• Journal of Institute of Control, Robotics and Systems
• /
• v.21 no.10
• /
• pp.944-950
• /
• 2015

Recent studies on ankle-foot prostheses used for transtibial amputees have focused on the adaptation of the ankle angle of the prosthesis according to ground conditions. For adaptation to various ground conditions (e.g., incline, decline, and step conditions), ankle-foot prostheses should first recognize the ground conditions as well as the current human motion pattern. For this purpose, the ground reaction forces and orientation angle of the tibia provide fundamental information. The measurement of the orientation angle, however, creates a challenge in practice. Although various sensors, such as accelerometers and gyroscopes, can be utilized to measure the orientation angles of the prosthesis, none of these sensors can be solely used due to their intrinsic drawbacks. In this paper, a time-varying complementary filtering (TVCF) method is proposed to incorporate the measurements from an accelerometer and a gyroscope to obtain a precise orientation angle. The cut-off frequency of TVCF is adaptively determined according to the human gait phase detected by a fuzzy logic algorithm. The performance of the proposed method is verified through experiments.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.1
• /
• pp.8-16
• /
• 2017

There are various methods to count the number of repetitive motions such as jump rope. Most of the methods use features extracted from the time-varying waves of acceleration or angular velocity, which is the main feature in the count of rotations in jump rope. However, there exist several variables and it is not easy to find the count with a single sensor. For example, accelerometer is susceptible to noise and vibration, and the angular velocity may cause a drift phenomenon, which is the main cause of the inaccurate count of jump rope rotation. In this paper, complementary filter is used to consider two sensors simultaneously and complement each other, which results in more accurate count in jump rope rotation. The proposed method can count the exact number of jump rope rotation compared to other existing methods only using one sensor value, which is confirmed through experimental results.