Particle Swarm Optimization-Based Multivariable Generalized Predictive Control for an Overhead Crane

The transient and residual vibrations in flexible underactuated mechatronic systems adversely affect the effectiveness and accuracy of performed tasks and movements. Moreover, in the case of crane operation the transient underactuated payload swing may present a safety hazard. In this paper, a novel control approach based on a multivariable model predictive control and a particle swarm optimizer is proposed for limiting the transient and residual swing of a payload transferred by an overhead crane. A control scheme is developed based on a discrete-time model approximating the decoupled dynamic of an actuated cart and an underactuated pendulum identified online using a recursive least-squares technique with parameters projection. A particle swarm optimizer is applied to determine the optimal sequence of control increments in the presence of constraints on input and output variables. The control scheme was successfully tested on a laboratory scaled overhead crane for different constraints and operating conditions. The experiments proved the feasibility and robustness of the proposed method.