Algorithm for estimating the orientation of an object in 3D space, through the optimal fusion of gyroscope and accelerometer information

Two suitable coordinate systems overlapping at the origin were established to estimate orientation by computing the spatial relationship between them. An Inertial Measurement Unit sensor (IMU), consisting of a tri-axial gyroscope and tri-axial accelerometer was used to define reference systems. This work describes an algorithm to estimate the orientation of an object in 3D space through the optimal fusion of gyroscope and accelerometer information. A tri-axial gyroscope was used as a main source of information for assessing orientation during movement. On the other hand, a tri-axial accelerometer was used to compensate drifting deviation on gyroscope measurements. Orientation estimation was performed by using a Direct Cosine Matrix (DCM) as a combination of three consecutive rotations through one of each main axis to the coordinate systems. Rotation matrices also expressed as approximations rather than identities were used to improving orientation computed through DCM. Three different algorithms were proposed to estimate orientation. For assessing differences between them, two different studies were realized: 1. - analysis of behavior in both static position of the sensor and during moving, 2. - estimation of the orientation of the sensor by representing the orientation of an object through a virtual model. The developed algorithm opens opportunities to be used in the evaluation of human body joints.