Gantry Robot Control with Observer based on Subspace Identification with Multiple Steps Data

When working with complex dynamic systems such as robotic systems needs to have a good estimate of the various terms that make them dynamic equation. From this estimate may provide various control algorithms (usually nonlinear) to compensate or take into account this dynamic equation. The problem that often arises is twofold: one part is not trivial to obtain the dynamic terms, and secondly, the control algorithms usually have high complexity, which may make it impossible to implement them in real time environments. This paper deals with obtaining the system state estimation using subspace-based system identification that not only get an estimate, but also can be used to calculate and implement control algorithms. This method has been validated on a gantry robot with multiple data sets. We have also designed a optimal control with a state observer using the estimated state space systems of the gantry robots. The results are very satisfactory