Control of Manipulators on Moving Platforms Under Disturbance
- Publication Type:
- Thesis
- Issue Date:
- 2020
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
Mobile robots, such as underwater vehicles, drones, and rovers, are now being combined with manipulators to perform a variety of work in the field. But current state of the art in control assumes that disturbances from the environment are minimal. However, the effects of wind, waves, and rough terrain may make it difficult for the vehicle to maintain a steady base for the manipulator. Or, in some cases, the vehicle may lack the control authority to negate disturbances in all directions. In this thesis, predictions of the base motion are used to formulate control strategies that enable a manipulator to proactively counter, and even make use of, these disturbances.
Time series and Fourier series are commonly applied to many predictive control methods in literature. However, there are contradictory results in performance for different applications. To clarify these discrepancies, an objective comparison in prediction performance is made between time series, Fourier series, and Gaussian Process Regression (GPR) using motion data from underwater robots in waves. Analysis of the forecast errors and uncertainties show that GPR can produce better short-term. Furthermore, time series was found to be overconfident in the prediction, whereas Fourier series had the largest uncertainty.
A predictive control method is then presented that enables a manipulator on a free-moving platform to maintain a steady end-effector pose. By using forecasts of the base motion, the manipulator can anticipate and negate this disturbance. Simulations and experiments are conducted, and it is shown that the proposed predictive control can reduce tracking error by 60% compared to a PID feedback controller. Moreover, kinematic constraints can be satisfied whilst simultaneously minimizing task error.
A control strategy is also developed that allows a redundant manipulator to use the inertial forces produced by base disturbance to reduce joint torque. Further improvements are made by predicting changes in gravitational acceleration with respect to the manipulator. It is shown that joint torques can be reduced by 25% compared to a local minimization of the weighted torque norm.
Lastly, a torque minimization method is presented for redundant manipulators handling large external forces. Most literature only addresses the internal dynamics. This thesis presents a method to minimize torque from both an external loading and the internal dynamics. This method can be applied to manipulators on moving platforms, and further enhanced by incorporating the base motion predictions.
Please use this identifier to cite or link to this item: