Nonlinear model predictive control of a rotary crane system using on-line optimization

This paper investigates the implementation of a predictive controller using nonlinear dynamic model derived from first principles, together with online optimization, for a rotary crane system. The system is driven by two motors in velocity mode, and a white LED is installed on the payload to be observed by a camera located on the rotary frame to derive swing angles for control feedback. The proposed prediction controller is featured with control horizon long enough for online computation of prediction and optimization algorithms. Possible control sequences within the control horizon are described using spline interpolation to reduce the number of parameters to be optimized online. Response of the proposed scheme in face of exogenous disturbances and inaccuracy in estimated string length is demonstrated by numerical simulations.