Improved Representations of Sensor Exploitation for Automatic Sensor Management
J.P. Hill · 2003
Advanced optimization-based algorithms for sensor resource management have been previously developed. These algorithms ofler the potential for improving the sensor control process for the Level I (track-level) fusion problem. Previous studies have indicated that sensor resource management algorithms may have limited value in certain operational scenarios involving multi-platform surveillance and strike missions because the response is optimized for track maintenance without any assessment of situation context. In this paper, we will develop a fi-amework for representing the expected information value of planned sensor measurements as it contributes to higher-level situation inferences. We will extend previously-developed algorithms and concepts for semor resource management to include target valuation as a function of not only discretized tracking and classijication states but also both high and low level fusion valuation, which will provide for improved tracking and classification performance when identihing higher-level object groups, such as convoys. This will rely on a computationally eficient implementation of Bayesian modeling and inferencing algorithms,