Sensor resource management with level 2 fusion using Markov chain models

Advanced optimization-based algorithms for sensor resource management have been a recent research focus area in multisensor tracking and fusion. These algorithms offer the potential for automating the sensor management process in response to level 1 (object or track-level) sensor fusion estimates. We have previously presented a hierarchical target valuation model that extends the target valuation model to include not only level 1 fusion information but also level 2 (group level) fusion information. We will use previously developed recursive composition inference techniques (specified using Bayesian inference techniques) that can efficiently and optimally reason about the identity of military units given partial observations of constituents in order to modify the sensor resource management target valuation algorithm. In this paper, we will develop new modifications to Markov state transition models that allow parameterization and approximate characterization of ground truth scenarios.