Comparación de desempeño de observadores de estado en sistemas lineales con aplicación a un motor de corriente continua

In many industrial processes, there are essential variables which need constant monitoring for process control tasks. However, often such measurement is complicated, or there is just no adequate sensor to perform this goal; this creates a significant challenge for engineers, who must somehow have measurement and control over system variables. An alternative for cases in which the access to the state vector is unfeasible is to obtain an estimate of the non-measurable states of the system through a state observer. The observer is a dynamic system whose states converge to those of the observed system, based on the system model and the observation of the input and output of the system. This information is used to estimate variables that are difficult to measure to make possible the full description of the behavior of the dynamic system. In this work, a set of state observers is designed and implemented to estimate the speed of a DC motor by measuring its stator current. Using different types of state estimators such as Luenberger Estimator, Kalman Filter for Linear Systems and the Sliding Mode Observer, we intend to estimate the DC motor speed correctly based on its mathematical model and stator current measurement. In both the simulations and the implementation of the observers, the response of these to unexpected perturbations that could alter the standard conditions of the system is analyzed and compared. The results suggest that each observer has specific conditions where their performance is enhanced.