Abstract
This paper presents a modeling of an automatic guided vehicle (AGV) to achieve a model-based control. The modeling includes 3 kinds of choices; a choice of input-output data pair from 14 candidate pairs, a choice of system identification technique form 5 candidate techniques, a choice of discrete to continuous transform method from 2 candidate methods. In order to obtain reliable plant models of AGV, an approach for calibration between a statistical model and a physical model is also here. In our approach, the models are combined according to the weight of AGV. As a result, our calibration problem is recast as a nonlinear optimization problem that can be solved by quasi-Newton's method.
