Investigation of an inertial-sensor-based dynamic position measurement system for a parallel kinematic machine

The key to the positional accuracy of a Parallel Kinematic Machine (PKM) is the precision to which the lengths of the individual parallel links (or struts) can be measured. The strut length can be measured to a high degree of precision using a laser interferometer but this is expensive. Alternatively, encoders may be used to measure the relative change in length of the strut but they can only record ball screw or motor rotation, and cannot detect any deformations that may occur within the machine’s structure. In addition, only limited information on any dynamic disturbances such as vibration can be obtained. To obtain this type of data, displacement should be directly measured at the end of the strut. The use of inertial sensors offers a possible solution since they are self-contained, relatively low cost and can be easily mounted on the machine structure. However, an inherent limitation of inertial systems is the growth, with time, of errors in measured velocity and position; in the system described within this paper, they are corrected by using an external reference measurement and Kalman filtering. Through the integration of the two measurement systems, a velocity profile containing improved dynamic information can be calculated. The key steps required for the integration of the inertial and encoder measurement systems are introduced, and include the formulation of a system and measurement model and Kalman filter estimation and testing. In order to investigate the feasibility of the proposed measurement system, it was evaluated using a single PKM strut. The experimental results are presented and analysed.