Hydraulic manipulator virtual decomposition control with performance analysis using low-cost MEMS sensors

Abstract

This paper presents closed-loop motion control of a heavy-duty hydraulic manipulator using non-linear model-based Virtual Decomposition Control (VDC), where the motion feedback is estimated solely with low-cost micro-electromechanical systems (MEMS) inertial sensors. By virtually decomposing the strongly non-linear and dynamically cross-coupled manipulator system into individually controlled subsystems, a significant improvement in overall control performance is achieved. The controller performance is analysed using planar Cartesian end-effector motion. The experiments show that the stability-guaranteed VDC approach based on low-cost MEMS sensor feedback yields a high-performance control solution: with a 0.85 m/s maximum velocity, the end-effector has a peak tracking error of 13 mm, which is a notable improvement by a factor of 3.6 compared to our previous work based on linear state feedback control.