Error reduction for an inertial-sensor-based dynamic parallel kinematic machine positioning system

This paper addresses the development of an error reduction strategy for an inertial-sensor-based system for measuring the displacement of parallel kinematic machine (PKM) struts and hence the tool centre point pose. The error reduction strategy can significantly improve the performance of low-cost accelerometers and has many potential generic inertial measurement applications. Inertial measurement systems are prone to errors introduced by manufacturing error, system set-up error and environmental factors. To accurately measure displacement the errors and noise contaminating the inertial data must be removed. Based on data analysis, methods are developed in this paper to remove bias error, and correct integration error through band limitation, error modelling and velocity reconstruction. The velocity reconstruction algorithm described significantly improves the performance of the system and is novel in its use of the bounded nature of the strut displacement. The inertial measurement system described is evaluated using a single PKM-Strut test bed for linear movement and the experimental results are presented and analysed.