Content uploaded by Daniela Ventura
Author content
All content in this area was uploaded by Daniela Ventura on Jul 08, 2015
Content may be subject to copyright.
An Approach for Monitoring and Smart Planning of
Urban Solid Waste Management
Using Smart-M3 Platform
Vincenzo Catania, Daniela Ventura
Department of Electrical, Electronics and Computer Engineering
Catania, Italy
first.last@dieei.unict.it
Abstract—Solid waste management is one of the most im-
portant challenges in urban areas throughout the world and
it is becoming a critical issue in developing countries where a
rapid increase in population has been observed. Waste collection
is a complex process that requires the use of large amount
of money and an elaborate management of logistics. In this
paper an approach to smart waste collection is proposed able to
improve and optimize the handling of solid urban waste. Context
of smart waste management requires interconnection among
heterogeneous devices and data sharing involving a large amount
of people. Smart-M3 platform solves these problems offering a
high degree of decoupling and scalability. Waste collection is made
by real-time monitoring the level of bin’s fullness through sensors
placed inside the containers. This method enables to exempt
from collecting semi-empty bins. Furthermore, incoming data can
be provided to decisional algorithms in order to determine the
optimal number of waste vehicles or bins to distribute in the
territory. The presented solution gives important advantages for
both service providers and consumers. The formers could obtain
a sensible cost reduction. On the other hand, users may benefit
from a higher level of service quality. In order to make users
feel closer to their community, they can interact with the system
to be aware about the fulness state of the nearest bins. Finally,
a mechanism for collecting “green points” was introduced for
encouraging citizens to recycle.
I. INTRODUCTION
Internet of Things will extend and revolutionise the concept
of “cities” making them more comfortable and able to give
intelligent responses to different kinds of events or needs, and
also common “things” and infrastructures will be integrated
with new hardware and software technologies. Interactive
sensing streets or active buildings will gather environment data
about the city and measure the level of noise, traffic, crowds,
temperature - literally everything. Data will be transmitted
to other “smart objects” or “smart controll systems” and
processed by them. This information will be aggregated to
achieve and provide more complete and intelligent services to
citizens, monitor the environment and act quickly in case of
natural disasters, improve decision-making both in public and
private sectors.
Research on smart cities has been conducted by many
organizations and many applications have already been im-
plemented. Typical examples include: smart parking which
supplies the position of a car park at any time [1]; smart
agriculture to improve agroindustrial production based on
weather and environmental conditions [2]; smart transport to
find the best route taking into account the current traffic
conditions [3]. All these applications are a step towards the
realisation of a complete smart city.
Another field of interest that should be made smart con-
cerns the waste collection. All cities, regardless their size,
their geographical location or their economic level, spend huge
amount of money every year for waste collection. The number
of bins located in the streets and the number of vehicles used
to empty them are generally estimated based on the number of
citizens, but the resulting estimation is sometimes either too
high or too low. The natural consequence is the provision of
poor service or to incur in high costs (e.g. the cost of fuel
for too many trucks). Furthermore, the collection of waste,
regardless the type of material (recycling or unsorted), is
typically fixed weekly without taking into account the actual
state of the level of fullness for each bin. The result is
the collection of semi-empty bins or the trash accumulation
degrading conditions of hygiene of the city.
Predicting best time to make the garbage collection and
optimizing the number of vehicles and containers placed on the
streets became feasible operations only if a stream of constant
information on the quantity of waste daily is provided. Less
recent studies on waste management have considered the load-
account analysis [4] considering the amount of waste collected
and disposed in the landfills. Conversely, the diffusion of low
cost mobile devices having longer battery life enables to data
collection on the amount of waste produced directly on-site
and in real-time. Monitoring the fullness of bins through the
use of various types of sensors, it is possible to achieve a
more efficient system. If the system is also able to respond
appropriately to events that occur in real-time by applying
different strategies depending on the event itself, the system
can be defined “smart waste system management”.
Moreover, in a smart city context, it is also important to
allow users to interact with ubiquitous information produced
by the city, anytime and from any device [5]. New user-
centric applications will be developed and they will bring new
forms of interaction, utilizing sensor data production and social
networking. The role of user should be promoted in the field
of the smart waste management as well.
In this paper we propose a solution for smart monitoring
and planning of waste management. We adopted the Smart-
M3 platform [6] that enables us to take advantage of some of
its characteristics such as interoperability, uncoupled commu-
nications, wide range of APIs and ease of implementation.
This solution consists mainly in two phases: a monitoring
phase in which the fullness levels of the rubbish bins is
constantly measured; a computation phase in which collected
data is elaborated for the optimization of the trash collection.
Beside, we have designed an application for users that provides
information about the nearest bins and handles the collection
of green-points with the aim to stimulate users in recycling.
The remainder of this paper is organised as follows:
Section II discusses the domain of interest for this paper and
gives some food for thought. Section III describes the case
study dealt with and the components used. In Section IV, we
summarize the basic concepts of Smart-M3 platform and the
characteristics of which we have most benefited. In Section
V, we discuss the architecture which we have defined and the
ontology that we have used is examined in depth in Section
VI. Finally, in Section VII, we talk about the implementation
and some results obtained simulating the system.
II. DOMAIN OF INTEREST
The planning of a smart waste system is a complex task and
it does not have an univocal solution, but it depends on many
factors and aspects which have to be considered and evaluated.
First of all it is important to narrow the domain of interest and
understand the goals to achieve. The following questions can
help to design and implement a smart waste system. We tried
to answer to these questions to better explain our domain of
interest that we have been taking into account in this paper.
A. What category of waste is a subject matter of the study?
Waste is surplus of various human activities and it can be
categorized in base of origin and degree of danger. The three
main categories are:
•solid waste: they are a highly heterogeneous class
and are non-hazardous waste produced by public or
private (household waste). This category covers both
the unsorted and the recycling and what differentiates
these two types of collection is just the process of
garbage disposal. So, the design of a smart waste
system that deals with both types of waste is equal
until to the transfer of waste in landfills;
•medical waste: this category includes waste from hos-
pitals, pharmacies, veterinarians or other health care
facilities. This type of waste may be liquid or solid as
also infectious. To design a system that deals with the
collection of medical waste, it is necessary to consider
that the collection usually occurs not through public
bins but directly from the manufacturers or sellers of
medical equipment;
•hazardous waste: they include toxic, explosives and
waste that can be harmful to humans, generally are
produced by factories. Designing a smart system for
these types of waste is very complex because the
transport requires the use of special resources able to
isolate them and to avoid that they could contami-
nate environments, lakes, rivers or aquifers and the
collection process is highly dependent on the type of
material to be collected.
In this paper, as said so far, we have dealt with only the
management of solid waste. It is more simple to apply and to
realize in a smart city context. The other two types of waste
management require special licenses or vehicles different from
one country to another and so it’s difficult to achieve a general
smart system for them.
B. What type of collection is used in the area where the smart
waste management should be realized?
Once confined the problem to be addressed only to a
specific category, for us the solid waste category, the next
step is figuring out what type of collection is carried out
in the place where the system should be achieved. In fact,
various collection and container systems are used depending
on the areas where reference is made to. So, it is possible to
distinguish two different modalities:
•door-to-door collection: it is a technique which pro-
vides a periodic pick up service;
•indirect collection: it’s based on the use of containers
or communal bins placed near markets, in apartment
complexes, and in other appropriate locations where
special vehicles go to gather waste.
In both cases, the goal is to plan a system that would to make
smart the collection in relation to the times and days of waste
collection according to the different types of materials. In this
paper, we focus primarily on indirect collection although we
think that the proposed system can be easily adapted also to
the door-to-door collection.
C. What goals does it want to obtain?
There are several aims that a smart system for urban waste
management may want to achieve. We list the main:
•optimize the routing operations;
•sometimes the containers are moved, stolen or dam-
aged. For companies responsible for the service is very
important to be able to trace the positioning of the
different waste bins within the urban or metropoli-
tan areas to reduce theft or monitoring of position’s
change;
•reduce unnecessary costs of companies like the fuel
of vehicle;
•provide a higher quality service to the citizen;
•identify and trace the profile of workers who work in
the collection;
•increase recycling and so to reduce the environment
impact of the waste dumping.
Being able to achieve all these objectives simultaneously is
very difficult because some goals could be in conflict each
other. In our system, we have focused on optimization of
the trash collection, costs, quality of service and increase
recycling.
III. HARDWARE ARCHITECTURE
To achieve a smart waste system, we have used very
different components each capable of interacting with the
smart-space, which will be deepened in the next sections.
Figure 1 shows graphically the hardware architecture used.
In accordance with the vision of the smart cities where it
is required the insertion of components capable to measure
and transmit data about the context in real time, inside the
bins we have inserted two types of sensors: a proximity sensor
located inside the lid or on the the internal and upper part of
the bin; and a weight sensor in the bottom of the container.
The proximity sensor has the task to measure the level of
fullness of the container; the weight sensor is used instead
to measure the amount of trash contained in the bin or thrown
by a user. Each bin has a Zigbee module that is able to
communicate the values of measured physical quantity to the
nearest light pole. Alternative protocols, instead of Zigbee,
could be used (e.g. BLE) because the system logic and the
protocol to communicate sensors data are decoupled thanks to
the use of a gateway component in the light pole.
The gateway was realized through a Raspberry PI and its
most important task is to collect, process and then transmit data
measured by the sensors to the control center. Each Raspberry
PI is powered directly by the power source used to produce the
light of the lamppost, has a Zigbee module to get the values
from the sensors, and a network interface, i.e. one gprs or
wifi shield. Furthermore it is also important to determine the
position of the bins. Since the light pole is close to a very small
group of bins, it is permissible to think that it is directly the
Raspberry PI that associates to each bin the GPS coordinates
recorded.
The control center is the component that uses sensor data
to implement effective and efficient optimization strategies and
to find solutions for problems linked resources organization to
solid waste management. Also it informs any vehicle whether
and when they have to empty bins.
Fig. 1. Hardware components used in the smart waste system
In our system, a city is divided into areas. A list of coor-
dinates’ points is a polygon, i.e. an area. Each waste vehicle
is mapped on an area of the city during the configuration of
the system. The waste vehicles have an onboard computer or a
tablet with the Internet connection. They receive in daily time
slots the list of bins that have to be emptied, according to their
area and to quantity of full bins.
Finally, we have considered each user can interact with
the smart waste system through his smartphones or tablets.
These devices are used for the authentication and login into
the system, to collect “green points” and to know the state of
the nearest bins and in base of waste types.
IV. SMART-M3 PLATFORM FOR SMART WASTE
MANAGEMENT
Smart-M3 platform is an open-source project that provides
an environment in which different entities can share informa-
tion and cooperate in a transparent way to the heterogeneities.
The space in which the agents interact, called “smart-space”,
is a virtualization of the real environment where relevant
informations about real world are stored and kept up-to-date
at every moment. It is based on the use of ontologies to
describe relationships between real entities and to contain
shared data. The exchange of messages happens modifing,
adding or removing subjects, predicates or complements from
ontology itself.
The components of a smart space are: semantic information
broker (SIB) which is the entry point to the RDF graph, and
knowledge processors (KPs), which run on physical devices.
KPs communicate to the SIB their data using the SSAP
(“smart space access protocol”) protocol which defines a set
of messages for reading or modifying the smart-space. More
details on the architecture of Smart-M3 platform are presented
in [7], [8].
Many applications in different domains were implemented
using Smart-M3 platform. Some of these use cases are: a
cross domain application for wellness domain and home
entertainment domain [9] that shows as independent agents
can cooperate each other to provide a high user experience;
another application is in the domain of smart-buildings where
interoperability between furnitures and smart objects made
by different manufacturers is rare [10]; and SmartRoom is
a system that provides a set of digital services for activities
localized in a room and during meetings [11].
The domain of waste management is a context character-
ized by a highly dynamic and fast data production, as well as
by the presence of many users that require this data. The smart
waste management proposed uses Smart-M3 platform to solve
these problems. In fact, first of all, Smart-M3 platform gives
a high level of decoupling between producer and consumer
of data. The participants in the smart-space are unaware of
other participants, their physical characteristics or capabilities.
Therefore, the smart-space acts as an intermediary between
different agents and provides a unique means of communica-
tion understood by all components. So, a producer can send
data to a consumer without the latter having to directly ask
the producer.
A second feature concerns the sharing of knowledge and
access to information on the monitored environment. First of
all, data sharing in the form of RDF triples and data access
through semantic queries, make communication independent of
the operating system or manufacturer. Also, Smart-m3 allows
to have direct access to the freshest and updated data among
multiple devices. An agent can make a subscription on an
information so as it can be notified when data value changes
without making an explicit request to obtain the newest data.
Each agent can customize the type of subscription in order to
obtain personalized services.
Moreover, Smart-M3 platform gives a good degree of
ease to extend and integrate different applications in similar
contexts, through integration of ontologies. This feature is
important in smart cities context in which many innovative
applications could be implemented. For example, our waste
system gives data about position of waste vehicles on the
streets. These data could be aggregate with other information
to realized a smart traffic monitoring system through easily
integration of two or more differents ontologies.
V. S OFTWARE ARCHITECTURE
According to the Smart-M3 view, devices that want to take
part to the smart-space have to implement one or more of KPs.
The smart waste system, shown in Figure 2, is composed by
many types of KPs: two KPs for each light pole, a lot of KPs
for the control center, two KPs for each vehicle and two for
the users’ mobile device. Each of them cooperate and share
data through the smart space, when certain events occur.
In the system that we have developed, some constraints
have been applied:
•waste collection frequencies are daily and it happens
in fixed time slots. In particular to simplify our tests,
we have supposed that each day there are two time
slots: at 5 a.m. and at 2 p.m.. Obviously through a
configuration operation it’s possible to change these
times;
•the types of materials collected are: glass, plastic,
paper and general waste;
•residual waste and recyclables are collected separately;
•separate fleets of vehicles are employed for different
types of materials.
To better describe the proposed architecture, we divide the
description into two sub-sections.
A. Real-time monitoring and intelligent planning of daily
collection operations
Each proximity and weight sensor, placed in each bin,
transmits the measured values to the Raspberry PI with its
own rate using the Zigbee protocol. But for the optimization of
the logistics, we have considered only the values of proximity
sensors, so as to identify which containers are close to their
fill level, while the values of the weight sensors are used in
the next subsection to interact with users. Each Raspberry PI,
placed on a light pole, has two KPs: SensorsLightPole-KP
and CoordsLightPole-KP. Each time that a SensorsLightPole-
KP or a CoordsLightPole-KP perform an update-query on
the smart-space, the control center is notified, but they have
differents goals. The SensorsLightPole-KP updates the sensor
data within the smart space. It may decide to adopt optimized
strategies to perform this publishing operation. For example,
if the value of a certain sensor does not change compared to
the previous measurement, it makes no sense to act on the
smart-space; or it might want to aggregate multiple data for
the same sensor making an average first to communicate with
the SIB. The CoordsLightPole-KP is responsible for updating
the coordinates of the bins. If the bins are moved, a KP of
another light pole will send a notification to the KP of the
control center so to update its personal list.
The control center uses the GarbageLevelManager-KP to
collect data from the various proximity sensors and evalu-
ate the level of filling of a bin. We have considered four
different levels: empty, half-empty, half-full and full. All
information related to the level of bins’ fullness are saved by
GarbageLevelManager-KP in a local database with aims to
provide useful information for offline processing of the data
collected. The history of real data and correspondent times-
tamp can be useful to a decision support system that can find
solutions to the problems of organization of resources related
to the management of solid waste. For example analyzing the
time taken to fill the bins in one of the areas of a city could
help understanding the best number of bins that should be
distributed in that area: if most of the bins are filled in a short
time, the analysis of data suggests the need to add new bins in
that area in order to provide a more efficient service to citizens,
otherwise if the bins are filled in a quite long time, it means
that it could reduce the amount of bins without affecting the
service.
After having known the level of garbage of a bin, one of
the following two cases can happen:
•if the level of a bin belongs to the full or half-
full band, the GarbageLevelManager-KP upgrades the
smart space by creating a connection between the bin
and the vehicle which should gather the area of the
city where the bin is located. Creating this connection
does not mean that the vehicle will surely recover the
bin. This phase of the decision will be implemented
by another component of the control center, that is the
VehicleStatusManager-KP.
•if the level of a bin is in the empty or half-empty
band, the GarbageLevelManager-KP removes the con-
nections between the bin and the associated vehicle to
the area where the bin is located.
Whenever the time band in which the vehicles should
carry out the collection springs, the VehicleStatusManager-KP
applies an algorithm to determine if an area must, may, or
doesn’t require the collection. The algorithm applied by us
compares the total number of bins in an area with full or half-
full ones and marks all areas in clean, little dirty or completely
dirty. More sophisticated algorithms can be integrated into
the system without changing in the structure proposed in this
paper. VehicleStatusManager-KP acts as publish to update the
state of the vehicles of the smart-space. So it assigns the status
in this way:
•All little dirty or completely dirty areas will make the
status of the vehicle to “work”.
•All clean areas will make the status of the vehicle to
“not work”.
The Vehicle-KP is subscribed to the status and according to
this information will carry out the collection or remain in the
garage. The result is a saving in terms of fuel costs for vehicles.
Finally, each vehicle, through the VehicleCoords-KP, updates
data about the position of the truck in real-time. At the moment
this information is not used by our system, but we think it
Fig. 2. Architecture of smart-space with all KPs and SIB
might be useful both to users, because if it is integrated with
others applications, it gives useful information to monitoring
traffic on the streets, and to the control center, to track the
movement of workers.
B. Real-time Monitoring and Incentives for Citizens
In accordance with the vision of smart user-centric cities, a
user has to be able to know the measured values by the sensors
in own city in accordance with his needs and interests. The first
operation that a user can perform through his device, is to know
the status of the closest bins to him. The LevelBinsForUser-
KP has the task of carrying out this functions. It executes a
subscription for the values of proximity sensors according to
the type of collection and level of fullness selected as filter. It
can be turned on and off by the user through the application.
The time interval to receive notifications is fixed by the user.
The second operation that the user can do through his
device, is log in to the system and to collect green-points. The
idea of providing incentives to users who help to carry out
a task or just to encourage them to use a service, it is still a
very popular thing and widely used. The most obvious example
is the collection of points provided by a supermarket chains.
Who hasn’t got at least one card to collect points in their own
wallet? In this case, the user is rewarded with physical goods
(e.g. household items), more quantity of money spent in the
supermarket. Even on the web field, the use of incentives is a
very common practice. In [14], it is investigated the effect
of incentives in web-based surveys showing that monetary
incentives increase the number of people who fill out a survey.
The Internet of Things will enable new type of incentives
for new type of goals. In a smart city, we believe it is possible
to implement a mechanism based on providing incentives to
users in order to increase the amount of garbage to recycle. In
fact, in the architecture described so far, it has been inserted
the chance for citizens to obtain “green points” on the base
of their behavior. At the moment the behavior of the user is
measured based on the amount and the type of waste thrown.
In the future, if it be possible to have a system to evaluate the
material thrown by the user, it will be possible to evaluate also
the qualitative aspect of the collection made for each user and
possibly “punish” him if he commits frequent mistakes.
As previously mentioned, each bin has a sensor to measure
the weight of waste in real-time and a QR-code, applied
on different edges, which encodes the ID of the bin. For
each new sensor values updated in the smart-space, the
BinsWeightManager-KP saves the data in the database. When
a user, through website of the company that manages the
collection, does the authentication (e.g. OpenID), he obtains an
access token. Through the webcam of his smartphone or tablet,
he frames the QR-Code and, at the time when the QR code is
decoded, the User-KP makes an insert-query in the smart-space
adding the user’s identifier connected with the bin’s indentifier
in the ontology. The UsersManager-KP, active in the control
center, is notified and it sent the ID of bid and the ID of
the user to the BinsWeightManager-KP. From this moment
onwards the BinsWeightManager-KP measure the amount of
trash thrown by the user and convert them in “green points” up
to the moment in which the user will do the logout. Obviously
these points are saved on the database of control center. The
green points could be used to provide users benefits or fees
discounts or simply by comparing each others achievements
with friends so create a social real game. The study of finding
the best financial incentives or other type’s incentives, is an
important area but it is altogether another area of research, not
treated in this paper.
Instead of the QR-Code, it could also use authentication
mechanisms through NFC tags. In this case, however, the
proposed smart waste system couldn’t change. The choice of
using the QR-Code is needed by the fact that it is simple to
use, all smartphones or tablet can decode it (while the NFC
is present only on the latest generation devices) and is also
very suitable for use with the so-called “wearable technology”
(such as the “Google Glass”).
As further improvements for user authentication manage-
ment, we want to evaluate the possibility of implementing the
mechanism of policies based on the exchange of public keys
described in [15]. Moreover, to manage the atomic lock on a
single RDF triple, we are considering to extend the architecture
through the mechanism described in [16].
VI. ONTOLOGY
The Smart-M3 platform requires the smart-space is repre-
sented as a RDF ontology on which KPs acting through update,
insert, or delete queries. The classes in the ontology are shown
with the shape of rectangle in the Figure 3, while the circles
are the data properties.
The Sensor class consists of two child classes that dis-
tinguish the two types of sensors in the system: Proximity
and Weight. The choice of logically distinguish the sensors
in smart-space is due to the fact that the two sensors are used
for different tasks. The parent class has two data properties that
allow to save the latest measured value and the corresponding
timestamps. Since each bin contains a proximity and a weight
sensor, it has been created the Contains object property, as it
is possible to see in the Table I.
TABLE I. LIST OF OBJECT PROP ERTIE S USED I N THE S MART-SPACE
Domain Property Range
Bin Contains Sensor
Bin IsLockedBy User
Vehicle Collects Bin
Vehicle AssignedTo Area
The class Bin has the type attribute to indicate the type
of dumpster between glass, plastic, paper or general waste.
In addition for the management of bin’s localization, the
Raspberry PI updates the coords attribute. The management
of the incentives given by the user takes place after the phase
of user login. During this operation, the system creates an
entry of type User with the access token of the user and adds
the object property IsLockedBy between this individual and
the corresponding bin in which the user is throwing waste.
The status of the bin is used by the control center to know
when it has to start the collection of green points of the user.
This attribute can assume the values “lock” when the user is
logged and “unlock” after user’s logout. During the logout
operation, the instance of type User containing the access
token is removed from the smart-space. It should also note that
the User class has no properties because for security reasons
we have avoided to enter passwords or other information
characterizing the user directly in the ontology.
We have decided to divide a city, object of interest, in
some areas, each of which is identified by a list of coordinates.
For this reason, the Area class contains the list of points that
form a polygon. The control center checks in which area the
various bins are placed by comparing their coordinates with
the coordinates of the area and updates its own local list. The
control center for each “full” or “half-full” bin, acts on the
smart-space by creating an object property Collects between
this bin and its area. Obviously, when the real dumpster is
emptied, the proximity sensor will report that the bin is now
empty, and therefore the property Collects will be removed.
Finally, the class Vehicle has the type property to indicate
what type of collection the vehicle is able to perform (depend-
ing on the material), the object property AssignedTo to indicate
the collecting area in which the vehicle has been assigned,
coords to indicate the current position of the truck and a status
that indicates whether the vehicle is going to make a collection
or not. In the case in which the vehicle is in the “work” status,
it will know the list of bins to empty via the object property
Collects made with the various bins.
VII. RESULTS AND APPLICATIONS
All components of the system were implemented using
Python language and the ontology was developed using Pro-
tege.
The implemented system was verified by testing several
use cases. First, communication between bins and the smart-
space was simulated using a weight and a proximity sensor
and a Raspberry PI. In this first test, each sensor transmit
its measured values to the Raspberry through a Xbee module
at regular time intervals. Afterwards, the Raspberry behaves
according the following rule: whenever the value measured by
a sensor exceeds a certain threshold, the measure is updated
on the smart-space along with a timestamp indicating the time
when the measure itself was taken.
Fig. 4. Map resulting of a simulation with three vehicles
Furthermore, we have conducted some simulations to ver-
ify the behavior of the system with a greater amount of nodes.
The Raspberry PI generated a random number of bins (and
hence proximity and weight sensors) assigning coordinates
also them random. The city taken into observation was divided
manually into 3 different areas, visible in Figure 4.
For the control center and the vehicles we have used a PC.
During the simulation, the values of proximity sensors have
been evaluated by the control center and only those with a
label full or half-full was sent to the vehicles. The system
has properly set the status of each vehicle and recognized the
dirtier areas (orange and purple in the image) respect to cleaner
ones (in blue in the figure).
Fig. 3. Graphical representation of the Ontology used in the smart-space. The picture shows only the Data Properties
Two Android applications for users and vehicles are in
phase of implementation. We are using Android SDK to
realize the graphical interface of these applications. The logic
is written in ANSI C. GUI elements are connected with
the application logic using Java Native Interface (JNI), as
explained in [17]. Moreover SmartSlog is used to develop data
structures and variables related with the ontology entities.
Fig. 5. Android Application for Users
In Figure 5, it shows the application for users. It consists
of two main sections: the first allows to view on the map the
closest bins respect to user’s location based on the base of the
filters inserted; the second the other allows to frame a qr-code
and if it is recognized, to lock the bin (c). In order to provide
the location of user we’ve integrated Google Map API. The
different types of bins are identified by different colors on the
map (b). A user can set filters (a) for a better search on the
base of the type of bin (plastic, glass, paper, general) and on
the level of fullness (empty, semi-empty, semi-full, full).
VIII. CONCLUSION
In this paper we have examined the issue of solid urban
waste management. We have proposed an approach based on
Smart-M3 platform through which it was possible to manage
the sharing of data between devices very heterogeneous and
achieve a high degree of scalability and flexibility, important
features for the management of highly dynamic and heteroge-
neous environments typical of smart city context. The smart
system described focuses on two aspects: first of all, it is
addressed to governments and private companies in order to
plan a better management of resources to be deployed in city’s
areas and an optimal planning of waste collection; secondly,
it is aimed at giving citizens the opportunity to know the
position and conditions of the nearest bins and encourage them
to recycling.
There are several future works and improvements for the
proposed system: change the system of users authentication
and atomic lock of bins during the collection of green-point
in accordance with Smart-M3’s features; implement graphical
interfaces for the control center and complete Android appli-
cations; possibility of extending the system adding other use
cases and applications for smart cities. Moreover, the proposed
solution is flexible and decoupled respect to the algorithm
to determine optimal number of bins and vehicles or to the
algorithm to define the best route for vehicles. Therefore,
future works can be made in the study of models that offer
the best results in terms of decision-making.
REFERENCES
[1] S. V. Srikanth, P. Pramod, K. Dileep, S. Tapas, M. Patil, and C. Sarat,
“Design and implementation of a prototype smart parking (spark)
system using wireless sensor networks,” in Advanced Information Net-
working and Applications Workshops, 2009. WAINA ’09. International
Conference on, May 2009, pp. 401–406.
[2] T. Qiu, H. Xiao, and P. Zhou, “Framework and case studies of intelli-
gence monitoring platform in facility agriculture ecosystem,” in Agro-
Geoinformatics (Agro-Geoinformatics), 2013 Second International Con-
ference on, Aug 2013, pp. 522–525.
[3] W. Q. Wang, X. Zhang, J. Zhang, and H. Lim, “Smart traffic cloud:
An infrastructure for traffic applications,” in Parallel and Distributed
Systems (ICPADS), 2012 IEEE 18th International Conference on, Dec
2012, pp. 822–827.
[4] Y. Huang, B. Baetz, G. Huang, and L. Liu, “Violation
analysis for solid waste management systems: an interval fuzzy
programming approach,” Journal of Environmental Management,
vol. 65, no. 4, pp. 431 – 446, 2002. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0301479702905669
[5] A. Vakali, L. Angelis, and M. Giatsoglou, “Sensors talk and humans
sense towards a reciprocal collective awareness smart city framework,”
in Communications Workshops (ICC), 2013 IEEE International Con-
ference on, June 2013, pp. 189–193.
[6] “Smart-m3: Free development software downloads at
sourceforge.net,” November 2013. [Online]. Available:
http://sourceforge.net/projects/smart-m3/files/
[7] J. Honkola, H. Laine, R. Brown, and O. Tyrkko, “Smart-m3 information
sharing platform,” in Computers and Communications (ISCC), 2010
IEEE Symposium on, June 2010, pp. 1041–1046.
[8] Korzun, Lomov, Vanag, Balandin, and Honkola, “Multilingual ontology
library generator for smart-m3 information sharing platform,” Interna-
tional Journal On Advances in Intelligent Systems, vol. 4, no. 4, pp.
68–81, 2011.
[9] J. Honkola, H. Laine, R. Brown, and I. Oliver, “Cross-domain interop-
erability: A case study.” ser. Lecture Notes in Computer Science, vol.
5764. Springer, 2009, pp. 22–31.
[10] K. Framling, A. Kaustell, I. Oliver, J. Honkola, and J. Nyman, “Sharing
building information with smart-m3,” International Journal on Ad-
vances in Intelligent Systems, vol. 3, no. 4, p. 347357, 2010.
[11] D. Korzun, I. Galov, and S. Balandin, “Development of smart room ser-
vices on top of smart-m3,” in Open Innovations Association (FRUCT),
2013 14th Conference of, Nov 2013, pp. 37–44.
[12] S. Longhi, D. Marzioni, E. Alidori, G. Di Buo, M. Prist, M. Grisostomi,
and M. Pirro, “Solid waste management architecture using wireless
sensor network technology,” in New Technologies, Mobility and Security
(NTMS), 2012 5th International Conference on, May 2012, pp. 1–5.
[13] M. Arebey, M. Hannan, H. Basri, and H. Abdullah, “Solid waste
monitoring and management using rfid, gis and gsm,” in Research and
Development (SCOReD), 2009 IEEE Student Conference on, Nov 2009,
pp. 37–40.
[14] J. Su, P. Shao, and J. Fang, “Effect of incentives on web-based surveys,”
Tsinghua Science and Technology, vol. 13, no. 3, pp. 344–347, June
2008.
[15] Boccacci, Bordoni, Cicora, Fabbri, and Prati, “Smart m3 a secure chat
application,” 2013.
[16] DElia, Honkola, Manzaroli, and Cinotti, “Access control at triple level:
Specification and enforcement of a simple rdf model to support con-
current applications in smart environments,” in Smart Spaces and Next
Generation Wired/Wireless Networking, ser. Lecture Notes in Computer
Science, S. Balandin, Y. Koucheryavy, and H. Hu, Eds. Springer Berlin
Heidelberg, 2011, vol. 6869, pp. 63–74.
[17] Kovyrshin and Korzun, “Programming android client for smart-m3
applications: Smartroom case study,” 2013.