1,721,056 research outputs found
A toolbox for input-output system inversion
Full text linkIn this paper a Matlab-based toolbox for the input-output system inversion of linear systems is presented. Different methods, based either on analytical or numerical approaches, are implemented. The toolbox can be exploited in the design of a feedforward action for control systems in different contexts in order to improve performances in the set-point regulation. The use of a pre-actuation and a post-actuation
time can be easily analyzed as well as the role played by the choice of the desired output function
Improving set-point following performance of industrial controllers with a fast dynamic inversion algorithm
No full textIn this paper, we propose a new method for the improvement of set-point-following performances of industrial controllers. Basically, this new approach consists of substituting the standard step signal to be applied to the closed-loop system (in which the controller has been previously selected) with a command input function that is determined by means of a dynamic inversion procedure. This procedure is based on a fast algorithm, and it is therefore suitable for application in an
industrial context, where the ease of implementation is of major concern. Simulation results show the effectiveness of the technique
Iterative feedforward tuning for residual vibration reduction
No full textAn iterative approach for the determination of an input-output inversion feedforward control law for residual vibration reduction is proposed in this paper. In particular, point-to-point motion planning of vibratory servosystems is considered. The method aims at estimating recursively the parameters of the system in order to determine the exact command input to be applied to the control system in order to achieve a predefined motion without oscillations. In this context, a gradient based minimisation of the integrated square error cost function is performed. Simulation results show the effectiveness of the methodology
Optimal dynamic inversion based control of an overhead crane
No full textA methodology is proposed to control the transient sway and residual oscillation of a payload carried by an overhead crane. The design approach is based on a linearised model of the
crane and consists of dampening the linearised system by an observer-based controller and applying a dynamic inversion procedure in order to assure a predetermined oscillation free
polynomial motion law for the payload. Polynomial functions are adopted in order to guarantee that the input function has a continuous derivative of an arbitrary order. Moreover, the motion time can be minimised, taking into account constraints on the actuators, by means of a simple bisection algorithm. Parameter uncertainties are taken into account during the whole design procedure. Simulation results, based on a nonlinear crane model, show how the method is also effective when the payload is hoisted or lowered during the motion, and when friction effects are considered
Optimal system inversion based motion planning for servosystems with elastic transmission
No full textOn the equivalence of model inversion architectures for control applications
No full textThe main inversion-based control architectures are the plant and closed-loop inversion architectures. For scalar continuous-time linear systems, these architectures are shown to be fully equivalent for both the minimum and nonminimum-phase cases when exact stable inverses are used. This equivalence, deduced by using a behavioral approach, dictates that the two architectures deliver the same performances for any disturbance and mis-modeling affecting the controlled plant. A simulation example highlights that the equivalence still holds in practice when a careful truncation of the preaction control is performed
- …
