Manoeuvring unmanned vehicles through confined three-dimensional environments using Model Predictive Control

Current Unmanned Vehicle (UV) navigation systems are capable of navigating UVs through cluttered environments. However, these systems fail to manoeuvre vehicles through highly confined environments (environments with a higher density of obstalces then cluttered environments) because they do not explicitly consider the geometry and the dynamic characteristics and constraints of the vehicle in the navigation task.;This thesis presents a novel methodology that enables the navigation of UVs in confined 2D and 3D complex environments. The proposed approach employs a plan globally, react locally hybrid manoeuvring architecture in combination with a Model Predictive Control (MPC) methodology. This approach allows for a modular approach to navigation applicable to any UV. The use of MPC enables the integration of the UV's dynamic characteristics and constraints as well as the geometry of any UV in the navigation task. Simulation results in 2D and 3D environments for a number of vehicles are presented to demonstrate the applicability of the proposed approach.