Smartfarm is a technical trend to improve agricultural value chain through ICT convergence. To
build up a smartfarm system, ICT components like sensors, communication modules are
important. However, commercial sensor systems for local farms have some management
problems such as data loss and high maintenance cost because they use personal computers as
their user interface and data storage. The objectives of our research were designing the threelayers
architecture that solves these problems, implementing it to an automated weather station,
and evaluating it.
The three-layers architecture has the sensor layer, the data management layer, and the user
interface layer. In sensor layer, a sensor node collects data and transfers it. The data is stored and
managed in the data management layer. A user can analyze the data and extract information in
the user interface layer.
Based on this three-layers architecture, an AWS system consisting of a weather station, a data
management unit, and a monitoring application was developed. The weather station measures
seven meteorological factors. The data management unit manages the weather data and
communicates with the weather station and the monitoring application. The monitoring
application shows weather information to users. In order to evaluate the AWS system under realconditions, two systems were installed at local farms and compared with two conventional
systems. The data loss rate of new architecture system was under 1%.
As a result, the three-layers sensor system architecture could lessen the management burden of