An optimal state estimator for a fuel valve actuator

We describe a digital implementation of a Kalman filter to estimate the true values of the angular position, the angular velocity and the coil current of a motor used to drive a fuel valve. By modeling the effect of coil current on the position sensor as a linear additive term, and the uncertainties in the actuator parameters as an additive random noise on the angular position, angular velocity and coil current, we have attempted to improve the robustness of the estimator. We believe that the storage and execution time requirements of this algorithm as implemented on the TI-320C2407 based actuator are comparable or less than the current model based algorithm.