Cubature H ∞ information filter and its extensions

Abstract State estimation for nonlinear systems with Gaussian or non-Gaussian noises, and with single and multiple sensors, is presented. The key purpose is to propose a derivative free estimator using concepts from the information filter, the H ∞ filter, and the cubature Kalman filter (CKF). The proposed estimator is called the cubature H ∞ information filter ( C H ∞ IF ); it has the capability to deal with highly nonlinear systems like the CKF, like the H ∞ filter it can estimate states with stochastic or deterministic noises, and similar to the information filter it can be easily extended to handle measurements from multiple sensors. A numerically stable square-root C H ∞ IF is developed and extended to multiple sensors. The C H ∞ IF is implemented to estimate the states of a nonlinear permanent magnet synchronous motor model. Comparisons are made with an extended H ∞ information filter.