Sensor resource management using cost functions

When integrating track data across multiple platforms, one tries to create a super-platform that has performance capability which is greater than an individual platform. The performance of the super-platform ideally would be the superposition of the individual performances of the single platforms. From an estimation perspective, one of the key problems blocking this is prediction which degrades significantly with time delays or bias which are caused by sensor misalignments and threat miss-modeling. While prediction functions can be accomplished seamlessly in the background in a single ship environment, they become more problematic in a networked ship environment. Prediction that is not adequately dealt with can lead to staleness in the usefulness of information which causes temporal decoupling of the networked data so it fails to meet requirements which can cause network failures. We discuss some sensor resource management approaches that can be used to avoid common problems which lead to both estimation inadequacies and the introduction of delays into the network. The introduction of cost functions into the management of sensor resources is the means we propose to deal with this problem, so we discuss their application in some detail