Attitude measurement of permanent magnet spherical motors based on adaptive mahony complementary filtering

•The drift problem of the gyroscope in the inertial measurement unit of Permanent Magnet Spherical Motor is resolved.•The method is unaffected by magnetic field variations and more suitable for three-degree-of-freedom mechanical structures.•The measurement accuracy and computation time of the algorithm are verified through experimental validation. An adaptive attitude measurement algorithm called adaptive mahony complementary filter(AMCF) is proposed to address the issues of drift and noise in gyroscope measurements, as well as the high computational complexity, error accumulation, and sensitivity to noise in the quaternion extended kalman filter(QEKF) within the permanent magnet spherical motor(PMSM) inertial measurement unit. Firstly, this method corrects the gyroscope’s angular velocity bias by utilizing the error between theoretical and actual accelerations. Subsequently, the covariance is adjusted by comparing dynamic accelerations to mitigate disturbances. The fourth-order Runge-Kutta method is employed to continuously update the quaternion, thereby achieving higher accuracy in attitude measurement. Finally, static and dynamic experimental investigations are conducted on the PMSM. The experimental results demonstrate that the AMCF offers faster computation speed compared to the QEKF and higher accuracy than the fast complementary filtering(FCF). This indicates that utilizing the AMCF enables better real-time attitude measurement for three-degree-of-freedom mechanical structures.