Fundamental limitations on the performance of navigation systems

摘要

Following a nonlinear filtering formulation, the performance of navigation systems, using the linear dynamics and nonlinear measurements, is evaluated using the Cramer-Rao lower bound (CRLB). The Bobrovsky-Zakai version or CRLB for discrete-time Markovian systems, together with some practical simplifying assumptions regarding dynamics modeling and measurement functions, results in closed-form expressions for bounds on the performances of navigation systems. A simple transformation is shown to enable the application of the results for linear tracking problems, as bounds on the performances of nonlinear navigation systems. For more complex navigation models, a class of approximations for CRLB is suggested.