Framework for Implementation and Evaluation of Estimation and Control Methods for Unmanned Aircraft Systems

Author(s):

Amoolya S. Tirumalai

Abstract:

A state estimation and control software package was developed that can be used to integrate unconventional sensing modalities, estimation methods, and control methods. The package accounts for asynchronous sensing and control and is implemented in C++ with the Robot Operating System 2 message-passing interface on the PX4 flight stack. The package is evaluated in Gazebo simulation. The modularity of this package is documented with respect to sensor modules, estimators, and controllers. To demonstrate basic capability, the extended Kalman filter for a UAS and a linear-quadratic-integral controller are implemented. The estimator and controller perform well in simulation, successfully allowing trajectory matching for a figure-eight path. Difficulty was encountered when integrating onto physical hardware due to limitations in the PX4 flight stack, necessitating that future work explores other flight stacks, such as Ardupilot.