Closed-Loop Linear Covariance Analysis for Hosted Payloads

Abstract

The concept of a hosted payload has numerous applications, ranging from remote sensing to turreted weapons and satellite communications. Components common to such systems include a host vehicle, a payload to sense or interact with the environment, an adjoining structure, and an accompanying navigation and control system. In practice, the performance of such systems is almost exclusively assessed via inefficient sample-based methods such as Monte Carlo analysis. In contrast, linear covariance analysis produces the same statistical information but in a single run. This paper develops a closed-loop linear covariance model capable of analyzing a wide range of hosted payloads. The current linear covariance theory is extended to model controllers typical of hosted payloads, with internal states, filters, and auxiliary control measurements. Nonlinear equations are developed to model the navigation and control systems of a hosted, gimbaled pointing payload. These equations are linearized about the reference trajectory to form a set of linear, time-varying differential and update equations to describe the closed-loop pointing performance of the payload. The accuracy of the linear covariance model is verified by comparison to Monte Carlo analysis and is then used to perform error budget analyses of the navigation errors and pointing dispersions.