Reduced-Order Model Based Control of the Flexible, Articulated-Truss Space Crane

The large size and extreme flexibility of an articulated space crane entail tremendous difficulties in precision positioning. We demonstrate the application of a low-order control strategy for the flexible system using a linear, Timoshenko beam finite element model. The model captures the system's low-frequency characteristics. From selected modes of the finite element model, we design a reduced-order model (ROM) controller based on an arbitrary crane configuration. The ROM controller rotates the fixed-shape crane about the base alpha joint to actuate position control. This constructive control design takes place in two pieces: design for performance and compensate for stability. Ignoring the presence of residual modes, we design the low-order control law to achieve our performance criteria. Eigenvalue analysis of the closed-loop system, however, shows the destabilizing interaction of the ROM controller with residuals. Adding a parallel bank of residual mode filters, we restore system stability and performance, without redesign of the original controller. Finally, numeric simulation of the controlled finite element model reveals the effectiveness of a ROM-based control law with residual mode filter compensation. Application to a large-scale data-base highlights the utility of this low-order, noniterative controller design.