Dynamic parameter identification of the rigid body model of an elastic laser cutting machine

Identifying dynamic parameters of an elastic robotic system can be a quite challenging task. This is because elastic robotic systems are often highly underactuated and sensor data may only be available for the driven states. To overcome this shortcoming, the identification process can be addressed by different experiments based on different reduced models. However, reducing a model leads almost always to redundant dynamic parameters. This work addresses a method how to deal with these redundant parameters and how to combine the results of different experiments based on confidence weights.