Locomotion mode identification for lower limbs using neuromuscular and joint kinematic signals

Recent development in lower limb prosthetics has seen an emergence of powered prosthesis that have the capability to operate in different locomotion modes. However, these devices cannot transition seamlessly between modes such as level walking, stair ascent and descent and up slope and down slope walking. They require some form of user input that defines the human intent. The purpose of this study was to develop a locomotion mode detection system and evaluate its performance for different sensor configurations and to study the effect of locomotion mode detection with and without electromyography (EMG) signals while using kinematic data from hip joint of non-dominant/impaired limb and an accelerometer. Data was collected from four able bodied subjects that completed two circuits that contained standing, level-walking, ramp ascent and descent and stair ascent and descent. By using only the kinematic data from the hip joint and accelerometer data the system was able to identify the transitions, stance and swing phases with similar performance as compared to using only EMG and accelerometer data. However, significant improvement in classification error was observed when EMG, kinematic and accelerometer data were used together to identify the locomotion modes. The higher recognition rates when using the kinematic data along with EMG shows that the joint kinematics could be beneficial in intent recognition systems of locomotion modes.