This paper introduces a robust measurement scheduling strategy forlinear time-varying systems with zero-mean white process and measure-ment noises. Relying on previous works, the measurement strategy con-sists of choosing the timing and the measurement noise intensity prolesuch as to minimize under an integral constraint an upper-bound of theestimation error covariance matrix in a Kalman lter. The measurementstrategy is applied against a process noise intensity profile designed tomaximize that upper bound under a similar integral constraint. Theproblem is solved iteratively and is presented here for the case of a scalarprocess noise intensity. The result is a sequence of (few) epoch times atwhich measurements should be acquired, along with the optimized accu-racy levels, and a sequence of (few) epoch times and intensities, at whichthe process noise should be active. The proposed planning is valuablein providing guaranteed performances under uncertainty on the processnoise intensity. The proposed robust measurement planning method-ology is applied to a simple harmonic system for validation. It is alsoapplied to a relative navigation problem for two low Earth orbit satellitesying in formation and equipped with laser ranging capabilities.
展开▼
