Automated tuning of Kalman filter: Kalman filter tuning in the Windows Azure Cloud environment

For accurate application of Kalman filters in a navigation solution (i.e., position, velocity and attitude), the tuning of its parameters is key, regardless of the quality of IMU. Dedicated tuning parameters must be established for every specific sensor by considering its unique characteristics and its underlying random noise. The problem of filter tuning has always been demanding in terms of operator and computer work loads. This study presents the automated tuning of Kalman filter parameters. In contrary to common tuning approaches, the presented procedure does not limit the number of parameter combinations to be examined. The method aims to evaluate filter performance with an extensive sample, in an efficient way - Azure Cloud framework is used. Moreover, the selection of the optimal set of parameters is based on minimizing multiple cost functions. As result, the tuning outcome is more robust and the tuning time is significantly limited.