Hydraulic servo system command shaping using iterative learning control

Fluid power systems are utilised in a wide variety of industrial applications. Such systems are predominantly controlled by PID controllers that are often very conservatively tuned. In applications such as structural test machinery where demanding reference trajectories are required to be tracked with minimum error, the standard feedback control approach is often inadequate. In this paper the performance of a command shaping approach based on iterative learning control is evaluated as a potential solution to such problems. An adaptive gradient algorithm, recently developed at Sheffield, is shown to perform well on a hydraulic test facility that exhibits significant nonlinearity in its dynamic response.