Control of sensor information in distributed multisensor systems

Provides an analysis of error covariance control techniques for allocating sensing resources in distributed, multiprocessor, multisensor systems. We present two algorithms for allocating sensing resources that manage the rates and resolutions at which sensor information from various nodes is processed. An elliptical annulus described by two covariance matrices is used to control the prediction and update covariances in the decentralized Kalman filter. These algorithms allow for nodal autonomy by letting each node control the usage of its own suite of sensors. With a single state filter, these sensor management techniques are shown to result in a discrete periodic Riccati equation