ABSTRACT
In DDF at the simplest level, sensors could communicate information and data directly to a CP where these data are combined [1]. The local processing of information/data is not required. Of course, the merit of having several sources of information/data is lost due to having a complete centralised control over the processing/interpretation of these data. When more processing occurs at local nodes, the computational and communication burden is removed from the fusion centre. But, this is at the cost of reduced direct control of lower level sensor information/data. Then the increase in the intelligence of the local sensor nodes results in a hierarchical structure for the fusion architecture. This has a merit of imposing some order in the fusion process. In yet some more distributed architectures, the sensor nodes have significant local ability to generate tracks and to do the fusion tasks, that is, blackboard and agent-based systems. Fully DCF have no CP and no common communication system, here, the nodes operate in a fully autonomous manner. They only coordinate via the anonymous communication information.
