Sensor Networks for Agriculture in South India
Jacques Panchard, EPFL, Hydram, email@example.com
The goal of the COMMON-Sense Net project was to design and develop a decision-support tool for small agriculture, based on environmental data as collected by a wireless sensor network.
Fresh water is a vital but scarce resource for mankind. However, it is often used very inefficiently. According to most estimations, 40% of the water used for irrigation around the globe could be saved without impact on yield, provided agriculture adapted its practices. Since this sector accounts for two-thirds of the fresh water consumption worldwide, the gains can be significant.
To save water in irrigation, one needs to know the hydric needs of the plant at a given time, which means knowing its environment with precision. As a consequence, monitoring the rural environment for a better use of water resources has become a crucial challenge, especially in semi-arid developing countries, which suffer most of water shortage.
With their low cost, low maintenance and ease of deployment, wireless sensor networks emerge as a promising technology that is ideally adapted to such a problem. With the help of sensors communicating with each other wirelessly, it is possible to build an integrated picture of the field's environment, on which one can apply crop prediction models that feed a decision-support system for farmers.
The project: COMMON-Sense Net
The scientific partnership was built around a multidisciplinary theme: the use of new information and communication technologies for poverty reduction in rural semi-arid regions. In particular the theme of enhanced environment monitoring was explored.
The questions that this research sought to address are
- Can a better knowledge of their environment help the poor rain-fed farmers improve the yield of their crop?
- What would be the precise context of use of an environment monitoring system targeted at this population?
- What are the technical requirements of such a system?
- Can we implement and assess such a system using new environment monitoring techniques?
- Can we reach out to the farming population, or are scientists or NGO a necessary intermediary at this point?
The challenges are threefold:
- Scientific: develop the capacity to use and process fine-grained environmental data in an innovative context
- Technical: design and implementation of a system
- Socio-cultural: the necessity to build capacity and ownership of a disruptive technical system in a rural context.
For this, a multidisciplinary and multicultural team was assembled, partly in Switzerland, partly in India (see bellow). In addition, a Karnataka-based non-governmental organization, the Chennakeshava Trust contributed its broad and deep knowledge of the rain-fed agriculture. Finally, repeated contacts were maintained with the University of Agriculture Sciences, Bangalore (UAS). The project partners defined broad use cases in collaboration with agricultural scientists and a local association of farmers. This led to the design, implementation and deployment of a prototype. Throughout the project, the developed system collected extensively data useful for a better understanding of the effects of water - and possibly other environmental parameters - on agriculture, making possible the implementation of replicable strategies. In the project’s concluding phase, extensive meetings with agricultural scientists led to the definition of precise applications to be conducted in a semi-controlled environment for the benefit of small rain-fed agriculture. The project also investigated possible direct use of these data by the targeted populations, as well strategic decisions necessary to reach them.
Special care was given throughout the project to build a transversal partnership with each partner, investing them with the same responsibilities and an equal involvement and participation to the project strategic and operational aspects. This was done through regular audio-conferences and the maintenance of an exhaustive project management web-site. This proved to be efficient in building trust and ownership among the different teams.
The involvement of partners coming from very diverse background proved to be a communication challenge at times, but lead to positive and promising results that could not have been reached without it.
Despite the pervasiveness of communication technologies, extended geographic proximity (at least several months) remains a must, especially at the bootstrapping phase of a project. This aspect was somewhat overlooked initially, which led to inefficiencies in project management.
In the framework of this project, the partners:
- identified user needs and use cases relevant to WSNs (Wireless Sensor Networks) in agriculture in developing countries, in particular as a decision support tool for crop selection, as a benchmarking tool for the efficiency of practices such as underground drip irrigation, or as an element of an early warning system against pests and diseases.
- designed and developed a system that can address the raised issues and the use cases
- ran a user survey with agricultural scientists coming from diverse background (agronomy, entomology, crop physiology, soil physics, etc.) based on the data gathered in the field, in order to assess the usefulness and usability of the application
- explored specific requirements of wireless sensor networks in the area of users’ acceptance and ergonomics
Implementation and Use
The project partners designed, implemented and assessed a concrete system that can be used for the benefit or rain-fed farming.
They realized that direct use of the system by the targeted population was still problematic.
However, they raised keen interest in the scientific community working on rain-fed farming, which saw such an environmental monitoring system as a potential tool for applied research in the near future.
A project proposal was submitted along these lines to the EPFL-SDC 2008 Cooperation Fund. Notification of acceptance will be given by the end of 2008.
- Rain-fed farmers need to know better their environment in order to increase their yield and reduce their risks.
- Enhanced environmental monitoring can help small agriculture to achieve this goal.
- However, due to the cost and complexity of the technology, applications targeted at applied research are more promising than direct application in the field, at least for the years to come.
- Wireless sensors are in need of new and innovative interfaces, if they are to be handled by non-specialists (people without significant background in wireless networking)
Academic partner in Switzerland:
Ecole Polytechnique Fédérale de Lausanne:
LCA: Laboratory for Computer Communications and Aplications (http://lca.epfl.ch)
HYDRAM: Laboratory of Hydrology and Land Development (http://hydram.epfl.ch)
Academic partner in India
Indian Institute of Science: Most important post-graduate studies institution in India
CEDT: Centre for Electronics Design and Technology (http://www.cedt.iisc.ernet.in)
Indian civil society partner:
Chennakeshava Trust: A local NGO of farmers in Karnataka.
Prof. Jean-Pierre Hubaux (EPFL) firstname.lastname@example.org
Prof. H.S. Jamadagni (IISc) email@example.com
Figure 1 Marginal Farmers in CKPura
Figure 2 COMMON-Sense Net: Elements of a wireless sensor
Figure 3 A Wi-Fi transmitter in the fields