The statistical convergence properties of Cramér-Rao bound in trajectory identification system with incomplete signals

The Cramér-Rao lower bound (CRLB) is discussed in the case where the signal measurements are lost in a random fashion. The statistical convergence properties of the CRLB as a function of the random arrivals of the signal measurements are investigated and a linear matrix inequation (LMI) approach of the steady-state CRLB is presented. The relation between CRLB and intensities of model noise is analyzed under the condition of a given measurement noise, then a new filter that admits the incomplete measurements system to have model noise with intensity as large as possible is designed with constrains of variance index in. The engineering sense in target tracking is that filter is designed to fit the maneuver area of target as large as possible with a given capability of detector. An illustrative numerical example of the trajectory identification system is provided to demonstrate the usefulness and flexibility of the proposed design approach.