Using iterated extended Kalman Filtering for estimation of a hexapod flight simulator motion state

Abstract

The six degrees-of-freedom Stewart platform, or hexapod, is in widespread use in the flight simulation industry for the generation of motion cues that are representative of those experienced in actual flight. For closed-loop control of such motion platforms, but also to be able to objectively assess the quality of the generated simulator motion, accurate measurement of the kinematic state of the motion platform is required. In current practice, the inference of such knowledge relies mainly on the isolated use of actuator length measurements and on, in certain cases, on-platform inertial sensors. The purpose of the current work is to extend a previously proposed and conceptually superior method, based on a tightly-coupled fusion of measurements provided by these sensors using the Iterated Extended Kalman Filter (IEKF). Results from computer simulations indicate that the IEKF has difficulty in converging to the true system state of a six degrees-of-freedom Stewart platform. This is because of the considerable increase in nonlinearity of the platform kinematics. Future research should therefore focus on the application of more advanced filters. In addition, further extension of the sensor fusion scheme using other types of sensors is investigated.