Non-identifier-based adaptive control for a mechatronic system achieving stability and steady state accuracy

Non-identifier-based concepts, as High-gain-control, can be applied to systems with uncertain parameters and reveal the advantage of simplicity compared to most other adaptive control concepts. To overcome the drawback of a nondecreasing gain function, Funnel-control was introduced. Because Funnel-control is a time-varying proportional control law only, a steady state error remains in general. Theoretically, if no measurement noise occurs, this error can be made arbitrarily small. But in practical applications a large control error remains however. In this paper, two appropriate extensions are presented, which obtain an improved performance of the controller, even in the presence of load disturbances and sensor noise. An integral control action together with a dynamic feedback are introduced, which yield a vanishing control error and provide active damping of mechanical oscillations in the shaft of an electrical drive system, even if the measurement signal is disturbed by noise.