Kalman Filters for Autonomous MAV Operations in Pitch Plane

Abstract

In autonomous control operations of unmanned micro air vehicles, the trajectories are generated first for a known mission profile and then they are reconstructed online. In some cases, a fixed-gain feedback controller is used to reconstruct the trajectories. However, for pitch plane autonomous operations of small fixed wing micro air vehicles, the human-made fixed gain feedback controller is sometimes inadequate. That is, the trajectory reconstruction process demands an online control system design with adaptive feedback gains. Although trajectory generation in pitch plane is a subject by itself, this paper focuses on the online control system design procedures so that the generated trajectories for autonomous control are reconstructed. It is shown that the extended Kalman filter in state and parameter estimation framework is an excellent tool to perform the control system design. The software that accepts an output from the trajectory generation module and performs trajectory reconstruction using a set of adaptive feedback gains is referred as ‘autonomous control agent’. A single input nonlinear micro air vehicle model is considered to illustrate the underlying principles of autonomous control agent with gains computed and tuned by the extended Kalman filter. Thus the human participation in micro air vehicle operation at pitch plane is minimized.