An attitude estimation algorithm for high dynamic movement

Abstract

Limited by the MEMS sensor's sampling rate, the non-commutable error that derived from finite rotation of rigid body should not be neglected. Aimed at high dynamic motion and high-speed rotation, this paper presents an optimized algorithm based on expert control system (ECS) which combined rotation vector algorithm and quaternion to solve attitude estimation. Compared with the traditional quaternion algorithm, it can improve the calculation accuracy effectively without reducing the calculation frequency significantly. Finally, according to the result of simulation, this method was applied to a bionic ornithopter with the low-cost inertial measurement unit (IMU), which had effective function in active control experiment. This method can also be used to wearable human posture platform, robot control system, etc.