Soft-constrained predictive control for an overhead crane

Reduction of transient and residual payload swing in crane systems is a key control objective to guarantee the safety and efficiency requirements. The fast and accurate payload positioning with swing suppression within the acceptable range to avoid accidents is the challenging problem due to the underactuated nature of crane systems. Since the actuated motion causes undesirable payload swing, the efficient control method should be developed to ensure fast and precise payload positioning and meet the safety requirements. The standard model predictive control method is not suitable for underactuated mechanical systems. In this article the two, soft and hard-constrained antisway predictive control strategies are compared in experiments carried out on a laboratory scaled overhead travelling crane. The both control schemes are developed based on the linear parameter-varying model of a planar crane system. The recursive least square algorithm with parameter projection is used to estimate the model parameters. The soft-constrained optimization problem is solved using the particle swarm optimization algorithm with the inertia weight linearly decreasing during iteration. The metaheuristic optimizer is applied to determine the sequence of optimal control increments subject to the hard constraint of the control input and soft constraint of the payload swing. The comparison of hard and soft-constrained predictive controllers is carried out on a laboratory stand for different payload deflection constraints.