by Jose Cazalilla, Marina Valles, Vicente Mata, Miguel Diaz-Rodriguez, Angel Valera
Abstract:
A model-based control system may produce a substantial increase in the overall performance of parallel robots, thus allowing less expensive manufacturing. Although dynamic parameter identification is an experimental technique that significantly improves data feeding dynamic models, some difficulties arise: the effect of the parallel robot's unmodeled dynamics, the variation of some dynamic parameters over time (e.g. friction parameters) and, even more importantly from an operational point of view, the payload, which is normally not included in the identification process for practical reasons. Adaptive control techniques offer the possibility to deal with these uncertainties. Therefore, an adaptive control in joint space has been developed in this paper in order to control a low cost 3-DOF parallel manipulator. The novelty of the proposed controller is that it considers, as a starting point, simplified dynamic models based solely on experimentally identified relevant dynamic parameters. In addition, an adaptive strategy has been applied to various scenarios where it is assumed that the uncertainties affect: 1) rigid body parameters, 2) friction parameters, 3) actuator dynamics, and 4) all the aforementioned. Simulation and experiments on an actual parallel robot were conducted to evaluate the performance of the controller, which was implemented in a modular way using Open RObot Control Software (OROCOS). Finally, experiments which consider the placement of a payload onto the platform were conducted over the actual prototype. Results demonstrate that the proposed adaptive schemes improve position-tracking performances when a payload is added to the platform
Reference:
Adaptive control of a 3-DOF parallel manipulator considering payload handling and relevant parameter models (Jose Cazalilla, Marina Valles, Vicente Mata, Miguel Diaz-Rodriguez, Angel Valera), In Robotics and Computer-Integrated Manufacturing, Pergamon, volume 30, 2014.
Bibtex Entry:
@article{cazalilla2014a,
title={Adaptive control of a 3-DOF parallel manipulator considering payload handling and relevant parameter models},
author={Cazalilla, Jose and Valles, Marina and Mata, Vicente and Diaz-Rodriguez, Miguel and Valera, Angel},
journal={Robotics and Computer-Integrated Manufacturing},
volume={30},
number={5},
pages={468--477},
year={2014},
publisher={Pergamon},
doi={doi:10.1016/j.rcim.2014.02.003},
url={https://www.researchgate.net/profile/Miguel_Diaz-Rodriguez/publication/261370562_Adaptive_control_of_a_3-DOF_parallel_manipulator_considering_payload_handling_and_relevant_parameter_models/links/53ecb02b0cf24f241f15984b.pdf},
gsid={https://scholar.google.com/scholar?oi=bibs&hl=es&cites=13532222029507023686,14535037694267713204},
abstract={A model-based control system may produce a substantial increase in the overall performance of parallel robots, thus allowing less expensive manufacturing. Although dynamic parameter identification is an experimental technique that significantly improves data feeding dynamic models, some difficulties arise: the effect of the parallel robot's unmodeled dynamics, the variation of some dynamic parameters over time (e.g. friction parameters) and, even more importantly from an operational point of view, the payload, which is normally not included in the identification process for practical reasons. Adaptive control techniques offer the possibility to deal with these uncertainties. Therefore, an adaptive control in joint space has been developed in this paper in order to control a low cost 3-DOF parallel manipulator. The novelty of the proposed controller is that it considers, as a starting point, simplified dynamic models based solely on experimentally identified relevant dynamic parameters. In addition, an adaptive strategy has been applied to various scenarios where it is assumed that the uncertainties affect: 1) rigid body parameters, 2) friction parameters, 3) actuator dynamics, and 4) all the aforementioned. Simulation and experiments on an actual parallel robot were conducted to evaluate the performance of the controller, which was implemented in a modular way using Open RObot Control Software (OROCOS). Finally, experiments which consider the placement of a payload onto the platform were conducted over the actual prototype. Results demonstrate that the proposed adaptive schemes improve position-tracking performances when a payload is added to the platform},
keywords={adaptive control, dynamic parameter identification, model-based control, Parallel robot},
}