Performance evaluation of a sensor fusion algorithm for attitude estimation using commercial IMU and scale Stewart platform
Costa, M. S. M.1; Meggiolaro, M. A.1; Speranza Neto, M.1; Albuquerque, A. N.1; Assad, M. M.1
1 PUC-Rio
doi:10.20906/CPS/COB-2015-0305
Resumo
IMU (Inertial Measurement Unit) is a set of inertial sensors normally composed of a triaxial accelerometer, a triaxial gyro and a triaxial magnetometer and is widely used in vehicles to evaluate their kinematic states (accelerations, velocities, positions and orientations). These sensors are indispensable in the control of Unmanned Aerial Vehicles (UAV), for example, and, with the recent popularization of these vehicles, there is a great effort in the development and price reduction of these units. In this work, a new sensor fusion algorithm is proposed to improve the estimated attitude (orientation) calculated from the data acquired by a low cost commercial IMU. In this algorithm, the measurements made by the gyro sensor are used to update the state transition matrix and, with an integration procedure, form one estimated attitude. In the other hand, the measurements made by the accelerometer and magnetometer are combined to form the second estimated attitude. Both solutions are fused to form an improved solution. For algorithm validation, the unit is coupled to a Stewart platform which in turn reproduces pre-programmed movements so that there is a known reference of the kinematic state of the IMU. With this reference, the calculated platform attitude values are compared with those obtained by the accelerometer combined with magnetometer solution, with those obtained by the gyro integration solution and with those obtained by the proposed fusion algorithm. The results indicate that the proposed algorithm minimizes noise and drift errors.
Palavras-chave: Inertial Measurement Unit; Stewart platform; Sensor Fusion; Attitude Estimation