Comparison of IMU Measurements of Curling Stone Dynamics with a Numerical Model. (2016)

• Record Type: Journal Article
• Authors: Lozowski, Edward; Maw, Sean; Kleiner, Bernard; Szilder, Krzysztof; Shegelski, Mark; Musilek, Petr; Ferguson, Dana
• Is Part Of: Procedia engineering. Volume 147(2016)
• Page Start: 596
• Page End: 601
• Publication Date: 2016
• Subjects: curling -- inertial measurement unit -- numerical model -- ice friction; Engineering -- Congresses; Engineering -- Periodicals; Engineering; Conference proceedings; Periodicals; 620.005
• DOI: 10.1016/j.proeng.2016.06.246
• Languages: English
• ISSNs: 1877-7058

Abstract

Abstract: Despite almost a century of research, the question of what causes a curling stone to curl (move perpendicular to its initial direction of motion) has no complete answer. Many hypotheses have been formulated, but none has been able to account quantitatively for the full magnitude of the observed curl. The objective of this research was to equip a curling stone with an inertial measurement unit (IMU) and measure its motion, in order to verify a previously published, numerical model of curling stone dynamics. Low cost, small size, accuracy, ease of programming and operation, wireless data communication, and a data-sampling rate near 1 kHz, were selection criteria and constraints for the instrument package. We used the MicroStrain 3DM-GX4-25 system. This is a MEMS-based IMU with a tri-axial gyroscope and a tri-axial linear accelerometer. It was mounted and interfaced with a Bluetooth transmitter, on a curling stone handle. The data were streamed to a host laptop and displayed graphically in real time. Post-processing of the data included filtering and time-integration in order to obtain linear and angular velocities, and displacements. We have compared our experimental results with trajectory data calculated using a previously published numerical model, based on a thermodynamic approach to ice friction. While the observed longitudinal and angular motions are captured reasonably well by the model (errors of about 5% or less), no curl is predicted by the model.