Linear and hammerstein subspace identification of a human controlling a ball and beam using the quanser engineering trainer haptic interface

Identification of human subjects has widespread applications. In this paper, we perform system identification of a human controlling a ball and beam setup through a Quanser engineering trainer haptic interface. The identification of the human in the loop is performed using subspace-based nonlinear identification algorithms. Several combinations of inputs and nonlinearities in inputs are tested. The identified controllers are implemented through Simulink. Results for two human subjects with and without the haptic feedback and a pre-programmed PID controller are compared