Zanahul Farnaz, Husain S. Sajith, Anuradha D. Abeysekara, Shirantha Welikala, Prabath J. Binduhewa, Parakrama B. Ekanayake, and Lilantha Samaranayake
Closed loop control, DC motor, Kalman ﬁlter, state estimation,torque control
In most robotics and automation applications, DC motors play a pivotal role as they have proven to be one of the easiest to control. Implementing feedback helps precise control with improved noise performance and robustness against parameter variations. DC motors with feedback control can be used for position, speed, or torque used for control. For closed loop position and speed control
applications, the position and speed of the DC motors can be obtained using an encoder connected to the shaft, which usually is a cost-eﬀective solution. Similarly, the electromagnetic torque of a DC motor can be obtained by using a shaft-connected torque sensor. However, this is usually an expensive solution with additional power
and interfacing requirements. Further, such torque sensors can be too bulky to insert in a limited space environment and they are generally sensitive to vibrations. To overcome these problems with a low-cost solution, we use a state estimator to estimate the torque of the motor. We compare the performance of closed loop state estimator with a Kalman ﬁlter implemented for the same purpose. In both cases, the input voltage to the motor and the
speed output of the motor are the input to the state estimator, while the estimated torque is the output of the estimator. The estimated torque output is used for closed loop torque control of the DC motor. The state estimators and the torque controller, containing pulse width modulation and PID controller, have been implemented
in low-cost microcontroller-based environment so that the solution can be developed for large-scale manufacturing and oﬀered as a ready-made solution.