On-line parameter estimation of PMDC motors using binary-valued speed measurements

Establishing real-time models for permanent magnet direct current (PMDC) motors is of importance for capturing authentic dynamic behavior of the systems to improve control performance, robustness, and diagnosis. Sensor types play an essential role in such an endeavor, affecting system costs, reliability, estimation accuracy, and control strategies. In this paper we develop an estimator that can be used to estimate the model parameters of PMDC motors using binary-valued speed measurements. A typical linearized model structure of PMDC motors is used as a benchmark platform to demonstrate the technology. The input/output relationship between the input voltage and the output speed is modeled by a suitable regression structure for binary system identification. Periodic input dithers are employed to identify the speed of the motor and to estimate the unknown parameters of the system. Simulations are performed to illustrate the utility of the technology. The advantage of this work is to support on-line estimation using cheaper and low-complexity binary sensors with guaranteed estimation accuracy.