Ho-Chuan HsuShang-Yu LinPo-Han ChenPEI-ZEN CHANGWEI-CHANG LI2025-03-212025-03-212024-10-20https://www.scopus.com/record/display.uri?eid=2-s2.0-85215323305&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/726001This study showcases the utilization of an intelligent vise equipped with integrated cutting force and audio sensors to predict surface roughness during the milling process. Specifically, a piezoelectric PZT-based force sensor and a MEMS microphone are embedded within the vise jaw, enabling the acquisition of cutting force and audio signals in close proximity to the workpiece without interference. The collected data is then used to train predictive models for estimating surface roughness. The results demonstrate a root mean square error (RMSE) of 0.038, which outperforms the use of either force or audio data alone, which have RMSE values of 0.072 μm and 0.103 μm, respectively.audio sensorCNC machiningcutting force sensormillingmulti-modalsurface roughness prediction[SDGs]SDG9[SDGs]SDG12conference paper10.1109/SENSORS60989.2024.10784677