Design and Implementation of Human Mobility Embedded System

for Urban Planning of Smart City

Anwar Al-Khateeb

Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioingegneria,

Piazza Leonardo da Vinci, 32, 20133, Milano, Italy

Keywords: Human Mobility, Smart City, Urban Planning, Embedded System, Optimization.

Abstract: The mobility model will help us to simulate the movements of people more realistic as in real case. This

paper developed human mobility system using random way point and random walk methods to model the

individual movement of person in different places within area of Baghdad, Iraq. There are a lot of types of

mobility model but we use Random Way Point (RWP) because it is simple, most common use and needs

less memory and time for computation. The geographical information and maps of the cities are used to tell

the person about the constraints and correct the direction of motion in different areas to generate more

accurate data. Tested areas are quite different in style, structure, social, history and culture. The most similar

thing between them is that they are both in continuously changing which is the one of biggest difficulties in

designing the city. Al-Sadar city is one of biggest population city in Baghdad, Iraq. It has three millions

people population. We choose 5000 users moves from Al-Sadar city to two different places in Baghdad: Al-

Khadmia and Al-Zawra. Al-Khadmiia is old and religion place and Al-Zawra is biggest public garden in

Baghdad. In our work, we find the best ways that connect Al-Sadar city with both places. We decide the

best ways depending on distances and traffic between them. It gives good method to build smart

transportation and smart city using mobility and traffic models. .This paper will help urban decision maker

to suggest the analytical model for urban space and have a clear picture about the city: how can it change,

how do the people move in the city, what are the problems and how can solve them to generate smart city.

1 INTRODUCTION

The smart cities market size is estimated to grow

from USD 312.03 Billion in 2015 to USD 757.74

Billion by 2020, at a Compound Annual Growth

Rate (CAGR) of 19.4%. The base year considered

for the study is 2014 and the market size is

forecasted from 2015 to 2020.

Smart cities market has been segmented into four

major focus areas across regions (Vicente Casares-

Giner et al., 2011; Ignacio Martinez-Arrue et al.,

2008). The focus areas include transportation,

utilities, buildings, and smart citizen services.

Technological advancements in Information and

Communication Technology (ICT) and growing

demographics & hyper-urbanization are the major

driving factors for the market. Cities across the

world are increasingly adopting smart solutions for

their various sectors, such as buildings,

transportation, utilities, and citizen services in order

to achieve better living standards, increased

efficiency, economic stability, optimization of

energy, and higher environmental protection based

on the technologies: IoT, cloud, mobile, and sensors.

Estimates forecast the smart energy technologies

market (including smart grid) reaching $220 billion

worldwide by 2020, whilst other sources estimate

smart transport to be $156 billion, and smart water

to be $22 billion globally by 2020 (Mir et al., 2006;

Tamás Szálka et al., 2009). The example of

Singapore next generation smart city. Singapore

aims to have 80% of all its buildings meet its

minimum ‘Green Mark Certified' energy efficiency

standards by 2030. We identified eight key aspects

that define a Smart City: smart governance, smart

energy, smart building, smart mobility, smart

infrastructure, smart technology, smart healthcare

and smart citizen.

Geographic information systems, often known as

GIS, are computer-based tools for the capture,

analysis, storage, manipulation and visualization of

geographic information. Companies, researchers and

individuals use GIS to analyze spatial data and

Al-Khateeb, A.

Design and Implementation of Human Mobility Embedded System for Urban Planning of Smart City.

DOI: 10.5220/0006374103370345

In Proceedings of the 6th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2017), pages 337-345

ISBN: 978-989-758-241-7

337

create maps of many different types. GIS-created

maps can store and display large amounts of data

within a single map (Yu-Liang Tang et al., 2010;

Son et al., 2004; Tamás Szálka et al., 2009; Peppino

Fazio et al., 2012).

According to the Environmental Protection

Agency, a GIS, which refers to Geographic

Information System, works by combining database

functions with computer mapping to map and

analyses geographic data. It uses a "layering"

technique to combine various types of data. Special

GIS software is used to analyse layered data and

create new layers of data (Broch et al., 1998;

Francesco Calabrese et al., 2010; Heiko Bauke et al.,

2007).

GIS maps can be used to show an estimated

number of people living in a given region. Because

these maps are created after careful considerations of

various data, the results are usually quite accurate.

These maps can be quite specific and show the

number of individuals in a region according to

profession. Users can actually tell the number of

doctors, lawyers or policemen in a region by simply

going through a GIS map (Injong Rhee et al., 2011;

Jae-Hyung Jeon et al., 2013).

GIS maps are capable of giving users a rough

idea of each region on the map and what it is prone

to. For example, a map can depict flood-prone

regions against the landmarks situated close by.

Researchers use these maps to analyze the

characteristics of a given region over a period. This

help in developing strategies to combat issues such

as crime, flooding and any other form of disaster

(Chellappa Doss et al., 2004; Berk Birand et al.,

2011; Matteo Leccardi, 2005).

Disadvantages of using a geographic information

system, or GIS, are that its technical nature might

portray results as being more reliable than they

actually are, and errors and assumptions can be

hidden, leading to a lack of questioning into the

results. Another disadvantage of analyzing the

results from a GIS is that the results will only be as

accurate as the data that they come from. Because of

this, the data may not be able to serve different

contexts, particularly if the data is not applicable

(Francesco Calabrese et al., 2010; Matteo Leccardi,

2005).

For instance, if the input data on a GIS is entered

at the county level, the results in the GIS will only

be usable for the county level, not any other level,

such as the district or ward levels. Data availability,

in itself, is also a major issue. If the data is not

available, than the GIS system is useless.

Furthermore, GIS systems are not like other

programs. They do not come "off the shelf," which

means that they must be assembled and constructed

to a user design. This could be a long, complex and

costly process. Because of this, many GIS systems

don't come to fruition or fail outright in their

implementation because their creation was rushed or

inadequately planned (R. Chellappa Doss et al.,

2004; Jae-Hyung Jeon et al., 2013).

GIS systems are often so complex, in fact, that it

becomes difficult to describe the intangible benefits

they may provide, making it difficult to find funding

for their creation. Also, the technology behind GIS

technology expands rapidly, causing GIS systems to

have a high rate of obsolescence. It's also very

difficult to make GIS programs that are both fast and

user friendly. GIS systems typically require complex

command language (Chellappa Doss et al., 2004;

Broch et al., 1998). Data fields and their

accessibility are also not very understood, and data

can become incomplete, obsolete or erroneous,

rendering the GIS misleading.

In this paper, we introduced the mobility model

which help us to simulate the movement of users

with more realistic as in the real case. There are a lot

of types of mobility model depending on the

characteristics of nodes such as historical model,

correlated model and geographical depending. Each

method has its advantages and disadvantages. In our

work we use random way point method (RWP). It is

simple; most common used in mobile networks and

needs less memory and time for computation. It fits

our requirements. Also we use random walk method.

It is more realistic and near to human walk.

But the goal for our system is to find the location

of users and decide the best way that can be used.

Therefore we need also to predict the position of

users. We use Polynomial method as a predicted

method. It gives accurate and fast prediction of user

position.

Beside the individual movement of user, we use

the relative movements (within the area). We used

geographical mobility depending using map from

openstreet program and parsed the geographical

constraints using Matlab (ways, nodes and tags) to

present the required area.

Using the mobility model is very useful to

decide the way taken by user and improve it. It helps

the city planner to have correct decision depending

on the data given from mobility model. It is one step

to model smart city with enhanced transportation

and improved the ways taken by users.

Using mobility model is much easier and

accurate compared with GIS. It is more realistic and

near to human walk. Also it takes into account the

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

338

nature of the tested area by adding geographical

mobility to the original mobility model.

In this paper, we show in section 2 the previous

works for mobility and prediction. Sections 3 and 4

study the mobility algorithms. They give the

mathematical models of RWP, random walk and

Polynomial methods. They explain the principles

and algorithms details of mobility and prediction

methods. The simulation results and figures are

shown in Section 5. We use Matlab and openstreet

map for Al-Sadar city in Baghdad, Iraq to simulate

mobility, prediction and urban planning

management algorithms. The mobility tested areas

include: Al-Khadmid and Al-Zawar. They are both

in centre of Baghdad, Iraq. The best ways between

these places and Al-Sadar city was founded using

mobility and traffic models. Finally the conclusion

and suggestion for future works is in section 6.

2 RELATED WORK

In (Ignacio Martinez-Arrue et al., 2008) a new

model to predict the location of a person over time

based on individual and collective behaviours. The

model is based on the person’s past trajectory and

the geographical features of the area where the

collectively data are organized in terms of land use,

points of interests and distance of trips.

The mobility aspects and some basic background

on mobility are provided in (Son et al., 2004). They

are being used in performance evaluation of relevant

mobility management procedures. While in (Mir et

al., 2006; Heiko Bauke, 2007) the same group

proposed a new mobility model as an extension of

the random walk model. It gathers mobility patterns

with several degrees of randomness, so that both

random walk and totally directional mobility

patterns are modelled. This model is used as input to

study and compare the location management cost of

the distance-based and movement-based strategies as

a function of the mobile terminal directional

mobility patterns.

In (Tamás Szálka, 2009; Matteo Leccardi, 2005)

authors proposed a mobility prediction algorithm

based on dividing sensitive ranges. In the future

distance prediction scheme proposed in (Peppino

Fazio et al., 2012; R. Chellappa Doss et al., 2004), a

node predicts its own future position from its current

position, speed, and direction.

Tuduce and Gross in (Matteo Leccardi, 2005)

present a mobility model based on real data from the

campus wireless LAN at ETH in Zurich. They used

a simulation area divided into squares and derive the

probability of transitions between adjacent squares

from the data of the access points.

MoVes (Berk Birand et al., 2001; Injong Rhee et

al., 2011) is an embedded system generating

vehicular mobility traces and also containing a basic

network simulator. The major asset of this project is

its ability to partition the geographical area into

clusters and parallelize and distribute the processing

of the tasks from them, which improves the

simulation performance. Although the mobility

model reaches a sufficient level of detail, the

project’s drawback is the poor network simulation.

MOVE (Injong Rhee et al., 2011) contains a

single graphical user interface for the configuration

the mobility modelling and network simulation.

However, MOVE does not itself include a network

simulator, but simply parses realistic mobility traces

extracted from a micro-motion model.

In (Mir et al., 2006), authors also proposed an

integrated vehicular and network simulator. As all

solutions proposed by this approach, the authors

developed their own traffic and network simulator.

In (Ignacio Martinez-Arrue et al., 2008;

Francesco Calabrese et al., 2010) the City Form Lab

has released a state-of-the-art toolbox for urban

network analysis. As the first of its kind, the

Centrality Tools this ArcGIS toolbox can be used to

compute five types of graph analysis measures on

spatial networks: Reach; Gravity; Betweenness;

Closeness; and Straightness. Redundancy Tools

additionally calculate the Redundancy Index,

Redundant Paths, and the Wayfinding Index. The

toolbox requires ArcGIS software with an ArcGIS

Network Analyst Extension.

3 INTRODUCTION FOR

MOBILITY MODEL

In this section we give the definition of mobility

pattern, types, analysis and effect on the

performance of the network. Studying mobility

models will help us to choose the best model that

can fit with our requirements. For example, the

nodes move with random walks are different from

the nodes with spatial dependency (move in group or

follow leader of group).

In our work, we use the movement of individual

users (location, velocity, direction) as well as the

movement relative to given area.

The mobility models are divided into different

classes. They are classified depending on their

characteristics such as history dependence, spatial

Design and Implementation of Human Mobility Embedded System for Urban Planning of Smart City

339

dependence and geographical dependence (Matteo

Leccardi, 2005; Jae-Hyung Jeonet al., 2013).

Random walks are example of history

dependence where the nodes change their speed,

destination and direction randomly in each step. In

temporal mobility model the velocities and

directions of nodes at different slots are correlated

such as Gauss-Markov Mobility Model. While the

streets, freeways and geographical restrictions are

important in designing the path mobility model

In our work, we use Random Way Point (RWP)

and geographical restriction model. RWP is most

common way used in different networks. It was first

proposed by Johnson and Maltz (Francesco

Calabrese et al., 2010). In simulation of RWP, the

nodes start by choosing the starting point and

velocity randomly.

The velocity is uniform distribution with ranges

[0, Vmax] where Vmax is maximum velocity

allowable by node. The velocities and targets points

for given node are not correlated with other nodes. If

Vmax is high then the node will move fast and

mobility will be dynamic otherwise it will be stable.

The average length of transition of all nodes is equal

to the average length of single node at different

times. Therefore RWP has mean-ergodic properties.

The PDF of transition length for rectangular field

(as in our case) is Eq. (1). A rectangular area with

length a and width b. Without losing the generality,

we assume that b≤ a. The mean and variance of

transition length L are given in Eqs. (2-3)

respectively (Jae-Hyung Jeon et al., 2013).

(1)

(2)

(3)

The expected time transition between two successive

steps is:

E[T]=(E[L].(ln(Vmax/Vmin)))/(Vmax-Vmin)

RWP is memoryless process because each node does

not correlate with other node. In our case we choose

RWP because it is simple and fit the requirements

for not correlated users move on map in random way

with velocities of range [Vmin Vmax].

4 POLYNOMIAL PREDICTION

ALGORITHM

If P(.) is a polynomial function with k + 1 unknown

coefficients as follows:

(4)

The previous location data at H-times compose an

original data sequence S(p) where (p = 1, 2, …, H).

If R

is the location of the previous h-th process,

the sum of the square of the difference between

actual R

and predictive P(t) of the previous n

processes is defined as

(5)

Each a

in each polynomial can be treated as a

variable. The coefficient a

can be determined by

taking the partial differential of each a

in equation

(5), and setting each partial differential equation to

zero. At the end, we can obtain the following

equivalent polynomials:

(6)

After determining all the summations of t in equation

(6), this matrix of polynomials can simplify to an

upper triangle matrix of polynomials. Each

coefficient a

can be determined by the native Gauss

elimination method, Consequently, the polynomial

function of P(t

n+1

) can be obtained,

In our work, we will use the polynomial

regression mobility prediction to predict the future

user position. For the velocity we take the second

derivative of Eq. (5) and equal the results for zero.

Also we can use geographical information after

applying the predicted values on map and see the

direction if there is forbidden area or error in

prediction direction.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

340

5 MOBILITY MODEL AND

SIMULATION RESULTS

In our system, we have the requirement of design the

smart city is with good quality of services as shown

in Figure 1. It is specified in abstract layer where all

arithmetic, optimized ways and solutions are used.

The constraints and the measurements are used to

choose the implantation ways in the abstract layer.

In first stages we chose the software platform and

architecture and in last stages we proposed the

services providing and advance performance

monitoring of city urban planning.

The lower layer (architecture and space design)

includes: fading, noise and shadowing models,

mobility management (RWP and RW models),

mobility prediction models (Polynomial model),

network layout and traffic model. Architecture layer

takes all these information, processes and sends the

performance to the upper layer (Application layer)

where evaluate, visualize and monitor the design

according to the requirements and constraints.

The simulation program is divided into three

phases: initial phase, processing phase and urban

planning management phase. In the initial phase, we

select the initial values such as number of users,

number of steps, x, y and velocities for all users. The

initial values of one user are not correlated with other

users. It used dependent on the traffic model after

entering the intra-segment components and inter-

segment components. They are taken from tested

field. The values of x and y are within the limit of our

selected rectangular simulation area. The velocities

are in uniform distribution and randomly selected

within the range [Vmin, Vmax]. The time for the first

phase is within t_int. The second phase is the

processing phase where all computations and

optimizations will run in this phase. It is divided into

two parts: mobility models and prediction of user

location, velocity and direction. The mobility models

RWP and random walk. For prediction side, we use

polynomial method. At the beginning, RWP chooses

randomly the destination and velocity and repeats

this in each step till it reaches to steps_end then it

will go to next which is prediction. The same we do

with Random Walk method.

The prediction methods will choose polynomial

method. It is accurate, simple and need less memory

storage for coding and processing. The mobility in

our work deals with individual movement of users

and relative movement within the selected simulated

area.

For prediction method, we first predict the value

of velocity using the last values of time, x and y of

each user and move to direction and the location of

users at selected time. The time required at the end of

this step is t_int+t_processing. At the end of this step

we have the predicted values of location, velocity and

direction of user movement at selected time.

The last and important step in our system is the

urban planning calculation, optimization and testing.

The saving and advantage for design the city and

how our work can help the city. The time required at

the end of this step is t_int+t_processing+t_manag.

The Dijkstra algorithm is used to get the nearest

and optimum way between the source and

destination points.

To test the system with all phases, we use

openstreet program. The simulated area is one place

in Baghdad, Iraq called Al-Sadar City. It has high

population in Baghdad. There is serious need to plan

and manage the ways to enter and leave it. In our

work, he users move from Al-Sadar City to two

places; Al-Khadmia and Al-Zawra. Al-Khadmia is

religion place and old city in centre of Baghdad. The

map of Al-Khadmia is shown in Figure 2-a. Al-

Zawra is biggest public garden in Baghdad. The map

of Al-Zawra is shown in Figure 3-a.

We export files with type xml.osm files for Al-

Khadmia city and Al-Zawra and parsed the data

using Matlab such as:

Bounds: [2x2 double]  axis limitations of our

areas

Node: [1x1 struct]  node_id, node_xy

Way: [1x1 struct]  combination of nodes,

node_id, way_id, way_tag

For our area we have

id: [1x14626 double]

nd: {1x14626 cell}

tag: {1x14626 cell}

Figures 2-a and 3-a show the parsed files with

ways for Al-Khadmia and Al-Zawar areas

respectively. Figures 2-b and 3-b show the area with

nodes and complete ways in Al-Khadmia and Al-

Zawar areas respectively.

The connectivity of nodes and ways are parsed in

connectivity matrix using Matlab as shown in the

following:

Connectivity matrix =

(68,5) 1

(98,5) 1

(100,20) 1

The connectivity matrix is parsed matrix and “1”

here means there is connection between the nodes.

To decide the connection and route between ways we

Design and Implementation of Human Mobility Embedded System for Urban Planning of Smart City

341

choose Dijkstra's algorithm to decide the shortest

path between nodes and ways.

Dijkstra's algorithm is invented by scientist

Edsger Dijkstra in 1956 and published in 1959

(Matteo Leccardi, 2005; Jae-Hyung Jeon et al.,

2014). It is used to find the shortest path between two

points in graph.

We use Dijkstra's algorithm to find the shortest

path in each step and the best route to reach the

destination point for RWP model in our limited area

(Baghdad). In each step there are number of hops

(nodes) to pass during the complete route. The parsed

data provides us with node_id and node_xy of each

hope.

The number of users in our work is 5000 users

move from Al-Sadar city to both Al-Khadmia and

Al-Zawara. For RWP point we use 400 steps for ech

user. After running the program with random 5000

user with 400 steps for each user in both places (Al-

Khadmia and Al-Zawara), we get the best ways from

Al-Sadar city to Al-Khadmia and Al-Zawar. The

program decides the best ways depending on

Dijkstra's algorithm for l and traffic in both tested

areas.

In this case, we solve the one of biggest problem

in Baghdad. Al-Sadar city has a population for more

than 3 million people. It needs to plan and manage in

smart way. It is one example how to solve the

problems and build smart city.

6 CONCLUSIONS

Transportation and choosing the best way are very

important factors in designing the smart city.

Needing for designing the smart transportation is

increased with increasing number of people and the

cars.In our work, we study the mobility of users and

predict their locations in the tested area. It will help

the designer to choose the best way in advance to

cover the users and have the chance to manage

congestion and traffic for different ways within the

tested area.

We use RWP and Random Walk models to

represent the individual mobility of a user and use

geographical dependent to represent the relative

mobility (corresponding to the tested area). The

mixing models provide more realistic representation

for the movement of users.

For prediction side we used polynomial method

to calculate the user’s velocity, direction and location

at given time slot. Also we deal with traffic and

number of users when decide the ways within the

tested areas and optimized the best way taken by

users.

Al-Sadar City in Baghdad, Iraq is chosen as

tested area. It has high population near 3 millions

people. The users move from Al-Sadar city to

different places: Al-Khadmia and Al-Zawra. They

are different areas. Al-Khadmia is religion and old

city. It has also big market. The second one, Al-

Zawra, is biggest public garden in Baghdad. It has

big green places, games and restaurants. In Baghdad

there is serious need to manage and plan the best

ways to to enter and leave Al-Sadar city.

In our work, we find the best ways to connect Al-

Sada city with both Al-Khadmia and Al-Zawra using

mobility model and traffic model in both places. It

will help the designer to take the correct decision

when design and manage the ways and transportation

of the cities. It will help us to build smart cites with

good planning and smart transportation ways.

REFERENCES

Vicente Casares-Giner_, Vicent Pla, and Pablo Escalle-

Garc´ıa,” Mobility Models for Mobility Management”,

Next Generation Internet, LNCS 5233, pp. 716–745,

2011.

Ignacio Martinez-Arrue, Pablo Garcia-Escalle and Vicente

Casares-Giner,” Location Management Based on the

Mobility Patterns of Mobile Users”, Wireless Systems

and Mobility in Next Generation Internet Lecture

Notes in Computer Science Volume 5122, 2008, pp

185-200.

Yu-Liang Tang, Chun-Cheng Lin, Yannan Yuan1 and

Der-Jiunn Deng,” Dividing Sensitive Ranges Based

Mobility Prediction Algorithm in Wireless Networks”,

Tamkang Journal of Science and Engineering, Vol. 13,

No. 1, pp. 107_115, 2010.

Son, D., Helmy, A. and Krishnamachari, B., “The Effect

of Mobility-Induced Location Errors on Geographic

Routing in Mobile Ad Hoc Sensor Networks: Analysis

and Improvement Using Mobility Prediction,” IEEE

Transactions on Mobile Computing, Vol. 3, pp.

233_245 (2004).

Mir, Z. H., Shrestha, D. M., Cho, G.-H. and Ko, Y.-B.,

Mobility Aware Distributed Topology Control for

Mobile Multi-Hop Wireless Networks, in Proc. of

ICOINS 2006, Vol. 3961 of LNCS. 2006, pp. 257_266.

TAMÁS SZÁLKA, SÁNDOR SZABÓ, PÉTER FÜLÖP,

”Markov model based location prediction in wireless

cellular networks”, Info-communication journal, 2009.

Peppino Fazio, Salvatore Marano,” A New Markov-Based

Mobility Prediction Scheme for Wireless Networks

with Mobile Hosts” International Symposium on

Performance Evaluation of Computer and

Telecommunication Systems (SPECTS), 8-11 July

2012.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

342

R. Chellappa Doss , A. Jennings, N. Shenoy, “A Review

on Current Work in Mobility Prediction for Wireless

Networks”, Proceedings Third Asian International

Mobile Computing Conference, Bangkok, Thailand,

26-28 May 2004.

J. Broch, D. A. Maltz, D. B. Johnson, Y.-C. Hu, and J.

Jetcheva, A performance comparison of multi-hop

wireless ad hoc network routing protocols, in

Proceedings of the Fourth Annual ACM/IEEE

International Conference on Mobile Computing and

Networking (Mobicom98), ACM, October 1998.

Berk Birand, Murtaza Zafer, Gil Zussman, Kang-Won

Lee,” Dynamic Graph Properties of Mobile Networks

under Levy Walk Mobility”, Eighth IEEE International

Conference on Mobile Ad-Hoc and Sensor Systems,

2011.

Injong Rhee, Minsu Shin, Seongik Hong, “On the Levy-

Walk Nature of Human Mobility”, IEEE/ACM

TRANSACTIONS ON NETWORKING, VOL. 19,

NO. 3, JUNE 2011.

Francesco Calabrese, Giusy Di Lorenzo, C. Ratti, “Human

Mobility Prediction based on Individual and Collective

Geographical Preferences”, 2010.

Matteo Leccardi,”Comparison of Three Algorithms for

Lévy Noise Generation”, 2005.

Jae-Hyung Jeon, Eli Barkai, Ralf M, etzler, ” Noisy

continuous time random walks”, final report, 2013.

Heiko Bauke,’ Parameter estimation for power-law

distributions by maximum likelihood methods’,

August, 2007.

Design and Implementation of Human Mobility Embedded System for Urban Planning of Smart City

343

Figure 1: Schematic block diagram of mobility and city urban planning system.

SMARTGREENS 2017 - 6th International Conference on Smart Cities and Green ICT Systems

344

a) Parsed the map using Matlab program for Al-Sadar City

and Al-Khadmia places in Baghdad, Iraq

a) Parsed the map using Matlab program for Al-Sadar City

and Al-Zawra places in Baghdad, Iraq

b) Parsed nodes from the map using Matlab program for

Al-Sadar City and Al-Khadmia places in Baghdad, Iraq

b) Parsed nodes from the map using Matlab program for

Al-Sadar City and Al-Zawraia places in Baghdad, Iraq

Figure 2: Openstreet Map for for Al-Sadar City and Al-

Khadmia places in Baghdad, Iraq with parsed information

using Matlab.

Figure 3 Openstreet Map for for Al-Sadar City and Al-

Zawra places in Baghdad, Iraq with parsed information

using Matlab.

44.34 44.36 44.38 44.4 44.42 44.44 44.46 44.48

33.33

33.34

33.35

33.36

33.37

33.38

33.39

33.4

33.41

33.42

Longitude (

)

Latitude (

)

OpenStreetMap osm file

44.36 44.38 44.4 44.42 44.44 44.46 44.48

33.28

33.3

33.32

33.34

33.36

33.38

33.4

Longitude (

)

Latitude (

)

OpenStreetMap osm file

44.36 44.38 44.4 44.42 44.44 44.46 44.48

33.28

33.3

33.32

33.34

33.36

33.38

33.4

Longitude (

)

Lati tude (

)

OpenStreetMap osm file

44.34 44.36 44.38 44.4 44.42 44.44 44.46 44.48

33.33

33.34

33.35

33.36

33.37

33.38

33.39

33.4

33.41

33.42

Longitude (

)

Latitude (

)

OpenStreetMap osm file

Design and Implementation of Human Mobility Embedded System for Urban Planning of Smart City

345