Passivity-Based Adaptive Control of a 5-DOF Tower Crane

Abstract

Tower cranes are often used in construction to transport and lift heavy loads. They are typically controlled by human operators, and the speed, efficacy, and overall cost of their operation could be improved by automation. However, this is a challenging task due to the fact that the system is underactuated and highly nonlinear, which has limited the development of practical dynamic models and control methods of tower cranes. This project contributes to that goal with the derivation of a nonlinear dynamic model and adaptive control method that requiring little knowledge of the system parameters for precise and robust reference tracking. An adaptive control input is derived that ensures the tower crane features a passive input-output mapping. A novel approach is developed to bound the time derivative of the system's mass matrix, which is a critical part of the proof of passivity. Robust closed-loop input-output stability is proven using the Passivity Theorem. Experimental tests are performed, showing the effectiveness of the control law on the three- dimensional tower crane.