Neural fractional-order control of telescopic truck cranes

A telescopic boom crane is a strictly underactuated system in which three actuators track five motions. The boom luffing and extension cause variations in the shape and structure of cranes. The dynamic reaction between payload and boom through a flexible rope and viscoelastic cylinders induces significant vibrations. Additionally, a crane contains many uncertain parameters, suffers wind disturbances and sometime encounters failure of actuators. On the basis of a complex model with five fully nonlinear differential equations, we construct an intelligent robust control system for such cranes, which integrates the benefits of advanced techniques. Fractional-order control combined with robust control produces a controller holding flexible derivative orders. A multilayer perceptron network provides a learning feature to the cranes, in which an adaptive mechanism estimates one equivalent component represented for individual influences. Simulation for a practical crane shows the effectiveness of the proposed control system. © 2022 Elsevier Inc.