Fuzzy estimation of DC motor winding currents

Fuzzy systems have been used extensively and successfully in control systems over the past few decades, but have been applied much less often to filtering problems. This is somewhat surprising in view of the close relationship between control and estimation. The paper discusses and demonstrates the application of fuzzy filtering to motor winding current estimation in permanent magnet synchronous motors. It is shown via simulation results that fuzzy techniques provide an attractive alternative to current estimation using analytic techniques. It is further shown how the fuzzy membership functions can be optimized using gradient descent.