An Optimal Moving Horizon Estimation for Aerial Vehicular Navigation Application

In this article, an optimal state is estimated using the moving horizon estimation technique (MHE), based on the minimizing the deterministic cost function defined for moving window with a finite number of samples at specific time interval. The optimal moving horizon observer was designed and implemented for the non-linear dynamic problem of aerial vehicle integrated navigation. The low grade commercial inertial measuring instrument (IMU) equipped with accelerometers and gyros sensors instrumented on-board in the strapdown configuration, is employed for collection of the real time experimental data. The data fusion algorithm of moving horizon estimation is realized and the results are collected from the offline algorithm testing on the Matlab software platform. Essential data processing and cleaning of data processing was conducted before algorithm application i.e. solving the multi rate sensors data synching and removing high frequency unwanted contents. Finally, the aerial vehicle dead reckoning integrated navigation was performed with recursive observer using IMU/GPS avionics. Contrary to the widely practiced extended Kalman filter results, recursive observer of MHE exhibited performance enhancement in the response and precision aspect, regardless of environmental noise and failure scenarios.