Ground Trajectory Control of an Unmanned Aerial-Ground Vehicle using Thrust Vectoring for Precise Grasping

This paper presents the development of a novel hybrid unmanned aerial-ground vehicle (UAGV) to improve the success rate in grasping tasks. With the proposed method, the vehicle lands near the object and attempts grasping in its ground modality. A passive propulsion mechanism using motorized deflectors is employed for the ground locomotion. A nonlinear thrust vector model for ground locomotion is derived and validated in SolidWorks simulations. A nonlinear model predictive controller (NMPC) is employed for precise ground navigation. Multiple experiments are conducted with different initial conditions to demonstrate the performance of the proposed UAGV and NMPC for autonomous grasping. It is shown that the system is able to reliably reach and grasp the object when initially aligned with the object at distances between 1 and 1.5 meters.