A Joint-Angle Estimation Method for Industrial Manipulators Using Inertial Sensors

Abstract

This paper focuses on the use of low-cost inertial sensors in industrial manipulators. In particular, an easy-to-install joint-angle estimation method is proposed. This scheme can work independently from the primary encoder-based sensor, for the creation of a fault-detection system. It can also be the primary joint-angle sensor in low-cost robotic manipulators. A robot with six degrees of freedom has been used, with three inertial measurement units set on its arms; one for each pair of joints. Three cascaded extended Kalman filters have been used to estimate the joint angles by the fusion of the outputs of triaxial gyroscopes and accelerometers. The results of three experimental tests are reported in order to show the performance and limitations of this approach, in comparison with the simple gyroscope measurements integration.