Active vibration reduction for elastic pick-and-place systems by observer-based state feedback

An approach is presented that allows the reduction of oscillations in a system to improve accuracy. The novelty lies in the modelling approach with two important advantages: First, the system structure in the observer is modified to remove the dependence of non-linear friction effects from the model. Secondly, and more importantly, the approach allows the necessary state information, in particular the displacement, to be estimated by an acceleration measurement (instead of a position measurement) of the machine base combined with a position measurement of the actively controlled axis. It is shown that these measurements and the knowledge of the system dynamics are sufficient to reconstruct the base position with high accuracy using the presented approach. To demonstrate its effectiveness, this information is then used in a controller to reduce system oscillations. The concept is implemented on a laboratory system to demonstrate the practicality of the presented approach.