Speaker
Ming Li
(Mechanical Engineering)
Description
In the inside engineering of DEMO, the robotic machines or manipulators are foreseeable to be widely employed, which often have to deal with the demanding working conditions. The construction of the dynamic model of the robotic machine or manipulator can not only benefit the performance evaluation of the manipulator in the early design stage, but also can be incorporated into the control system of the robot or manipulator, in practical level, to gain the high control performance. However, in practice, it is rather difficult to construct accurately the analytical dynamic model for the robots or manipulators. The reasons behind include, but not limited to, lacking the physical insight of some dynamic phenomenon, the inaccuracy or infeasibility of the direct measurements or the deviation of some dynamic properties after the robot or manipulator’s assembly and deployment.
A method of constructing the dynamic model of robot with the unknown parts is proposed. The method can identify the unknown parts of the dynamic system by incorporating a BP neural network that will substitute the unknown parts in the system after the well training. A modified Levenberg-Marquardt algorithm is developed for the training of BP neural network, which can back propagate the errors between entire actual system and the constructed model into the training process of neural network.
For general application, an example of constructing the dynamic model for a general second order mechanic system with unknown dynamic component is presented. For the further validation on the complicated structure, the method is applied to a 10 DOF robotic machine. The friction models in the robot are taken as the unknown dynamic parts. After incorporating the BP neural network, the dynamic model of the entire robotic machine are successfully established.
The proposed dynamic model identification method can also be applied to a general case.
Co-authors
Antony Loving
(Remote Applications in Challenging Environments, UKAEA, Culham Science Centre, Abingdon, Oxfordshire, OX14 3DB, United Kingdom)
Heikki Handroos
(Mechanical Engineering, School of Energy Systems, Lappeenranta Uinversity of Technology, Lappeenranta, Finland)
Huapeng Wu
(Mechanical Engineering, School of Energy Systems, Lappeenranta Uinversity of Technology, Lappeenranta, Finland)
Matti Coleman
(EUROfusion Consortium, Boltzmannstr.2, Garching 85748, Germany)
Ming Li
(Mechanical Engineering, School of Energy Systems, Lappeenranta Uinversity of Technology, Lappeenranta, Finland)
Robert Skilton
(Remote Applications in Challenging Environments, UKAEA, Culham Science Centre, Abingdon, Oxfordshire, OX14 3DB, United Kingdom)
Yongbo Wang
(Mechanical Engineering, School of Energy Systems, Lappeenranta Uinversity of Technology, Lappeenranta, Finland)
