Robust object-mass measurement using condition-based less-error data selection for large-scale hydraulic manipulators

A practical framework for measuring the mass of an object grasped by the end-effector of a large-scale hydraulic manipulator, such as construction manipulators, is proposed. Such a measurement system requires high accuracy and robustness considering the nonlinearity and uncertainty in hydraulic pressure-based force measurement. It is thus difficult to precisely model the system behaviors and completely remove error force components, so our framework adopts a less-error data selection approach to improving the reliability of the measurand. It first detects the on-load state to extract reliable data for mass measurement, including evaluating measurement conditions to omit sensors in indeterminate conditions and redefining three-valued outputs such as on, off, or not determinate, to improve robustness, then extracts data during the object-grasp state identified by the grasp motion model and removes high-frequency component by a simple low-pass filter, to improve accuracy, and finally integrates date from plural sensors using the posture-based priority and averages all selected data, to improve reliability. Evaluation experiments were conducted using an instrumented hydraulic arm. Results indicate that our framework can precisely measures the mass of the grasped object in various detection conditions with less errors.