A Middleware for Managing Sensory Information in Pervasive Environments

Abstract

Context and context-aware computing have attracted remarkable attention in recent years. Research into wireless sensor networks is rapidly moving from simulations to realistic testbeds. The work presented here investigates the architecture and design of an agent-based sensor network for monitoring and spraying control of olive fly in an ubiquitous precision farming environment. It contributes by providing a generic, flexible and extensible model for handling heterogeneous sensor data, which can be managed using a simple user interface. The innovative aspect stands for the development of a multi-agent system and focuses on the communication aspects of the proposed middleware architecture. The modeling allows the encapsulation of different modules as agents and integrates them, by utilizing a standardized XML schema in order to hide complexity to the users. The proposed design adopts a layered architecture, which is based on some software agents that solve different tasks and communicate among themselves their results and requests