A MEMS inertial sensor-based onboard system with calibrated vector-synthesis attitude estimation for dynamic testing of automatic loading systems

Purpose The purpose of this paper is to design an onboard storage and testing device for automatic loading systems in large-caliber heavy machinery, to accurately measure acceleration, angular velocity and attitude under high impact and vibration, overcoming the limitations of wired sensors and attitude algorithms that assume ideal axis alignment, thereby supporting system optimization, fault diagnosis and performance enhancement. Design/methodology/approach First, an equipment-mounted storage testing device was developed using a CA-YD-180 ICP accelerometer, an ADIS16365 six-degree-of-freedom MEMS inertial sensor and an STM32 microcontroller for data acquisition, storage and communication. System-level calibration was then performed: gyroscopes were calibrated using the static angular rate method with least-squares error compensation, and accelerometers were calibrated using the six-position method. Next, to address non-ideal mounting conditions where the sensor’s sensitive axis is not perfectly parallel to the rotation axis, an angular velocity calculation method based on vector synthesis was proposed. Finally, the device was validated through static tests, accuracy verification and live loading experiments, with results compared against ADAMS multi-body dynamics simulations. Findings After calibration, the gyroscope’s zero bias decreased from 1.386°/s to −0.018°/s. Under intersecting rotation and sensitive axes, the typical absolute error for a 90° rotation angle calculation was 1.2°. Actual loading experiments revealed a 25 Hz vibration during feeding mechanism operation, more pronounced than simulation predictions; the coordinator arm swing angle calculation error was 3.3°; and the payload tray exhibited significant angular velocity fluctuations with periodic residual vibrations after rotation. The measurement system achieved 10 effective bits and successfully captured overload accelerations. Originality/value This paper presents a complete methodology for system-level calibration of MEMS inertial sensors tailored to automatic loading systems. It proposes a vector-synthesis-based attitude calculation method that tolerates installation deviations. The comparative analysis between experimental results and ADAMS simulations provides new insights into dynamic characteristic differences and error sources. The developed device fills a gap in real-time, in situ dynamic testing for automatic loading systems and has significant engineering application value for similar complex mechanical systems.