Sensor and Actuator Modelling Issues in Achieving the Tracking Control of a Pendulum

R. Madhavan


Pendulum, sensors, actuators, tracking control, parameter estimation


The states associated with a particular model are not always available, and thus the estimation of such states becomes inevitable where the missing states are reconstructed from measurements provided by a sensor or sensors. Such measurements are usually corrupted to some degree by the presence of noise, biases, and device inaccuracies. Sufficiently accurate modelling and understanding of the sensors and actuators is necessary to arrive at numerically efficient state estimation and control algorithms that enable online implementations. This article describes the sensor and actuator modelling issues that were encountered while trying to achieve the tracking control of a simple pendulum. Five combinations of three sensors, the encoder, the rate-gyro, and the tilt-sensor are utilized to measure the pendulum’s joint angle or the joint velocity or both. A brushless DC motor is employed for actuation purposes. The sensors are analyzed for sources of errors that are detrimental to achieving the tracking control of the pendulum and are quantified via experiments.

Important Links:

Go Back