2 network layouts; a series 2

network requires the coordinator, router and end devices assignment, while the nodes in a

digimesh network require no assignments. What this means is all nodes are at the same level as

there is no parent-child relationship between the nodes and hence the nodes can interchange

about without any communication breakdown (Fig 3.5).



Fig 3.5: Interchangeable nodes

DigiMesh offers these advantages:

• Network setup is simpler

• More flexibility to expand the network

• Increased reliability in environments where routers may come and go due to interference or

damage

It is important to note that Digimesh is only meant to be used with ZigBee series 1 chips, to give

the series 1 chips more flexibility and the ability to connect to more ZigBee series 1 chips rather

than just peer to peer mode. It gives a different option to a series 2 ZigBee chip and allows users

to choose either series 1 with Digimesh or a series 2 depending on usage.

The Digimesh would have to be flashed /downloaded onto the ZigBee 1 chips in order for it to be

used. This is done with the help of the Arduino boards and software mentioned in chapter one.

3.7 Digimesh vs ZigBee (Series 2)

Allowing a node to sleep reduces power consumption. This is especially essential for nodes that

are battery powered like the robots being used in this project. Currently, Zigbee only allows for

End Devices to sleep but not Routers or Coordinators. If Coordinators or routers sleep,

connectivity will be lost.

DigiMesh allows all nodes to sleep, thereby increasing battery life. Sleeping is allowed by time

synchronization. A significant advantage of DigiMesh is it eliminates the single point of failure

associated with relying on a coordinator or gateway. Instead, DigiMesh establishes time

synchronization through a nomination and election process, enabling the network to operate

autonomously.


Node 1

Node 2

Node 3

Node 1

Node 2

Node 3


Since Zigbee is an open standard, it offers the potential for interoperability with devices made by

different vendors. This provides the ability to have over-the-air firmware updates. Furthermore,

Zigbee offers established profiles for common applications such as energy management and

lighting controls. A good selection of diagnostic support tools, like RF packet sniffers, is also

available.

DigiMesh, as a proprietary protocol, allows for tighter control of code space and therefore more

room for growth in features. DigiMesh is available on platforms with longer range and more RF

data rate options. Frame payload is generally larger, which can improve throughput for

applications that send larger data blocks. Additionally, DigiMesh uses a simplified addressing

method, which improves network setup and trouble shooting. [11]



Below is a table which summarises and compares ZigBee series 2 and the Digimesh.

Features ZigBee series 2 Digimesh

Node types, Benefits Coordinators, routers and end devices. End devices would cost lesser as it has less functions.

Only one type, homogenous. More flexibility to expand the network as it simplifies the network. Increased liability as there is no need for routers, especially when the routers lose connectivity due to some reason.

Sleeping routers and battery life Only end devices can sleep. Coordinator and router cannot and hence consume more battery.

All nodes can sleep. No single point of failure associated with relying on gateway or coordinator to maintain time synchronization

Range Less than 3.2 km for each hop Range up to 64km per hop with the help of range extenders.

Frame Payload Up to 80 bytes Up to 256 bytes, depending on product, suitable for larger blocks of data.

Code Size Larger, less room for growth in features

Smaller, so more room for growth in features

Frequency and data rates 2.4 GHz (250 kbps) 900 MHz (10, 125, 150 Kbps)

2.4 GHz (250 Kbps)

Interoperability Potential for interoperability between vendors

Proprietary

Security AES encryption. Can lock down network and prevent other nodes from joining

AES encryption.

Addressing Two layers, MAC address (64 bit) and Network address (16 bit)

MAC address (64 bit) only.

Table 3.1: ZigBee series 2 vs Digimesh [11]



4. EXPERIMENTS AND METHOD OF DEPLOYING ROBOTS

This chapter would explain the experiments conducted in order to determine how to deploy the

robots. It also shows why the Digimesh network was selected instead of a ZigBee series 1 or

series 2.

4.1 Experiments without Robots

4.1.1 ZigBee series 1 experiments

Experiments were conducted with Zigbee series 1 between 2 Zigbee nodes and

information/signal was able to be sent and received as seen in figure 4.1. The nodes here are

represented by ZigBee chips mounted onto the Arduino Boards. Node 1 is set as the broadcasting

node, while the others are the end nodes. The LED on the Arduino boards would blink on and off

every second if the signal is received. The LED would not light up is the signal is not received.

For experiments with more than 1 ZigBee chip however, when Node 1 tries to broadcast the

signal to Node 3, it was unable to do so due to the limitations of series 1 as seen in fig 4.2. There

is no ‘router’ function in series 1, hence the signal was not able to be transmitted. This shows

that series 1 only works in peer to peer mode.

Fig 4.1: Series 1 peer to peer connection Fig 4.2: Series 1 Failed broadcast


Node 1

Node 2

Node 1

Node 2

Node 3


4.1.2 ZigBee series 2 experiments

Similar experiments were conducted with Zigbee series 2 and information/signal was able to be

sent from 1 Zigbee to another. In order for the signal to be successfully ‘broadcasted’, the nodes

had to be assigned, i.e. there had to be a coordinator, router and the end devices. However for

this project, it is not practical for the nodes to be assigned as such as the robots are meant to be

mobile and would not be able to follow the same assignment as figure 4.3. If the nodes were to

move around like in Fig 4.4, the network would not be able to function and hence the signal

would not be able to be transmitted. Another limitation is that as long as either the coordinator or

the router is down, the network or part of the network would fail. So this would also be a concern

for this project. Hence the Digimesh network was preferred.

Fig 4.3: Series 2 network

Fig 4.4: Series 2 network failure

4.1.3 Digimesh experiments

Lastly, experiments were conducted with the ZigBee series 1, upgraded with the Digimesh

software. 4 nodes were used for this experiment. 1 ZigBee node was assigned as a broadcasting

node, while the other 3 were the receivers. The program used shows if the robots are able to

receive the signal by the help of LEDs on the Arduino boards. If the robot is able to receive the

signal from the base station, the LED would light up for about 5 seconds and the delay in

between lighting up is also 5 seconds.


Node 1(Coordinator)

Node 2(Router

)

Node 3(End

device)

Node 1(Coordinator)

Node 2(Router

)

Node 3(End

device)


The broadcasting node was deployed at a fixed location right at the end on Laboratory T1539.

For the 1st experiment the rest of the nodes were spread out throughout the 1st half of the

laboratories. Node 2 was placed inside T1538, node 3 outside T1537 and node 4 in the middle of

T1535. Signal transmission was successful.

The 2nd part of the experiment sees node 2 taken out and moved into T1534, which is in the 2nd

half of the laboratory. Transmission was successful after the nodes were spread to maximum

distance whereby they can receive the signal. If they are unable to receive the signal, the node

would be moved back to the previous position to ensure that the signal is received. This sets the

precursor to the experiments with robots. The layout of the Robotics laboratory (Fig 4.5) and

the final positions of the nodes are shown below (Fig 4.6).

Fig 4.5: Layout of the Robotics Laboratory in SP.



Fig 4.6: Final layout of the nodes, experiment without robots, using digimesh.

4.2 Experiments with Robots

From earlier experiments conducted, it was concluded that it was not necessary to conduct any

experiments with ZigBee series 1 or series 2 with the Robots as it would fail. Hence only

Digimesh was used for this section.

Initially only 1 type of robot, the four-wheeled Omni robot was supposed to be used for this

experiment. But to show how flexible ZigBee is, different types of robots were used instead of

just one type, as seen in the pictures (Figures 5.5, 5.6 and 5.7) below. Robot 3 (Fig: 5.7) is a Co-

space robot [12], which is a small and fast moving robot. It is also known as RE-Wheelie02 and

is bidirectional. It has sensors for obstruction detection and 2 line sensors for searching missions.

This makes it suitable to be the ‘searching’ robot mentioned in the introduction.



1 ZigBee node was assigned to be the broadcasting node (base station), while the others were

receivers, which were deployed with the robots. These robots are remote controlled by students

from SP for a controlled experiment but the concept would the same if these robots were

automated. The program used is similar to the one used for the experiments without the robots,

the LED blinking when there is a signal received. The base station fixed location and the robots’

starting point are at T1537.

The Robot 1 is deployed and starts moving out of T1537 and into the corridor. It continues to

move along till the signal is lost. The signal got lost after the entrance of the T1534 and moves

back to the last working position. This shows the maximum distance the robot can be deployed

without losing signal. The distance was estimated to be about 12-15 meters.

Robot 2 is then deployed and both 1st and 2nd Robot move concurrently as long as the signal is

received and the LED continues to light up. Once robot 2 loses signal, it moves back to the last

known working position, which happens to be also near the entrance of the T1534. The Robot 1

waits for the Robot 2 to regain the signal before continuing to move till it loses signal as well.

This occurs at the midway point of T1533. Total distance covered would be close to 25 meters.

Finally the Robot 3 is deployed and now all 3 robots move concurrently till signal is lost. The

Robot 3 loses the signal right at the entrance of T1534 and moves back to the last working

position. Then Robot 2 loses signal at the halfway point of T1533, while the Robot 1 still has

signal right at the entrance of T1531. Total distance covered is about 40 meters and maybe more

if there was more area for Robot 1 to move.

The connectivity and signal flowchart was determined with the help of this experiment. As there

were there were limitations to technology available, the author was unable to determine the

signal strength. But however, judging by the way LED lights up, he was able to judge if the

signal is getting weaker. The longer delay between the LED lighting up shows if the signal is

indeed getting weaker. This is due to the fact it takes longer for the signal to be received from the

base station to the nearest robot, which then forwards the signal to the next robot and so on. The

further the robot, the weaker the signal.



From this experiment, it has shown that, as long as the number of robots is increased, the

network area would be widened and hence more ground could be covered. This also shows that

any amount of robots could be used and this network would still function and expand.

The flowcharts below represent the algorithm of the ZigBee mounted robot. The 1st flowchart is

the connectivity flowchart while the 2nd flowchart is the signal flowchart. Both flowcharts would

work concurrently. The connectivity flowchart shows how the robots would link-up with a

known and unidentified robot, while the signal flowchart would react according to how strong or

weak signal. The signal flowchart would control how the robots would move, for example, in the

above scenarios, if the robot loses the signal from the base station, it would move back to the

previous position where there was a signal and remain there until the next robot sets off.

The connectivity flowchart however, will allow unknown ZigBee Robots to join the network

once the known ZigBee Robot/base station identifies it and lets it expand the network and

coverage, which was also explained and conducted in the experiments. Refer to the color of the

boxes highlighted in yellow for the difference between the 2 flowcharts.

The final layout position of experiment, with the robots is shown below.

Fig 4.7: Robot 1 Fig 4.8: Robot 2



Fig 4.9: Robot 3, Co-space robot [12].

Fig 4.10: Final layout of the experiments with robots. using digimesh



The Zigbee Connectivity Flowchart


Is the Zigbee robot able to receive any signal/information?

Signal sent from another source, i.e. Other Zigbees

The source tries to resend the signal

Yes

No

Zigbee A robot responds and moves further ahead/away

If no response after 2 tries

Zigbee Robot moves back to last known working position and remains there till the next

Robot starts moving off


Yes

No

Is it a known source, e.g. Zigbee B?

NoZigbee sends a signal to

identify and join the network with the newly identified

Zigbee XYes

Fig 4.11: The ZigBee connectivity flowchart


The ZigBee Signal Flowchart



Signal sent from another source, i.e. Other Zigbees

The source tries to resend the signal

Yes

No

Zigbee robot responds and moves further ahead/away

If no response after 2 tries

Zigbee Robot moves back to last known working position and remains there till the next

Robot starts moving off


No

Is the signal strength strong or moderate?

NoIf signal is weak, robot moves slowly till it loses signal then

slowly moves back to last working position and remains there till the next robot starts

moving off

Yes

Yes

Fig 4.12: The ZigBee signal flowchart


5. PROJECT MANAGEMENT, PROBLEMS ENCOUNTERED

AND REFLECTIONS

In this section, I would explain on my initial project planning and time management. However,

no project would be able to run smoothly without any hiccups along the way. There were some

problems I encountered along the way, which affected my initial planning and hence faced some

time constraints and made some changes accordingly.

5.1 Project Planning

The diagram below represents the proposed method and approach planned for this project. It acts

as a rough guide and gauge as to how he would approach carrying out this project.


Research on Zigbee and wireless topologies

Run tests to enable communications between 2 robots.

Run tests on 3 or more robots to establish communications with one and another.

Test the Zigbee signal strength on various topologies and distances.

Analyse the data collected from the tests conducted.

Fig 5.1: Flowchart of Proposed approach and method


The project was divided into the following parts, namely

Project Planning, literature and research

Familiarization of Zigbee technology

Testing and experimenting with two Zigbees

Testing and experimenting with three or more Zigbees

Experiments with the Robots

Analyzing the data collected from all the experiments/tests conducted

Final Year Project Report, Poster and Presentation



5.2 Project Management

The Project was divided into several parts as mentioned above. A Gantt chart was used to

provide an overview of the scheduling of activities. The Gantt chart helps to manage the

project’s timeline and helps to keep track of important dates like the project submission and

deadlines. However as things do not always go according to plan, the Gantt chart had to be

amended to reflect the changes. It would be noted that Topologies experiment was conducted

concurrently with the experiments with 2 Zigbees rather than after those experiments due to time

constraints.

Below is the summary of the project tasks list (Table 5.1) and the final Gantt chart (Fig 5.2). The

initial Gantt chart can be found in Appendix A.

No Task Start(DD/MM/YY) End Wireless network for mobile robots 01/08/10 04/06/11

1 Research, Project Planning and Management 01/08/10 06/09/10 Project Management and planning 01/08/10 06/09/10

Project Proposal 01/08/10 06/09/102 Initial work & Literature Review 07/09/10 02/10/10

Familiarization with Zigbee Technology 07/09/10 02/10/103 Experiments with 2 Zigbees 03/10/10 06/11/10

Communication between 2 Zigbees 03/10/10 30/10/10Collection and Analyzing data 03/10/10 06/11/10

4 Experiments with more than 2 Zigbees 28/11/10 23/03/11 Communication between more than 2 Zigbees 28/11/10 23/03/11 Collection and Analyzing data 28/11/10 23/03/11

Topologies Research and Experiment 02/01/11 20/03/11 5 Experiments with Robots 21/03/11 17/04/11

Collection of data 21/03/11 17/04/11 6 Report, Poster and Presentation 04/04/11 04/06/11

Report writing 04/04/11 16/05/11 Poster Preparation 17/05/11 03/06/11 Oral presentation Preparation 04/06/11 04/06/11

Table 5.1: Summary of Project tasks list



Fig 5.2: The final Gantt chart


Note: The boxes highlighted in various colours are the initial planning, while the black boxes are the actual completed datesWeek 1: Aug 1 - Aug 7

Week 2: Aug 8 - Aug 14



Week 5: Aug 29 - Sep 4

Week 6: Sep 5 - Sep 11



Week 9: Sep 26 - Oct 2

Week 10: Oct 3 - Oct 9




Week 14: Oct 31 - Nov 6

Week 15: Nov 7 - Nov 13



Week 18: Nov 28 - Dec 4

Week 19: Dec 5 - Dec 11



Week 22: Dec 26 - Jan 2 2011

Week 23: Jan 3 - Jan 9




Week 27: Jan 31 -Feb 6

Week 28: Feb 7 - Feb 13



Week 31: Feb 28 - Mar 6

Week 32: Mar 7 - Mar 13



Week 35: Mar 28 - Apr 3

Week 36: Apr 4 - Apr 10



Week 39: Apr 25 - May 1

Week 40: May 2 - May 8




Week 44: May 30 - Jun 5


5.3 Technical Problems

Experiments were mainly conducted in Singapore Polytechnic (SP) with the guidance of my

supervisor. In order to save some time and learn more as well, I had to practice and conduct

some experiments at home too. I had very limited knowledge with Arduino and ZigBee and also

my Personal Computer (PC) was giving some problems initially as it was unable to read the

Arduino boards and ZigBee chips for some reason. After resolving the problem with the help of

my supervisor and some online software, I was able to conduct some of the experiments

successfully. For those unsuccessful ones and those I did not know how to do, I tried again in SP

with my supervisor.

There is also very few books available about ZigBee in the libraries and they were hard to get

hold of. Hence most of my references were found online.

Another issue faced was the fact that ZigBee series 1 and 2 had limitations to their capabilities,

as mentioned in chapter 3 and 4, which meant they were not suitable for the project. This was

solved with the help of Digimesh network.

5.4 Personal Problems

I encountered some personal problems along the way as well. Firstly, I had time constraints

throughout the project. For the 1st semester of the capstone project, I had to juggle between my

shift work, school workload as I was taking other modules last semester and lastly the project

itself. As the modules were also very tough, I had to spend more time studying for them. I was

hoping to have more time once the exams were over and concentrate on more on the project

since I was not taking any other modules for the 2nd semester of the project.

However my grandfather’s untimely demise during his overseas holiday towards the end of last

year disrupted my momentum and mood for this project. It was a very trying time as the process

of bringing back his body took a longer time than usual and took a toll especially on my mother

and I wanted to keep an eye on her.

Also, I had taken on some more responsibilities at work and hence had to spend some time

learning to handle the new workload as well.



Lastly, it was hard to get hold of my supervisor at times due to his shift work and also the fact

that my supervisor had to travel often. So we often had to compromise their timings in order to

meet up.

5.5 Reflection

Overall, I think I could have managed time a little better despite the events and some

complacency but I would treat this as a lesson learnt. And I would hope to fare better with time

management in the future.

6. CONCLUSION AND FUTURE DEVELOPMENTS

This chapter is where the conclusion to this project and some suggestions for the future

developments would be discussed.

6.1 Conclusion

After the numerous experiments and amount of research done for this project, I would conclude

that the project was relatively successful.

The objective of the project, which is enabling the moving robots to communicate with the other

robots in the group as well as maintaining communications with the base station using the

ZigBee technology, was met.

My personal objectives were somewhat met as I really learnt a lot from this project about my

abilities and limitations. However, I felt that I could find out and learn more about ZigBee and

wireless technology as it really began to interest me. I also believe I could improve my

programming skills as I realized I am still quite weak in it.

The fact that ZigBee is affordable and flexible could pave the way for more future developments.



6.2 Future Developments

As mentioned in the introduction, robots have already been used for some risky operations like

the bomb disposal robot and the radiation detecting robot in Japan. These are good examples of

remote controlled robots at use and there is no reason why more can be done to make them more

flexible. Some of these robots are wired, so using ZigBee or other wireless means, would make

the robot more flexible, as the user would not worry about tripping on the wire or the wire

getting entangled among the debris.

As mentioned in my experiments, the ZigBee chip can be mounted on any type of robot, so it

could also be fitted onto the robots mentioned above or the search and rescue robot.

Since the experiments were conducted in the confines of the laboratory, the distance gauged was

relatively small. The laboratory had many obstructions like walls and heavy tables which

affected the distance, so it is possible to cover more distance in open spaces. Another possibility

is to use ZigBee pro chips and even range extenders. Range extenders, as the name says, extends

the range of ZigBee chips, especially suitable for the Digimesh network. And hence increase the

area of coverage.

Future experiments can be conducted on the open space and closed space with ZigBee pro chips

and range extenders to see its effective range and add to the usage of ZigBee in the risky

scenarios.

Making the robot automatic and working like a team instead of remote controlled robots would

be an ideal future plan as well. The research done and the algorithm derived here could help

bring the automation one step closer.

The four wheeled Omni directional robots and the co-space robots would have been ideal for the

searching of survivors in a disaster struck area, as they are small, agile and fast and would be

able to cover more distance in a shorter time. But because of space constraints on the robots

itself, they may not able to carry cameras and infra red devices among others. Bigger robots

could be used instead and these robots could also have ‘hands’ or claws to remove smaller debris

if required. This could be another possible future development to look into.



REFERENCES

[1] Digi International – Making Wireless M2M Easy

http://www.digi.com, January 31st 2011

[2] Wireless – Wikipedia, the free encyclopedia:

http://en.wikipedia.org/wiki/Wireless_technology, August 19th 2010

[3] Wireless sensor network – Wikipedia, the free encyclopedia:

http://en.wikipedia.org/wiki/Wireless_sensor_network, August 22nd 2010

[4]Wireless sensor network – Google books: http://books.google.com.sg/books?

id=hor1lIYt5f8C&printsec=frontcover&dq=Wireless+sensor+network&source=bl&ots=04y7GA

4AYY&sig=CYHxE17cLenE6BfTkgWoczDQ0L4&hl=en&ei=1np9TIWRFYW8vgPmieHbAg

&sa=X&oi=book_result&ct=result&resnum=3&ved=0CDEQ6AEwAg#v=onepage&q=Wireless

%20sensor%20network&f=false, August 25th 2010

[5] Wireless sensor networks research group: http://www.sensor-networks.org/index.php?

page=0823123150, August 28th 2010

[6] Network topology – Wikipedia, the free encyclopedia:

http://en.wikipedia.org/wiki/Network_topology, September 15th 2010

[7] Webopedia: The 7 layers of the OSI model:

http://www.webopedia.com/quick_ref/OSI_Layers.asp, September 2nd 2010

[8] Zigbee – Wikipedia, the free encyclopedia:

http://en.wikipedia.org/wiki/Zigbee#cite_note-2, September 12th 2010

[9] Xbee-PRO 802.15.4 (Formerly Series 1) OEM RF Modules

http://www.digi.com/products/wireless/point-multipoint/xbee-pro-series1-module.jsp, November

1st 2010

[10] Xbee-PRO ZNet 2.5 OEM RF Modules- Digi International

http://www.digi.com/products/wireless/zigbee-mesh/xbee-pro-series2-module.jsp, November 2nd

2010

[11] The White Paper, Wireless Mesh Networking ZigBee vs Digimesh

http://www.digi.com/pdf/wp_zigbeevsdigimesh.pdf, February 25th 2011

[12] The Real Robot

http://www.robocupsingapore.org/cospace/cospace-platform/the-real-robot, April 23rd 2011





Appendix A: The initial Gantt chart

*The End*






Week 5: Aug 29 - Sep 4




Week 9: Sep 26 - Oct 2





Week 14: Oct 31 - Nov 6




Week 18: Nov 28 - Dec 4




Week 22: Dec 26 - Jan 2 2011





Week 27: Jan 31 -Feb 6




Week 31: Feb 28 - Mar 6




Week 35: Mar 28 - Apr 3




Week 39: Apr 25 - May 1





Week 44: May 30 - Jun 5

SubramaniamSenthilKumar_FYP

Documents

project supervisor

project report

rescue workers

subramaniam q0704622

technology wireless

abstract wireless technology

type of wireless technology

dr carlos acosta project