APPROACH MULTI-AGENTS EMBEDDED ALARM IN POTROOMS

David C. Wyld et al. (Eds) : ICAITA, CDKP, CMC, SOFT, SAI - 2016
pp. 83– 95, 2016. © CS & IT-CSCP 2016 DOI : 10.5121/csit.2016.61308
APPROACH MULTI-AGENTS EMBEDDED
ALARM IN POTROOMS
Luis Carlos Costa Fonseca1
, Alan Robert da Silva Ribeiro2
and Jonh
Selmo de Souza do Nascimento3
1,2,3
Department of Computer Engineering and System,
State University of Maranhão, São Luis, Maranhão, Brazil
1
luiscarlos.uema@gmail.com
2
alan.robert2617@gmail.com
3
jonhselmo@pecs.uema.br
ABSTRACT
Industrial Shop Floor environments require fast intervention of the controller’s computers and
operators to ensure high industrial production efficiency. This work focuses the electrolytic
potrooms process control efficiency. The main goal of this work is to design an embedded
solution for detection and alarm, using multi-agents system technologies so that controllers can
alert plant operators about the problems in the electrolytic pot still malfunctioning after
controller intervention. If the controller action was unsuccessful due to a feeder and pot bus
problem, an audio alarm is immediately issued to the potrooms so that the operator can be
notified about the specific problem, independently of the potrooms location.
KEYWORDS
Alarm, Potrooms, Agents, Embedded Systems, Pot
1. INTRODUCTION
In the aluminum production process, the cost and the quality of the product is strongly affected by
the amount of alumina added to the pot production. The resistance and variations of first and
second orders are used to assess the percentage of the alumina concentration in the bath [1]. The
resistance variation is the primary variable used as an evaluation metric of vessel behavior. In
practical terms, the monitoring and control of alumina in the bath changes are implemented in
accordance with the assessment of the resistive variation. There are situations, however, where
the electrolytic process control computers, even sending signals to the alumina feeding actuators
and movement of buses that regulate the resistance of the electrolytic cells allow better control of
the amount of alumina in the pot; this guarantee control may be interrupted by any mechanical
failures in actuators and busbars.
Therefore, rapid actuation of the vessel operator to reestablishment of normal operation of the
equipment and continuity of the aluminum production process with higher production gain
becomes a daily preoccupation of management, to the extent that an aluminum production room
comprises hundreds of devices distributed over large production halls. Figure 1 shows the Alumar

84 Computer Science & Information Technology (CS & IT)
potroom which has three production lines with 204, 250 and 256 electrolytic cells, each, for a
total of 710 electrolytic cells.
Figure 1. Reduction Alumar
Note that potrooms are arranged in large halls in which the electrolytic cells are disposed in
series, side by side. These pots are controlled by the process control computers, which send
signals to drive the feeders allowing it the addition of alumina in the pot. In addition, the control
computer is responsible for sending signals to drive the mechanical drive bus, responsible for the
regulation of the electrolytic tank resistance variation.
When, perhaps a mechanical failure occurs in one of these devices, we have as a consequence a
delay in the correction and re-establishment of the operation of the electrolytic pot, causing the
excess in the bus drive and alumina supply in the pot, due to unauthorized drives by computer
control generating the accumulation of alumina or the lack of it, causing increased resistance to
passage of electric current and negatively affecting the efficiency in aluminum production.
Thus, it becomes necessary to adopt a technique that best fits the application of electrolytic
process control computer, in order to provide the necessary information and alerts and can thus
overcome and minimize problems that may arise, for example, anode effect, unscheduled
shutdown pot, bored pot, or quality time requirements and life for the potrooms.
Thus, an embedded solution with communications technology based on multi-agent systems is
built based on stimulus-response model, in order to map and detect all abnormalities prevent the
correct operation of the electrolytic pot, even with sending signals performance controllers. If
signs of activity are not sufficient to establish the acceptable level of resistance change, and
consequently, of alumina in the electrolytic pot, then audio alarms are generated immediately for
operators and technicians of the potrooms (factory floor) can understand the situation and carry
out the necessary intervention to solve the physical problem of equipment. The system is
accessible and can be easily implemented in the automation industries that have controllers such
as PLC (programming logic controller) and process control computer.
The technologies investigated in this research are the Process Control System, Multi-Agent
Systems and Embedded Software, all used in the construction of a monitoring and real-time alarm
solution. To perform this analysis, it is used as a case study alarm management system for
aluminum production halls, rooms of tanks, which will be built using these technologies.

Computer Science & Information Technology (CS & IT) 85
Is approached, therefore, the need for detection and alarm automation environments and control
in which controllers can not function correctly and precisely, due to mechanical problems in the
equipment that receive the command signals for actuation, as the emitted pulses the controllers. In
the case of large factory floor environments where each electrolytic tank requires their respective
actuators work in real time, the extent to which the actuators are triggered, the delay in
identifying problems of this nature can lead to loss of productivity and own electrolytic pot of
aluminum production, generating as a result of damage to the production process.
The contribution of this article is to develop an embedded management solution and alarm based
on the standard observe and react, using multi-agent systems for monitoring and conversion of
messages generated by the controllers, allowing the broadcast in real time, state and
environmental behavior, allowing adoption of preventive actions and operational in the shortest
time interval, by the vats operators working there. It consists in building a software solution
embedded in a raspberry PI plate for automation and control environment, based on the response
stimulus model, able to communicate with controllers electrolytic cells through the use of multi-
agent systems, allowing communication between centralizing agents and controllers for
transmitting / receiving text messages sent, generated by computer control, warning of each of the
potential problems identified in several electrolytic cells. These text messages are received by the
centralizing agent that converts audio messages transmitted to speakers of channels distributed
over the potrooms, allowing the technician and in charge of the process in it can identify a pot
with problem despite being in a distant location of the problem.
2. CONTROL SYSTEM POTROOM
In general, the process of aluminum production is based on the amount of alumina (Al2O3), the
raw material added to the electrolytic bath. In a bowl, electricity circulates between a positive
(cathode) and a negative pole (anode). That of the alumina mixing process, existing electrolytic
bath in the pot, and passing electric current anode cathode, it is apparent that liquid aluminum is
removed from the vessel for the production of ingots. Figure 2 below shows the generic
functional architecture of the aluminum production process, from the perspective of control
systems. This architecture is applied to the modeling of a system of pots, specifically for
monitoring purposes of their states and implementation of its controls.
Figure 2. Block diagram for distribution of alumina in vessel

The control system comprises sensor, controller and actuator. First, it is an indirect measurement
system, because from voltage and current signals and other information such as operation status
pot, to estimate quantities or rates those are used by the control device. The control system
processes the signals and these are used to making decisions that interfere with the production
process behavior. For example, control can promote actions to modify the frequency of operation
of Al2O3 feeder.
This control system is run on computer called process control computer, and has as main
objective to control the supply of alumina to the electrolytic cells, whether pre-baked type or
Soderberg point-feeder [2]. The intention is to keep a quantity of alumina dissolved in the bath to
ensure maximum production of aluminum. In pots of pre-baked type, the optimum concentration
of alumina in the bath is around 3% in the electrolytic bath. However, there are large variations in
this concentration, depending on the operations carried out in vats and variations of other control
parameters [3]. In practice, there is no method to directly measure this quantity or alumina
concentration in the electrolytic bath in the process of speed control requires [4]. Therefore, the
process computer controls the amount of alumina should meet the demand of alumina is added to
the vessel within the rated power range. The feed interval set in seconds, is the time window in
which the computer will send a pulse to drive the alumina feeders, wherein said range is from pot
to pot. It means that each production cell (electrolytic tank) has its power gap target.
When the control computer sends a signal to drive the feeders, and for some reason, such feeders
are not triggered, it means that some physical problem may be occurring. In this case, a physical
intervention such equipment is required. For this task, they are designated process technicians
who perform the repair and the reestablishment of the operation of the feeders and voltage
busbars. The problem becomes complex when technicians deal in a potroom environment with
large amounts of electrolytic cells arranged in large rooms and corridors in an aluminum
production plant.
3. MULTI-AGENT SYSTEMS
The definition of an agent can be expressed in many different ways. [5] defines an agent as a
computational entity software, located in a given environment, which has the perception that
environment via sensors has reasoning ability and acts autonomously in that environment through
actuators so to perform a specific function for which it was designed. [6], defined as something
that can be seen as having the perception of an environment through sensors and acting in this
environment through actuators. For [7] would be a piece of hardware or a computer system based
on software that enjoys some properties, such as learning autonomy, responsiveness, pro-activity
and social skills.
Figure 3 shows the interaction between two or more agents in a computer system and configures a
Multi-Agent System where these work together to perform certain tasks or perform a set of goals.
Scientific research and the practical implementation of Multi-Agent Systems is focused on
building standards, principles and models that allow the creation of small and large companies of
semi-autonomous agents, able to conveniently interact in order to achieve their objectives [8].

Figure 3. Structure of a Multi-Agent System
3.1. Jade
JADE tool (Java Agent Development Framework) began in 1998 through a CSELT (Telecom
Italia), whose development was motivated by the need for validation of specifications of FIPA
(Foundation for Intelligent, Physical Agents). AvJADE in 2000, became an open source platform,
being distributed by Telecom Italia on the LGPL (Library Gnu Public License) [9].
JADE provides certain features, such as distributed applications that exploit the use of software
agents. Being built on the Java platform, enables deployed agents using the JADE framework,
which can run on any operating system (OS), which makes it ideal for environments where you
can not get information about the OS used [10].
Other JADE’s features are the development of agents and control of them via a graphical
interface [9].
3.2. Auml
Unified Modeling Language Agent (Auml) is a standardized graphical modeling language by
FIPA (Foundation Intelligent Physical Agents). Auml was proposed as an extension of UML
(Unified Modeling Language), which uses decomposition, abstraction and organization to reduce
the complexity of software development, decomposing a system in small parts of objects, models,
use case or class, several operational actions. In relation to abstraction, it provides a specialized
abstract view of modeling (class, use case diagram of interface etc.) and is used to create a set of
semantics and operating service conditions and infrastructure [11].
The Auml offers structures as a class diagram and interface to show how agents can be erected in
an agent system. The model focuses on a point at a time, increasing the ability to understand the
issues to complex problems during the lifetime of the system design.

The main parts of Auml are the mechanisms for modeling multi-agent interaction protocols. This
is accomplished by introducing a new class diagrams UML diagrams protocol. These extend
diagrams and sequence diagrams of UML state, as shown in Figures 4 and 5.
Figure 4. AUML – Diagram Sequence
Figure 5. AUML – State Diagram
4. EMBEDDED SYSTEMS
The embedded systems are a combination of computer hardware and software, sometimes having
mechanical parts designed to perform a dedicated function, wherein such systems are often only a
small part of a larger system [12].
Typically, they are implemented from different technologies such as microprocessors,
microcontrollers, DSP, reconfigurable circuits, analog circuits and microwave and even
microelectromechanical systems (MEMS - Micro Electro Mechanical Systems) [13].
4.1. Raspberry PI
The Raspberry Pi model B is a small computer that uses the Broadcom BCM2835 multimedia
processor, the SoC type (system-on-chip), with 700 MHz 32-bit, built on the ARM11 architecture
[10]. The board is fed by a micro-USB port, with 5V voltage and requiring at least 700mA, with
waste of energy varying in 3.5 W [14].

The Raspberry Pi, as shown inFigure 6, has no hard disk on your hardware composition, using
only one memory card. For operating system installation it is necessary that the card has at least 2
GB capacity for all the necessary files. The operating system commonly used is the Raspibian, a
distribution recommended by Raspberry Foundation, based on Debian [15].
Figure 6. Raspberry B+ model
5. RELATED WORK
Binnicker [16] in 2015 presents a tool to detect and alert with temperature sensors for vehicles
occupied by passengers. The system alerts drivers if a child is left unattended in the vehicle in
order to prevent the child from being exposed to high or low temperatures inside the vehicle. To
this end, sensors are used to capture data via occupancy sensors, temperature and distance. The
data captured by the sensors determine the need to alert sent to the vehicle owner (driver) if an
occupant has been left inside the vehicle, still controlling some functions of the vehicle, allowing
foreign individuals to rescue him, with the unlocking of doors, windows opening, with possible
activation temperature of the automobile system.
M. Kumar et al. in [17] designed and developed an intelligent detection and intrusion alert system
designed to increase security, with real positive identification probability of attackers and
intruders, compared to other electronic security systems already used. Through multiple sensors,
the system evaluates the extent of danger exhibited by a person or animal within the confines of
the residence, conveying the owner of the residence, the various information about critical events
generated.
Ramya [18] presents a construction work involving containing a microcontroller system for
detecting and warning of toxic gases, particularly propane and LGP. Specific sensors for
identification of these two noxious gases are added to the microcontroller, which through the use
of analog / digital converter capture and analyze the information, which are also shown in the
display. If the gases exceed acceptable limits, the system immediately generates an alarm and also
send SMS alert message to authorized persons using GSM technology.
Gaspar et. al. [19] addresses the adoption of an alarm management philosophy for incident
management that avoids the generation of a huge amount of alarms in order to make life easier
for technicians who must manage in real time complex and dynamic environments and good
management alarms can provide crucial information to identify the cause of the fault and replace
the plant in normal operating state.
Morales [20] created and patented a monitoring and alarm system in a network audio link that
detects sound alarms triggered and inform the responsible authorities. The system and method

created based on the audio produced from alarms which are triggered. Implemented for PC type
computers or other premises of processors, different sound coming from different alarms,
classifies the sound as a specific alarm by FFT processing and distributes the alarm over the
Internet, public telephone switches, or other communication links to central station, which then
distributes the alarm to the authorities. In another embodiment, the alarm reports are generated
directly from the PC to the residence or specific company where authority is located.
6. PRELIMINARY RESULTS
As preliminary results, diagrams were built in Agent UML (Auml) for multi-agent embedded
proposed solution, which includes the deployment diagram, use cases, activities, block, and an
embedded prototype raspberry pi with jade platform running on environment simulation, where
the centralizing agent is embedded in the board and the controlling agents are available in
environments that simulate controllers. All diagrams were constructed using the free tool-day
standard 0.97.2 (http://live.gnome.org/Dia). The source code of the solution multi-agent was
developed in Java with Netbeans IDE version 8.1, where the environment is multi-agent on the
implementation platform JADE, in version 4.4.0.
6.1. Diagram Use Cases
Figure 7 shows the actors are placed in the context of the proposed approach and its use cases,
showing the set of paths that agents can run through the system, each with a discrete goal. The
actors representing the controlling system of tanks in practice can be represented by any
controller technology, this PLC (programmable logic controllers) to real-time vats control
applications that interact with Analog Stick / digital converters, analog and digital cards
Mustangs.
Figure 7. Diagram Use Cases of the proposed system
The use case Send audio alarm, converts the textual words received by dividing each of them into
audio files (eg. .wav) that are sent to their rooms vats according to the parameter that identifies
the number of pot with problem.
6.2. Transition State Diagram
The Transition State Diagram shown in Figure 8 represents the state and associated transitions to
the communication protocol between the developed agents and between the controller and agents,

which at the end of processing, transfer to the vat room audible alarm identifying a potential
problem in an electrolytic tank. Is represented in this case, the level of interaction between
players by exchanging messages and activation through behaviors.
Figure 8. State Transition Diagram of the Proposed System
Through this diagram, we see the main actors interacting in the sending and receiving of text
messages parameterized (Controller and Speech) that is validated, converted to audio and
broadcast.
6.3.Diagram Deployment
The deployment diagram shown in Figure 9 detailing the distribution of the proposed solution in
a generic environment potroom controlled by the controlling computer. It is observed that the
bundled solution raspberry pi plate is arranged in the process control network, in the same
environment controllers. The audio channel raspberry card is connected to speakers arranged in
the potroom. Therefore, the power pulses, ascent and descent bus, when they jammed by
mechanical failures in the vats can be signaled in the form of beep, in which the solution proposal
report to the operator exactly what existing physical problem.
6.4. Block Diagram of the Proposed Solution
Each controller agent is responsible for receiving the parameterized messages generated by the
control computer in which the agent is inserted. From the generation and reception of such text
messages, with the use of multi-agent platform JADE, the controlling agents send text messages
with the problem identified in a given tank to the centralizing agent, which queues the message
and is in charge of is converted it audio, transferring them to the speakers of the respective room
vats, where the pot with problem lies. The solution proposed is the block diagram shown in
Figure 10.

Figure 9. Deployment Diagram of the Proposed System
Figure 10. Solution Block Diagram proposal
Computer Science & Information Technology (CS & IT)
Figure 9. Deployment Diagram of the Proposed System
Figure 10. Solution Block Diagram proposal

7. IMPLEMENTATION
Alarms tanks with high voltage, much increase and / or bus down, swings, and its operating status
were detected by the controlling agents and sent to the centralizing agent of the respective
potroom, residing in raspberry pi. The centralizin
messages that are sent in FIFO order (first
warning system, which helps in faster diffusion of the critical situation found. The prototype of
the proposed system is shown in Figure 11.
Figure 11. Prototype of the Proposed System
8. CONCLUSIONS
An embedded system using multi
potrooms was implemented. For demonstration purposes they were
pot and controlling and centralizing agents responsible the generation and transmission of alarms
in audio format for the tanks room, respectively. The extent to which the controllers identify
some kind of problem in a particula
parameterized for the pot number identifier is generated and transmitted to the controller agent
forwards to the centralizing agent. This solution may be suitable for any shop floor environment
where there is a process automation for large equipment and controllable mechanisms that need to
maximize their production with minimal stop as a function of real problems identified the
equipment, enabling operators and technicians act to minimize downtime.
REFERENCES
[1] Fiona, Stevens McFaden & Bearne, Geoffrey (2001) Apliccation of advanced process control to
aluminium reduction cell - a review, TMS
[2] Braga, Carlos A. P., Martin, S., Nagem N. F, Ritter, C., Silva, Ari,
Anode Effect Kill. TMS Letters, v. 50, p. 417
[3] Silva, Ari (1995) Alumina feed control system of line 1 and line 2 potrooms at ALUMAR. ABAL
Associação Brasileira de Alumínio.
potroom, residing in raspberry pi. The centralizing agent identifies the queue textual alarm
messages that are sent in FIFO order (first-in, first-out) to the potroom in an audible format, as a
ystem is shown in Figure 11.
Figure 11. Prototype of the Proposed System
An embedded system using multi-agent platform for detection and alarm electrolytic cells in
potrooms was implemented. For demonstration purposes they were made simulation controllers
some kind of problem in a particular pot, textual message of the problem in question,
is a process automation for large equipment and controllable mechanisms that need to
equipment, enabling operators and technicians act to minimize downtime.
Fiona, Stevens McFaden & Bearne, Geoffrey (2001) Apliccation of advanced process control to
a review, TMS- The Mineral, Metal & Materials Society.
Braga, Carlos A. P., Martin, S., Nagem N. F, Ritter, C., Silva, Ari, Verlihay, Mark (2007) Faster
Anode Effect Kill. TMS Letters, v. 50, p. 417-422.
Silva, Ari (1995) Alumina feed control system of line 1 and line 2 potrooms at ALUMAR. ABAL
Associação Brasileira de Alumínio.
93
g agent identifies the queue textual alarm
out) to the potroom in an audible format, as a
agent platform for detection and alarm electrolytic cells in
made simulation controllers
r pot, textual message of the problem in question,
is a process automation for large equipment and controllable mechanisms that need to
Fiona, Stevens McFaden & Bearne, Geoffrey (2001) Apliccation of advanced process control to
Verlihay, Mark (2007) Faster
Silva, Ari (1995) Alumina feed control system of line 1 and line 2 potrooms at ALUMAR. ABAL –

[4] Silva, Ari, Carlos Braga, Eliezer Batista (2002) Smelter personnel technical training on aluminum
production. ALUMAR - Consórcio de Alumínio do Maranhão.
[5] Reis, Luis Paulo (2003) Coordination in Multi-Agent Systems: Applications in University
Management and Robotic Soccer. PhD, Electrical and Computers Engineering, FEUP, Porto.
[6] Russel, Stuart. "NORVIG, Peter." (2013) Inteligência Artificial. Editora Campus. Tradução da 3a
edição.
[7] Wooldridge, Michael, and Nicholas R. Jennings (1995) Intelligent agents: Theory and practice. The
knowledge engineering review 10.02: 115-152.
[8] Lesser, Victor (1999) Cooperative Multi-Agent Systems: A Personal View of the State of the Art,
IEEE Transactions on Knowledge and Data Engineering, Vol. 11, Nº 1.
[9] Telecom Italia - JADE - http://jade.tilab.com/
[10] Bellifemine, F., G. Caire, D. Greenwood, e I. NetLibrary (2007) Developing Multi-agent Systems
with JADE. Springer
[11] Gatti, M., Von Staa, A., Lucena, C. (2007) AUML-BP: A Basic Agent Oriented Software
Development Process Model Using AUML., PUC-Rio, n. 21/07, p. 25.
[12] William, Stallings (2010) Arquitetura e Organização de Computadores.
[13] Dias, Ailton Fernando (2001) "Concepção Conjunta Hardware/Software de Sistemas Embarcados de
Processamento de Imagens." CDTN/CNEN. Belo Horizonte, Brasil.
[14] Richardson, Matt, and Shawn Wallace (2013) "Primeiros passos com o Raspberry Pi." Primeira
Edição. Novatec Editora Ltda.
[15] Raspberry PI Foundation: http://www.raspberrypi.org/downloads/
[16] Binnicker, G. (2015) Temperature-Sensitive Vehicle Occupancy Detection and Alert System.
https://www.google.com/patents/US9428109
[17] S. Kumar, et al. (2015) "A Pattern Matching Model For Misuse Intrusion Detection,", pp. 1-11,
(found at http://www.cs.purdue.edu/coast/archive/data/categ24.html).
[18] Ramya V., Palaniappan B. (2012) Embedded system for Hazardous Gas detection and Alerting.
International Journal of Distributed and Parallel Systems (IJDPS) Vol.3, No.3.
[19] Gaspar, D., Ferreira, D., Coutinho, J., Correia, J., Albuquerque, P. (2010). A Política de Alarmes na
Gestão da Manutenção de uma Indústria Farmacêutica. Millenium, 39: 135‐151.
[20] Morales, F. (2001) “Network audio-link fire alarm monitoring system and method”
https://www.google.com/patents/US6215404
[21] Braga, Carlos A., Nagem,N.F, Fonseca, J. V. (2007) “Bandwidth adjustment of digital kalman filter
for indirect measurement of electrolytic bath state variables”. 2nd International Conference on
Sensing Technology.

AUTHORS
Luis Carlos Fonseca graduated in Technology in the Data Processing of Maranhão
University (2000), Master in Engineering Electricity from Federal University of Maranhão
(2003) and PhD in Computer Science in Education from the Federal University of Rio
Grande do Sul (2009). He is currently assistant professor at the State University of
Maranhão. He has experience in computer science with an emphasis in Information
Systems, mainly in the following areas: Artificial Intelligence, Information Technology in
Education, Software Engineering, Information Retrieval, Recommender Systems.
Alan Robert da Silva Ribeiro graduated in Bachelor of Computer Science from Federal
University of Maranhão (1999), specialization in Systems Analysis and Design from the
Federal University of Maranhão (2006), specialized in Advanced Systems for Internet and
Intranet at University Center Maranhão (2007) and is Masters Student of Computer
Engineering and Systems at the State University of Maranhão. It is currently a Ministerial
Analyst and Coordinator of the Public Ministry of the Information Technology area of the
state of Maranhão, having worked as a Senior Process Control Systems Analyst at
multinational ALCOA, which was part of SDG global group (Service Delivery Group), responsible for
development and installation of process control systems in the company's factories in the world. He has
experience in the area of Computer Science, Aluminum Process Control, Information Technology
Management, acting on the following topics: Electrolytic Control Computer,
Information Technology Governance, Database, Ubiquitous Computing and Driven Engineering by
models.
John Selmo de Souza do Nascimento
emphasis on Automation and Process Control at the
Currently is Masters Student of Computer and Systems Engineering from State University
of Maranhão. He has experience in Geosciences with an emphasis on Georeferenced
systems, robotics, mechatronics and automation.
graduated in Technology in the Data Processing of Maranhão
University (2000), Master in Engineering Electricity from Federal University of Maranhão
(2003) and PhD in Computer Science in Education from the Federal University of Rio
is currently assistant professor at the State University of
Maranhão. He has experience in computer science with an emphasis in Information
Systems, mainly in the following areas: Artificial Intelligence, Information Technology in
neering, Information Retrieval, Recommender Systems.
graduated in Bachelor of Computer Science from Federal
University of Maranhão (1999), specialization in Systems Analysis and Design from the
(2006), specialized in Advanced Systems for Internet and
Intranet at University Center Maranhão (2007) and is Masters Student of Computer
Engineering and Systems at the State University of Maranhão. It is currently a Ministerial
the Public Ministry of the Information Technology area of the
state of Maranhão, having worked as a Senior Process Control Systems Analyst at
installation of process control systems in the company's factories in the world. He has
Management, acting on the following topics: Electrolytic Control Computer, Artificial Intelligence,
John Selmo de Souza do Nascimento holds a degree in Computer Engineering with
emphasis on Automation and Process Control at the State University of Maranhão (2015).
Currently is Masters Student of Computer and Systems Engineering from State University
of Maranhão. He has experience in Geosciences with an emphasis on Georeferenced
systems, robotics, mechatronics and automation.
95
installation of process control systems in the company's factories in the world. He has
Artificial Intelligence,

APPROACH MULTI-AGENTS EMBEDDED ALARM IN POTROOMS

More Related Content

What's hot (20)

Similar to APPROACH MULTI-AGENTS EMBEDDED ALARM IN POTROOMS (10)

Recently uploaded (20)

APPROACH MULTI-AGENTS EMBEDDED ALARM IN POTROOMS