Non-linear MPC based Navigation for Micro Aerial Vehicles in Constrained Environments
Björn Lindqvist, Sina Sharif Mansouri, George Nikolakopoulos
Abstract
This article establishes a novel Non linear Model Predictive Control (MPC) scheme for the navigation of a MAV (Micro Aerial Vehicle) in constrained environments, such as narrow passages, multi-obstacle populated spaces and tight openings. The proposed NMPC optimization framework is based on the Proximal Averaged Newton type method for Optimal Control (PANOC) and has the merit to employ a penalty method for the proper consideration of the obstacles and other environmental constraints during the navigation. The proposed scheme has the ability to be a fast solution for the navigation of MAVs that can be directly applied online and thus it is creating a powerful navigation framework for demanding flights. For achieving such an agile and fast aerial navigation, the article will also present the proposed penalty creation methodology for dealing with the obstacle avoidance and the space constraint navigation. Finally, the efficacy of the proposed scheme will be demonstrated by multiple simulation results under constrained and demanding environments.