Chandrasekhara B. Panathula, Farbod Fahimi, and Yuri B. Shtessel
Model-predictive control, input saturation, mobile robot, 3D terrain, slip prevention
Nonlinear model predictive control method is used for dynamic control of Hilare robots on slippery 3D terrains, while preventing wheel slip. The dynamics of the Hilare robot is deduced by considering 3D pose of the robot while moving on a hilly terrain. So, the dynamic equations and the control law consist of roll and pitch angles, representing the slope of the terrain. These Euler angles are fed forward to the system in real-time to attain more reliable controller performance. The control inputs are the longitudinal and the lateral traction forces of the wheels. It is assumed that a higher-level estimator finds the static coefficient of friction if sliding happens. The maximum affordable longitudinal and lateral static friction forces are used as the saturation constraints on control inputs. When the saturation limits are met by the controller, driving forces are less than the maximum affordable longitudinal and lateral static friction, which leads to the prevention of the wheel slip. The Hilare robot is simulated while tracking a circular path on a terrain of known static friction. The simulation results show that the designed controller is successful in allowing the Hilare robot to track the desired trajectory on the slippery hill of known static friction, while eliminating wheel slip. It is shown that a conventional controller fails under the same simulated conditions.
Important Links:
Go Back
