Abstract
This paper presents a method of measuring knee joint motion for clinical diagnostic procedures. MRI or optical tracking methods are examples of existing methods to quantify diagnostic procedures for assessing knee joint ligament injury. However, these methods require a large amount of space and are quite expensive. An alternative method is to use multifunctional sensors which contain tri-axial accelerometers and gyroscopes. Accelerometers can measure the tilt angle against the gravity vector and estimate the translational acceleration. Gyroscopes can measure the angle of rotary motion. An angle obtained by a gyroscope is normally influenced by noise; causing a bias in angle. This work presents a system using a Kalman filter to remove the bias angle obtained by the gyroscope. Experiments that measure three rotary motions and a translational motion of the knee joint were conducted. Sensors are mounted on the Tibia and Femur bone of a knee joint model. It was found that the bias of the gyroscope's output was minimized in all three rotary motions by applying the proposed system. In addition, translational distance was calculated by integrating the accelerometer's output. However, distance obtained by acceleration is significantly influenced by noises due to double integration. A method to process the acceleration data and obtain translational distance is also proposed.
