final report LACCIR Project pons-luna-szasz

final report LACCIR Project pons-luna-szasz

PROJECT TITLE: “LOW COST COMPUTER BASED SYSTEM FOR
QUALITY EVALUATION AND PRESERVATION OF GRAINS STORED IN
POLYMER BAGS”
FINAL REPORT
PRINCIPAL RESEARCHERS:
Claudia Pons (UNLP, Argentina), Carlos Luna (UR, Uruguay), Nora Szasz (ORT, Uruguay)
1. Abstract
Silobag is an ad-hoc low-cost storage technique that consists in keeping the grains into hermetic
polyethylene bags which are stored in the same field of crop (on-farm). It is likely that several
million tons of grain will be stored in harvest bags within the next years, in Latin American
countries and also in other parts of the world. Research confirms that although silobags have some
limitations, they offer to growers a relatively cheap and reliable grain storage solution. Therefore,
it is important to work towards the improvement of such technology; in particular, the
incorporation of computer technology would be a valuable asset. In this report we describe the
results of our project aimed at developing a low-cost wireless software system for monitoring and
adapting the internal conditions of the grain stored into silobags in order to improve their
preservation state. This system is addressed to small and medium growers and it was developed
using modern techniques of software engineering.
Project website: http://www.lifia.info.unlp.edu.ar/laccir/
2. Deliverables
To reach our goals we have developed the following deliverables:
low-cost sensors to measure the humidity and temperature of the grain.
mathematical models of the behavior of the humidity and temperature levels into the
silobags.
a software control system for monitoring and adapting the state of the silobags.
a software development process fitting this kind of systems in order to favor productivity,
portability, reuse and maintenance.
Due to the fact that the results of this project are addressed to small and medium growers,
to develop a low-cost solution was a mandatory requirement of our work.
We were not able to build the aeration systems being suitable to be incorporated into the bag to
modify the inner conditions at the different levels, as we proposed. This limitation was caused by
the characteristics of the silobag that turn very complex the design of such aeration systems.
The following deliverables can be downloaded from www.lifia.info.unlp.edu.ar/laccir/
•
SilobagConfigurator-src-v1.0 (download) Silobag Configurator Source (DSL Tools
•
SilobagConfigurator-v1.0 (download) Silobag Configurator Default Editor (Visual
Project)
Studio Extension)
•
SilobagConfigurator-app-v1.0 (download) Example User Application (Standalone
•
Silobag Configurator Installation Manual (download) (in spanish only)
•
Silobag Configurator User Manual (download) (in spanish only)
App)
3.
Evaluation Methods
We run a number of test suites to validate the mathematical models and the control system. The
system was deployed in 2 farms in Argentina, with the collaboration of INTA. We plan further
experiments. The improvements in the quality of grains will be studied in these cases. Following
the results we will adjust details as required.
4.
Data Collected
Currently we have accomplished the construction of a set off low-cost sensors to measure the
humidity and temperature of the grain stored in silobags. We have applied those sensors to collect
data for 8 moths, using real silobags deployed in the field. This data was used to create
mathematical models of the behavior of the humidity and temperature levels into the silobags.
Results were presented in [1].
5.
Conclusions
Currently we have proven the following hypotheses defined for this project,
Hypothesis: It is possible to develop low-cost testing methods for harvest bags that allow users to
measure and supervise humidity and temperature levels of the grain that is stored into harvest
bags.
Hypothesis: The humidity level into the bag will reach the optimal values as a consequence of the
layered homogenization process. The proposed optimization procedure will favor the preservation
quality of grains for long periods of time.
Hypothesis: Modern techniques of Software Engineering can be applied to the development of
intelligent control systems in order to favor their productivity, portability and maintainability.
In this project we have developed an infrastructure for the construction of wireless monitoring
systems for silobags. In particular, we have defined a Domain Specific Language (DSL) which allows
us to specify the specific configuration of the monitoring system. Then, the executable code is
derived automatically from the DSL .
Given the independence between the models and the code, the DSM methodology smoothly
supports the evolution of the underlying architecture (execution platform and domain framework)
on which you deploy the system, reusing the generated models in a large part. In this way, the
specified software is timeless in technological terms and the high-level representations that define
the systems remain valid as long as the problem domain remains the same. The proximity
between the model and the domain also has the advantage that the models work as an artifact for
both capture of requirements and specification of software. As a result, customers have greater
participation in the system development.
Because the source code is produced by generators automatically, that code is error-free and
efficient in the use of execution and memory resources, so it does not need be modified or
optimized. In addition, all the generated code has a unique style of programming and design. The
unfavorable aspect of this proposal is that it is necessary to invest extra work in the development
of the components of the architecture (meta-models, editors, generators, domain framework,
etc.). This initial effort can be amortized considerably due to the increase in productivity in the
development. This increases proportionally to the amount of "repetition" of software
developments carried out within the same domain.
The DSL that we implemented is sufficiently extensible to incorporate different types of sensors,
actuators and measures to be taken in the future. In addition, the final system has been designed
to incorporate prediction of the behavior of the internal conditions of the silos bag, generation of
alarms to certain risky behavior and implementation of specific actuators from the configured
alarms.
The medium-term objective is to have an intelligent low-cost system for monitoring and automatic
adjustment of the internal conditions of the grain stored in silos bag. In pursuit of this goal, the
incorporation of some type of actuator and the inclusion of the mathematical model for the
prediction of the behavior of the humidity and temperature, developed for this project is planned.
We believe that given the importance and dissemination that silos bag currently have in countries
of the region, this constitutes a significant contribution which, through the incorporation of
technology, will make it possible to improve the conditions of storage of grains in silo bag, identify
risky conditions of collection, take action to improve critical factors and as a result, extend the
time for the conservation of the production, without degrading the quality of the same.
This system presents an improvement on the way farmers take data today, facilitating and
expediting the collection of measurements. By placing sensors fixed inside the silos that transmit
wirelessly eliminates the need of having personnel dedicated exclusively to this task, which is a
strong economic advantage. It is possible to obtain measurements with a frequency greater than
the current without the need for unnecessary perforations that alter the conditions of the silobag.
Regarding the scalability of the system, one of the advantages of this design with respect to
current options is that it allows measurements in a field with more than a silo bag. It is
theoretically possible to deploy 65536 sensors that can work together to send data when they are
required to the receiving device. It is also possible to send data of temperature and humidity to
the Internet via a server dedicated to this task; this represents one of the most interesting
improvements to deal with in the near future. Nowadays Internet is a mass media and is
increasingly in a greater number of devices; the possibility of having this information in the web
will allow producer to carry out continuous monitoring of the status of the stored grain.
List of Publications
6.
“Un Lenguaje Específico de Dominio para la Generación Automática de Código en Sistemas
de Monitoreo Inalámbrico de Silos Bolsa”. Pons, Luna, Calegari, Szasz. Ninth LACCEI Latin
American and Caribbean Conference (LACCEI’2011), Engineering for a Smart Planet,
Innovation, Information Technology and Computational Tools for Sustainable
Development, August 3-5, 2011, Medellín, Colombia. (2011)
"Ingeniería Dirigida por Modelos Aplicada al Control Automático del Almacenamiento en
Silos Bolsa”. Daniel Calegari, Carlos Luna, Mauro Canabé, Federico Sierra, Nora Szasz,
Claudia Pons. Congreso Argentino de Agro Informática JAIIO/CAI 2010. Buenos Aires.
Septiembre 2010. ISSN 1666 1141 (2010)
"Sistema de Medición Inalámbrico de Temperatura y Humedad para silo-Bolsa”.A. Fueyo,
J. Epeloa. Congreso Argentino de Agro Informática JAIIO/CAI 2010. Buenos Aires.
Septiembre 2010. ISSN 1666 1141. Buenos Aires, Argentina, Agosto 2010.
“Características Higroscópicas de Forrajes Almacenados en Silo Bolsas”, Juan Pons;
Josefina Marinissen; Claudia Pons; Sebastian Orionte. Congreso Argentino de Agro
Informática. JAIIO/CAI 2009. Mar del Plata. Argentina. Agosto 2009. ISSN 1666 1141.
“Model transformation languages relying on models as ADTs”. Jerónimo Irazábal and
Claudia Pons. Software Language Engineering Conference (SLE 2009). Denver USA. Lecture
Notes in Computer Science. Springer Verlag.
“Model transformation as a mechanism for the implementation of domain specific
transformation languages”. Jerónimo Irazábal and Claudia Pons and Carlos Neil.
Electronic Journal of SADIO. 2010.
“Transformación de modelos como mecanismo de implementación de DSLs”. Jerónimo
Irazabal y Claudia Pons. Argentinean Symposium on Software Engineering. JAIIO/ASSE
2009. Mar del Plata. Argentina. Agosto 2009. ISSN 1666 1141
“Experiment with a Type-Theoretic Approach to the Verification of Model
Transformations”. D. Calegari, C. Luna, N. Szasz, A. Tasistro II Chilean Workshop on Formal
Methods (ChWFM), Chile, November 2009.
“Specification of Products and Product Lines”, C. Luna and A. Gonzalez. 9th
International Workshop on Reduction Strategies in Rewriting and Programming, Brasil,
June 2009. EPTCS.
A Two-level Calculus for Composing Hybrid QVT Transformations. C. Pons, R. Giandini, G.
Pérez, G. Baum. II International Workshop on Advanced Software Engineering (IWASE).
Chile, November 2009.
7.
List of Presentations
2do Congreso Argentino de AgroInformática (CAI). Talk title: "Ingeniería Dirigida por
Modelos Aplicada al Control Automático del Almacenamiento en Silos Bolsa”. Speaker:
C.Pons. Universidad Argentina de la Empresa (UADE), Buenos Aires, Argentina, Agosto
2010.
2do Congreso Argentino de AgroInformática (CAI). Talk title: “Sistema de Medición
Inalámbrico de Temperatura y Humedad para silo-Bolsa”. Speaker: Javier Epeloa.
Universidad Argentina de la Empresa (UADE), Buenos Aires, Argentina, Agosto 2010.
Microsoft Research Faculty Summit 2009. May 15, 2009. Sofitel Hotel - La Reserva
Cardales. Province of Buenos Aires, Argentina. Talk title: “Low Cost Computer Based
System for Quality Evaluation and Preservation of Grains Stored in Polymer Bags".
Speaker: C.Pons.
7mo Congreso Internacional de Innovación Tecnológica Informática, CIITI. Capitulo
Rosario. Santa Fe. Argentina. Noviembre 2009. Talk title: “Tendencias de investigación en
TICs”. Speaker: C.Pons.
Software Language Engineering International Conference (SLE 2009). Denver, USA.
October 2009. Talk title: “Model transformation languages relying on models as ADTs”.
Speaker: Jerónimo Irazábal.
8.
Teaching´s Impact
In the context of this project we successfully completed the following thesis:
Title:
“Especificación
y
Verificación
de
Transformaciones
de
Modelos”.
Students: F. Varesi, H. López, and M Viñolo. Thesis for obtaining the degree of Ingeniero en
Computación. Facultad de Ingeniería, Universidad de la República, Uruguay. Concluded in July
2010.
Title: “Aplicando MDD al desarrollo de sistemas agropecuarios: modelado de sistemas de
control de calidad de granos almacenados en silobolsas”.
Students: Miguel Martínez and Sebastián Lavie. Thesis for obtaining the degree of Licenciado
en Informática. Facultad de Informática, UNLP. Concluded in June 2010.
Title: “Sensores inalámbricos de temperatura y humedad para silobolsas”.
Students: Alexis Fueyo and Javier Epeloa. Thesis for obtaining the degree of Electronic
Engineer. Facultad de Ingeniería, UNLP. Concluded in August 2010.
9.
Research and Development Team
Universidad Nacional de La Plata, Argentina: Claudia Pons, Juan Pons, Gabriel Baum, Paul
Puleston, Jerónimo Irazabal, Javier Epeloa, Alexis Fueyo.
Universidad de la República. Uruguay: Carlos Luna, Daniel Calegari
Universidad ORT. Uruguay: Nora Szasz
INTA. Argentina: Josefina Marinissen
Universidad Abierta Interamericana (UAI): Carlos Neil