Research on detection state recognition of articulated arm coordinate measuring machine based on inertial sensors

Abstract The articulated arm coordinate measuring machine (AACMM) is widely used for three-dimensional measurements due to its flexibility and portability, which are crucial for industrial inspection. However, the operation of the AACMM is heavily reliant on manual handling, which can introduce both random and gross errors, potentially compromising measurement accuracy. To minimize the influence of human factors on measurement outcomes, this paper presents a method for detecting the measurement state using inertial sensors. By installing inertial sensors at the probe tip, the method collects real-time motion data during operation. It conducts measurement experiments to analyze the sources of gross errors caused by operational mistakes and employs deep learning algorithms to automatically recognize the measurement state. By identifying the operator’s specific state at the moment of data collection, the method can effectively assess the likelihood of gross errors, thereby monitoring measurement accuracy. Experimental results demonstrate that this approach significantly mitigates human errors, thereby improving the reliability and precision of AACMM measurements.