A Scalable Agent Architecture for a Dynamic Sensor Network
Mikael Brännström, Erland Jungert · Electronic Notes in Theoretical Computer Science · 2003
Information systems for networks of small and inexpensive sensors attached to computational nodes in distributed networks will in the future be required for applications where the focus is to assess information about different ground based targets in motion. Operations of concern to these applications are, for instance, to determine the type and track of these targets, processes that generally require methods for sensor data fusion. Of special importance is also that such networks must be able to expand and shrink with respect to the number of sensors and nodes, that is, a property called scalability. Clearly, traditional and centralized architectures are not suitable since they will lead to very complex systems that eventually will be very difficult to maintain and accommodate to changing requirements. For this reason novel systems architectures are required. In this work, an architecture based on systems of intelligent multi-agents is proposed. The architecture has been implemented and tested in a simulated environment.