Modeling and Control of the UGV Argo J5 With a Custom-Built Landing Platform
This thesis aims to develop a detailed dynamic model and implement several navigation controllers for path tracking and dynamic self-leveling of the Argo J5 Unmanned Ground Vehicle (UGV) with a custom-built landing platform. The overall model is derived by combining the Argo J5 driveline system with the wheels terrain interaction (using terramechanics theory and mobile robot kinetics), while the landing platform model follows the Euler-Lagrange formulation. Different controllers are, then, derived, implemented to demonstrate: i.) self-leveling accuracy of the landing platform, ii.) trajectory tracking capabilities of the Argo J5 when moving in uneven terrains. The novelty of the Argo J5 model is the addition of a vertical load on each wheel through derivation of the shear stress depending on the point’s position in 3D space on each wheel.
Static leveling of the landing platform within one degree of the horizon is evaluated by implementing Proportional Derivative (PD), Proportional Integral Derivative (PID), Linear Quadratic Regulator (LQR), feedback linearization, and Passivity Based Adaptive Controller (PBAC) techniques. A PD controller is used to evaluate the performance of the Argo J5 on different terrains. Further, for the Argo J5 - landing platform ensemble, PBAC and Neural Network Based Adaptive Controller (NNBAC) are derived and implemented to demonstrate dynamic self-leveling. The emphasis is on different controller implementation for complex real systems such as Argo J5 - Landing platform.
Results, obtained via extensive simulation studies in a Matlab/Simulink environment that consider real system parameters and hardware limitations, contribute to understanding navigation performance in a variety of terrains with unknown properties and illustrate the Argo J5 velocity, wheel rolling resistance, wheel turning resistance and shear stress on different terrains.