abstracts - IASRI

Theme 3: Agricultural Informatics
T3.1I
Pansy: A Portable Autonomous Irrigation System
Benjamin Beckmann and Ajay Gupta
Wireless Sensor Network (WiSe) Laboratory, Department of Computer Science,
Western Michigan University, Kalamazoo, Michigan, USA
Temperature, humidity, chemical composition and soil moisture are some of the factors that
need to be carefully monitored to insure the longevity and health of the crop. Over thousands
of year’s humans, farmers have learned that this monitoring needs to be done frequently and
is labor intensive, time consuming and expensive when many crops are simultaneously being
grown. We introduce a design of a preliminary system using wireless networked sensors,
which makes monitoring and irrigating the crop efficient and reduces, if not, eliminates
human interaction with the crop during its growth period.
Intelligent wireless sensors can be deployed in a greenhouse environment and effectively
monitor the environment and irrigate crops when needed for varying durations. This
automated irrigation is ideal for a greenhouse application since slight changes in the climate
within the greenhouse can cause irreversible damage to sensitive crops. By introducing smart
wireless sensors into a greenhouse environment the growth of the plants can be controlled
dynamically with very little human intervention. The intelligent wireless sensors produce an
environment where every plant in the greenhouse can be monitored at a rate hundreds of time
higher than a grower could possibly do alone.
The system introduced in our work uses a minimal amount of environmental factors to
monitor and sustain crops. It reduces human contact, water and chemical usage, and labor
costs, while producing crops with solid root structures, healthy foliage, and high bloom
counts. Once the wireless sensor devices are in place they self-organize and dynamically
control the local climate according the crop species. Positive side effects of such a system are
highly accurate inventories, species location, and elimination of barcodes on pots.
We will present the design of architecture for a portable autonomous irrigation system,
named PANSY, using intelligent wireless sensor devices. Experimental results are discussed,
which show PANSY’s potential and current ability to produce crops equivalent to those
produced by a professional grower. We also describe how it can be deployed in a greenhouse
to achieve this goal. Deployment of PANSY-like systems is also beneficial in other situations
where water is scarce or chemical fertilizers need be carefully controlled.
T3.2I
Hierarchical Censored Production Rules (HCPRs) System – A Frame Work for
Intelligent Systems
K.K. Bharadwaj
School of Computer & Systems Sciences, Jawaharlal Nehru University, New Delhi

In the real world, both humans and computers often have to reason using insufficient,
incomplete, or tentative premises. Moreover both are subject to constraints of time and
memory. Variable Precision Logic (VPL) (Michalski and Winston, MIT, 1986) is concerned
with problem of reasoning with incomplete information, subject to resource constraints and
the problem of reasoning efficiently with exceptions. VPL offers mechanism for handling
trade-off between the precision of inferences and computational efficiency of deriving them.
Specificity and certainty are the two aspects of precision. Certainty refers to the degree of
belief in a statement, whereas specificity refers to the degree of detail of a description.
As an extension of standard Production rule, Michalski and Winston proposed the Censored
Production Rule (CPR) of the form:
IF THEN UNLESS,
as an underlying representation and computational mechanism to enable logic based systems
to exhibit variable precision in which certainty varies while specificity stays constant. To
address various problems and shortcomings with CPRs system, Bharadwaj & Jain (1992)
introduced a concept of Hierarchical Censored Production Rule (HCPR) of the form:
Decision IF
UNLESS
GENERALITY
SPECIFICITY ,
which can be made to exhibit variable precision in the reasoning such that both certainty of
belief in a conclusion and its specificity may be controlled by the reasoning process. Several
extensions/generalizations of the system have been proposed and recently efforts are being
made towards knowledge discovery and knowledge management - automated
discovery/management of knowledge in the form of HCPRs (see the references). The HCPRbased
system would be able to adjust the certainty/specificity of inferences to conform to
time constraints. Such a system has numerous applications in situations where decisions must
me made in real time and with uncertain information-examples range from medical expert
systems for operating rooms to domestic robots.
T3.3I
Digital Networks for Farmers: A Step towards Faster and More Inclusive Growth in
Rural India during 2007 – 2012
Madaswamy Moni
National Informatics Centre, New Delhi
There have been concerns about persistent rural poverty, unemployment and inequality, and
resulting social tensions at grassroots level in India. The strategies and policies developed by
the Planners and Policymakers, adopted two approaches: one focusing on the overall
economic development (through percolation, trickle down and spread effect), and the other
poverty alleviation (direct intervention). Though these two approaches reinforce each other,
there has been no effect to integrate them. At the beginning of the new millennium, 260
million people in the country did not have incomes to access a consumption basket, which

defines the poverty line. Of these, 75 per cent were in the rural areas. Agricultural wage
earners, small and marginal farmers, casual workers engaged in non-agricultural activities,
rural women (especially women-headed households), among the others, constitute the bulk of
the rural poor. The growing populations need food, clothing, shelter, fuel and fodder for their
livestock.
As market oriented economic development proceeds, Indian farmers in rural areas continue
to experience great disparity in income compared with other sectors. They revert to natural
resources as the most accessible sources of livelihood. Degradation of natural resource is a
key threat to socio-economic development, and to global environment (e.g., climate change
and loss of biodiversity). After decades of limited success in eliminating rural poverty, new
ideas about rural development (i.e. viable society in a rural space, livelihood approach,
sustainable livelihood approach, sustainable agricultural and rural livelihood approach,
sustainable community concept, Multiple Livelihood Opportunities, etc) are emerging so as
to reduce the vulnerability of the rural poor. Poverty Alleviation, Livelihood Opportunities
and Gainful Employment are closely linked.
Science & Technology offers tremendous opportunities in simultaneous achievements the
goal of sustainable agriculture and improving rural livelihoods. “Doubly Green Revolution”
adheres to “growth in agricultural production” and “improved livelihoods” now. Digital
opportunities have emerged as “power tool” for fostering agricultural growth, poverty
reduction and sustainable resources use in developing countries. Rural India requires to be
given a chance through DNF for moving towards “faster and more inclusive growth”
during the 11 th plan period, i.e. 2007-12.This paper deals with the Indian Government
initiatives and progress on digital networks for farmer (DNF) and digital inclusion for
fostering rural prosperity and reducing spatial disparities in India.
T3.4I
Information Kiosks in Rural India:
An Opportunity for Agricultural Research to Reach the Un-reached
V.P. Sharma and K.V. Rao
National Institute of Agricultural Extension Management (MANAGE), Hyderabad
ICTs have started to make their presence felt in Rural India. The farmers and farm-families
are browsing the net and getting general, technical and marketing information from the
Information kiosks in over 26000 villages across the country. If the rural India can be
connected and the “masses” are empowered with “Information”, the Indian Economy will
take a leap forward into the Digital Millenium with a great speed. The process has already
started. The focus on “e-Governance” and “IT for Masses” is also emerging as front runner in
all these State vision documents. Technologies specially suitable for Rural Areas are being
developed and deployed.Portals on Rural Markets and Agricultural Services are being
hosted. District level Web Sites are being hosted, Information Kiosks are being established at
block/ Mandal and village levels and the technical and other need-based information is being
collected, digitized and hosted on the Internet. Private Sector has taken lead in providing
comprehensive “Extension” and “Marketing” content services to the Farmers. The “echaupals”
of ITC and “Agriwatch.com” web-portal are the trend-setter scaled-up across the
country.

The ICTs are thus supplementing the efforts of Agricultural Extension system (both in public
sector as well as in private sector) to become a more dynamic instrument of continuous 2-way
dialogue with the Farmers, on various issues. There are broadly 4 types of ICT enabled
interventions for improving rural communication system, with focus on delivery of
Agricultural Information, in India. These are:
‣ National Government initiated Projects;
‣ State Government promoted Projects;
‣ Private Corporate Sector promoted Projects; and
‣ Non-Government Sector (not-for Profit) sector promoted projects.
The authors (along with his team two support Faculty members) have undertaken a research
study on assessing the “Imapact of Information Kiosks on Farmers”, during June – September
2005. In depth study was undertaken for 60 Information Kiosks in three states of the country,
besides personally interviewing 270 farmers. The paper discusses in brief the basic
objectives, funding pattern, outreach, content (quality, timeliness and completeness),
community orientation of the projects, capacity building, linkages with other public and
privte institutions for information support and sustainability issues.
T3.5I
Assembling the Maize Genome on Large-Scale Parallel Computers
Srinivas Alluru
Bioinformatics and Computational Biology Program, Iowa State University USA
Maize genome is currently being sequenced by a consortium of Washington University,
University of Arizona, Iowa State University, and Cold Spring Harbor Laboratories. Repeats
span 65-80% of the maize genome, which creates particular challenges in its sequencing and
assembly. In this talk, I will describe parallel bioinformatics methods we developed for this
purpose, especially for assembling sequences resulting from gene-enrichment sequencing
strategies. I will present our assembly efforts on the IBM BlueGene/ L massively parallel
supercomputer, and discuss current progress on the NSF/USDA/DOE interagency project for
maize genome sequencing.
T3.6I
Rural Knowledge Centers as Facilitators of New Learning Opportunities for the
Rural Families: a case study
Dileep Kumar Guntuku 1 , Sreenath Dixit 2 and Balaji Venkataraman 3
1
Department of ICT for Agriculture, DA-IICT, Gandhi Nagar
2 Division of Resource Management, CRIDA, Hyderabad
3 Knowledge Management and Sharing Group, ICRISAT
There are various names and descriptions for rural knowledge centers, but their main role is
facilitate the access to managed or direct information services for rural families. India is
home to a very large number of pilot projects in this sector over the last about ten years. The
number of internet-connected rural knowledge centers varies but is believed to be close to
10000 as of mid-2006. The recent decision of the Government of India to establish about
1,00,000 Common Service Centers has generated altogether new possibilities and

opportunities to design and make available a wide range of information services for the rural
poor in India.
Earlier studies indicated that the access to government information, and information on
education were the two priority information demands. A number of projects include provision
of governance-related information as a key service. Education mostly relating to IT literacy is
offered on many projects although systematic efforts to measure the importance and impact
of such information are not yet available.
One of the significant challenges that India faces is the recurrence of disasters that lead to
large scale disruption of economic and development activities and causes considerable
distress among the victims. Over a period of time, relief measures have become affordable in
many situations but the cumulative losses are staggering. The Disaster Management
Authority of India has identified drought and earthquakes as among the phenomena that can
cause deep and lasting distress among the victims while generating massive economic losses
to the system as a whole. There is a need to identify new systems that combine early warning
arrangements with access for appropriate support services.
Over the last two years, ICRISAT has made an effort in partnership with CRIDA to study the
possible use of rural knowledge centers in enhancing drought preparedness at the microregional
level among the rural families. The pilot is premised on the assumption that a
country such as India has reasonable arrangements for early warning communication in a topdown
manner. International experience shows that such top-down flow of important
information must combine with a bottom-up process for its rapid and effective use by the
intended recipients, namely the rural families. The rural knowledge centers, operating in an
interactive hub-and-spokes model for local value-addition and dissemination and capture,
could provide the right interfaces to generate such blended communication.
ICRISAT and partners started out with a detailed information needs assessment process
covering all the villages in the Addakal Mandal, Mahbub Nagar District of AP State. The
information needs assessment revealed that there were significant gaps in the availability of
and access to natural resource management information by the rural families. There was
striking level of lack of awareness about the general extent and consequences of drought, and
most families had assumed that out-migration was the only easy option. Thus, as part of the
drought preparedness information system, opportunities for learning about NRM issues at a
basic level needed to be available.
ICRISAT’s Virtual Academy for the Semi Arid Tropics (www.vasat.org) attempts to
combine the advantages and power of contemporary process in technology-mediated open
and distance learning to provide such education-based communication arrangement in
Addakal Mandal. The principal local partner is a community-based NOGO called the
Adarsha Mahila Samaikhya, which is a federation of village-level micro-credit groups in the
mandal. The coverage extends to all the villages while the federation has a membership of
about 6000 individuals , all women (as of September 2006). The AMS has provided good
quality space and furniture and electricity for the rural knowledge centers in three villages
while its own premises act as the local hub that has access to the Internet. The village access
centers act as channels for two-way communication between the hub and the rural families.
The hub itself is supported by an online content management system, modification of a
standard learning management system hosted by ICRISAT. The CMS allows questions from
the rural families to be logged for viewing by ICRISAT and CRIDA-based experts.

Over a period of time, we found that it took an average of seven days for a farmer to receive a
response that he considered satisfactory. This was entirely due to the way the questions were
phrased and experts often sought more information prior to developing a solution. This
resulted in cycles of information flow prior to a solution being delivered. ICRISAT scholars
conducted a study with the AMS women volunteers and helped them learn the essentials of
pest management using technically simple, literacy-imparting modules. The learning process
took place over a period of 18 hours spread over almost a week. The process facilitator used
CDROM-based modules. The lessons learnt were evident in the way the quality of questions
changed. More information that describes the context was added to every question by the
AMS volunteer, and this resulted in dramatic reduction in the time taken to deliver a useful
solution. The average time now is about one day, often solutions being delivered within the
same day.
This experience has led us to design a larger learning process that involves rural knowledge
center facilitators in many other parts of India, in collaboration with the M S Swaminathan
Research Foundation through their Virtual Resources Centers project. In the initial trials, a
two-way video-conferencing facility was used, and we found that the quality of both expertfarmer
and farmer-farmer interactions was appreciable. In the second set of trials, we have
started online learning process for the Addakal volunteers on groundnut cultivation with
video-conferencing as an important tool. The process is continuing and the initial results
show that with sufficient literacy imparted, the rural knowledge center volunteers of the AMS
are able to decide on the aspects of learning modules that should be localized.
Thus, it emerges that rural knowledge centers can facilitate local learning provided adequate
efforts are made to generate such material. Learning capacity even outside a classroom milieu
is considerable if flexible methods are followed and learning is delivered as granules.
Technology-mediated learning is thus a possibility for rural families in their quest to find
more sustainable resource management options.
We are grateful to the ICT-KM program of the CGIAR for partial financial support.
T3.7I
Geoinformatics Applications in Agriculture
Anil Kumar Singh
Water Technology Centre, Indian Agricultural Research Institute, New Delhi
The twenty first century is the era of globalization where the rate of development and
industrialization are changing the atmosphere and surface of earth like never before. The
compulsion to produce more compounded by poorly planned developmental activities and the
rapid change in consumption patterns is not only resulting in environmental degradation but
also leading to non-sustainable agricultural production. The need to keep up with the
demands of the ever increasing population is necessitating development of technologies
which can retrieve, process and disseminate information in an effective and reliable manner
at the fastest pace possible. The phenomenal development of information and communication
technologies (ICT) is, therefore, going to play a very vital role in all developmental activities
and the agricultural sector will be no exception. Geoinformatics is a rapidly developing
science which encompasses all modern tools and techniques like remote sensing, GIS, GPS,

photogrammetry along with computer modeling to systematic data storage, analysis and
retrieval for research as well as documentation and enables extraction and analysis of relevant
information in the shortest time possible. It has virtually a limitless potential in its
applications in agriculture which can extend from assessing the state of natural resources,
monitoring the impact of human, industrial and other development activities on agricultural
production base and also monitor and evaluate the impact of not only anthropogenic activities
but also the remedial measures taken. The diverse applications range from coupling the
ground based GPS link information through remote sensing, inventorization, precision
farming, water-logging and soil salinity estimation, resource based development plans,
information systems, real time decision support systems etc. This paper presents an overview
of the vast scope and potential applications of geoinformatics in agriculture and the need to
adopt them for sustainable agricultural development.
T3.8I
Sensors, Systems and Networks for Precision Agriculture
Pawan Kapur
Central Scientific instruments Organisation,Sector 30, Chandigarh
Agro-based sector has witnessed phenomenal change in the recent past in almost every area
of activity. Over last few decades, the impact of science and technology on industrialization
and ecosystem has intensified the deterioration of the ecosystem, leading to depletion of
natural resources. The agriculture of the forties, which was eco-friendly, has now become
fully chemicalized with new farming technologies for effective commercialization of its
products. The pressure is further increasing due to population growth putting new challenges
especially for Indian agro-based sector to meet the domestic demands and also open up
avenues for export. Regulatory requirements and stiff quality parameters call for the
development of advanced measurement techniques, new sensors & devices and associated
instrumentation solutions for sustainable and equitable utilization of natural resources.
Sensors and associated systems for detection of key process variables in agro based industries
are vital in optimizing the performance in terms of productivity and quality. The applications
range from pre-harvest parameters such as soil nutrients monitoring, tissue culture and
subsequent crop health monitoring, water and micro- nutrient control; post-harvest operations
which call for optimum setting of process conditions for individual unit process control,
storage automation and mandi-level requirements in terms of quantification of quality for
maximum returns. The various technologies are being explored for sensor development
which includes optics, photonics, NIR spectroscopy, fluorescence, SAW, electronics,
electrochemical and new materials (conductive polymers).
Communication technologies, especially wireless and networking, have witnessed
phenomenal development in recent past. Types of wireless technologies being developed
range from simple short-range Infrared communications (IrDA), point-to-point
communications, to wireless personal area network (WPAN), point-to multi-point
communications, such as Bluetooth, to mid-range, multi-hop wireless local area network
(WLAN), to long-distance cellular phone systems, such as GSM/GPRS and CDMA.
Similarly, Ethernet to wireless Ad-Hoc networks and Mainframes to Ubiquitous systems,
networking technology has also progressed manifolds. Developments in these technologies
have been attributed to the growing mobile phone market, but a far greater potential exists for

other application areas, especially wireless sensors and sensor networks. A wireless sensor
network is a system comprised of radio frequency (RF) transceivers, sensors, microcomputers
and power sources. Wireless sensor networks with self-organizing, self-configuring, selfdiagnosing
and self-healing capabilities have been developed to solve problems or to enable
applications that traditional technologies could not address. Once available, these
technologies would allow us to find many new applications that could not have been
considered possible before. Wireless sensor technology is still at its early development stage.
Applications of wireless sensors in agriculture and food industry are still rare.
CSIO’s programme in the area of agri and food instrumentation deals with the complete food
chain from harvest of raw materials till the ultimate utilization of foods and food products by
consumers. It emphasizes on the instrumentation for utilization of efficient processes for
preparing and preserving raw agricultural materials, changing, shaping, and processing
agricultural materials into high quality food material, pharmaceutical materials and industrial
materials, extracting and purifying agricultural materials by using techniques which take into
consideration initial material properties, material properties during processing and final
product properties to ensure maximum production rate with highest quality for purpose of
consumption and manufacturing, This paper intends to give an overview of available sensors
and systems that may be applicable to agriculture and food industry, including development
trends, challenges to be faced and applications of sensor networks in sustainable precision
agriculture.
T3.9I
Prediction of Forest Cover Using Decision Trees
B. Chandra 1 and Pallath Paul V. 2
1 Department of Mathematics, Indian Institute of Technology- Delhi, New Delhi
2 Department of Computer Science, Institute of Technology, Nirma University. Ahmedabad
Information regarding forest land is highly required for developing ecosystem management
strategies which will facilitate the decision-making process. It is often difficult to get the
relevant data for forest lands that are outside the immediate jurisdiction of the concerned
authorities. One of the approaches for obtaining this information is through the use of
predictive models like Decision Trees and Neural Networks. Blackard et al. (2000) have
shown that Neural network approach outperforms the traditional Discriminant analysis
method in predicting forest cover types. The accuracy achieved by Neural Network was
70.58%. Decision trees algorithms have been proposed in the past for classification of
numeric as well as categorical attributes. SLIQ algorithm was proposed (Mehta et al, 1996) as
an improvement over ID3 and C4.5 algorithms (Quinlan, 1993). Robust Algorithm for
Decision Tree Classification was proposed (Chandra et al 2006) as improvement over SLIQ
where the decision tree is built by examining reduced number of split points and maintaining
the same classification accuracy. Prediction of forest cover types using Decision Trees is
discussed in this paper. Maximum accuracy of about 84% is achieved using Decision Trees.
T3.10I
e-Sagu: The Next Generation IT-based Query-less, Cost-effective
and Personalized Agro-Advisory System

P.Krishna Reddy and GV Ramaraju
Media Lab Asia Project, IIIT, Hyderabad
eSagu is a tool for IT-based personalized Agro-Advisory system. (“Sagu” means cultivation
in Telugu language). It aims to improve farm productivity by delivering high quality
personalized (farm-specific) agro-expert advice in a timely manner to each farm at the
farmer’s door-steps without farmer asking a question. The advice is provided on a regular
basis (typically once a week) from sowing to harvesting which reduces the cost of cultivation
and increases the farm productivity as well as quality of agri-commodities. In eSagu, the
developments in IT such as (database, Internet, and digital photography) are extended to
improve the performance of agricultural extension services. The eSagu system offers next
generation agro-advisory tool, and supplements and integrates into the existing agricultural
extension system. The development of eSagu was started during Kharif season of 2004. The
eSagu system was implemented by delivering advisory to 1051 cotton farms for the farmers
of three villages in Warangal district in Andhra Pradesh. During 2005-06, eSagu system for
5000 farms has been implemented for Cotton, Chilies, Rice, Groundnut, Castor, and Redgram
farms in 35 villages spread over six districts in Andhra Pradesh. In addition, the eSagu system
is developed to deliver advisory for fish. The farmers are happy with the expert advice as it is
helping the farmers to improve input efficiency by encouraging integrated pest management
(IPM) methods, judicious use of pesticides and fertilizers by avoiding their indiscriminate
usage. The turnaround time for advice delivery is 24-36 hours.The details of benefits accrued
to farmers for 2004 experiment were Rs. 3,820/- per acre with cost benefit ratio of 1:3 and
for 2005 experiment these were Rs 3,874/- per acre with cost-benefit ratio of 1:4.1.
T3.11I
Use of ICT in Agri Extension, Research, Development & Management:
Prospects and Challenges
Sanjiv Sharma
ITC International Business Division, Secunderabad
India is one of the dominant players in the world IT market. ICT and its benefits to rural India
are not new, infact it has been around for the last 10 yrs now. However, the efforts in this
direction have been minuscule in size. This paper discusses the prospects and challenges
relevant to use of ICT in AgriBusiness with special reference to the ITC imitative Echoupal.
T3.12I
Risk Analysis with Business Intelligence
Satyajit Dwivedi
SAS Institute India (Pvt) Ltd VIPPS Center, Greater Kailash-2 New Delhi
Effective Implementation of subsidy measures and management of quotas and licenses
become a crucial factor in the development of agricultural marketing. Large number of
applications is received from varied degree of beneficiaries from across any geography.
Manual selection process for approvals requires a huge battery of resources to manage the
mountains of data that are generated through the application process. It needs intensive

support from the computer department to analyse the data. New measures were expensive and
could only be integrated into the existing system with difficulty. A selection category could
only be rule based and the criteria forming the selection of beneficiaries were more
judgmental than scientific.
Data Mining based case allows a more specifically targeted selection, enabling the specialist
department to carry out data preparation and analysis on its own, so that it could be
independent of the IT department.
Selection of cases for Allocation and Monitoring of Area Subsidies, “Monitoring of Animal
Subsidies and Animal Marking is based on agricultural operations and their selection for onthe-spot
testing. Various models, such as: decision tree, neural network, regression analysis is
used for the variable selection and an efficient scoring model can be designed using factors
such as number of animals or areas covered by the application, the level of the possible
payment, changes over the previous year or proximity to limit values.
The data mining approach for efficient selection of application cases allows:
• a easier, more flexible approach to handling of data.
• a long-term solution
• simpler integration of new measures into the system
• faster availability of results
• development of an interdepartmental team of experts
• reduced dependency on the IT department
This approach of case selection is implemented in other audit scenarios such as Income Tax,
Customs, Property Tax and Sales Tax.