Skip to Main Content
A method for estimating limb orientation, during static, quasi-static, and dynamic motions, by using a combination of gyroscopes and accelerometers is presented. The method uses two tri-axis accelerometers and one single axis rate gyro to calculate an estimate of angle relative to the gravitational vector independently from the rotational accelerations. This unbiased inclination estimate is blended with the angular velocity and acceleration measurements using a Kalman filter to obtain the final estimated orientation. Initially developed for implementation in a feedback control loop for control of sit-to-stand transitions in felines during direct electrical stimulation of the sciatic nerve, the method can be directly applied to other limb angle measurement tasks such as human gait analysis. The algorithm and sensor is tested on a rotational linkage with a predefined trajectory and compared to an encoder measurement with good agreement.