Research on Error Characteristics and Fusion Optimization of MEMS Sensor for Manipulator

Abstract

With the rapid development of Industry 4.0 and intelligent manufacturing, lightweight collaborative manipulators are increasingly used in precision assembly, intelligent sorting, flexible handling and human-robot collaboration, and their operating accuracy and stability depend heavily on the detection performance of MEMS sensors. However, factors such as ambient temperature fluctuations, mechanical vibration, and long-term accumulated deviation often cause MEMS sensors to experience zero-point drift, data jitter, and detection bias. Moreover, the limitations inherent in single-sensor systems can further amplify operational errors, thereby restricting the precision capabilities of small- and medium-sized manipulators. To address this problem, this paper takes small- and medium-sized collaborative manipulators as the research object, systematically studies the error mechanism and coupling characteristics of MEMS inertial sensors, force sensors and infrared ranging sensors, and proposes an integrated optimization scheme from three dimensions: hardware calibration, algorithmic filtering and multi-sensor fusion. The results show that, compared with the conventional single-sensor detection and mean-filtering scheme, the proposed integrated optimization scheme reduces the static attitude angle error by 62.4%, the dynamic trajectory error by 44.2%, and the compliant grasping force error by 67.6%, while the workpiece slippage and damage rates are both reduced to zero. This significantly improves the manipulator's positioning accuracy, motion stability and compliant operation capability, providing a theoretical basis and technical reference for error optimization and engineering application of MEMS sensing systems in manipulators.