Optimized State Estimation Algorithms for Fast Hit-to-Kill Engagement

Propagation Research Associates, Inc., (PRA) introduces two innovative algorithms to reduce fire control and missile filter transients in short time-of-flight guided missile engagements. A class of Smoothed Iterated Filters is developed as a more efficient implementation of the forward-backward filter and is shown to significantly reduce the impact of filter transients due to initialization error. PRA proposes a Maximum Likelihood Estimation (MLE) algorithm that is computationally efficient and minimizes filter transients. PRA also proposes to investigate the utility of other approaches such as Levinson Auto-Regression and use of external fire control radar data uplinked to the missile to provide improved filter initial states which, in combination with the Smoothed Iterated Filter and/or MLE, will achieve even more reduction in filter transient response. Basically, filter transients are a function of the accuracy of the initial parameters of a filter. For the Kalman filter, the initial parameters consist of a state vector and a covariance matrix. The Smoothed Iterated Filters and MLE provide algorithms that improve the initial state and covariance estimates using the available measurements onboard the missile interceptor. Also, using external data that can be uploaded to the missile either pre-launch or post-launch has the potential to improve filter initialization.