In this paper, we proposed an IMU-based locomotion mode identification (LMI) system for ankle-foot prostheses. Specifically, an IMU sensor was mounted on the heel to collect the foot's dynamic information during walking. Then processing the dynamic data can estimate the foot trajectory for calculating the inclination grade of the terrain. It is noteworthy that our environment is constructed according to the inclination grade for ergonomics. For example, when the inclination angle ranges from 3 degrees to 11 degrees, the environment should be a ramp. On the other hand, when walking on different terrains, people prefer to move their feet around the ground's exterior. It is helpful for people to get the required foot clearance and, at the same time, minimize the energy needed for transporting the lower limbs. Therefore, with the estimated inclination grade, the presented method can precisely predict/identify the locomotion mode. Experimental results show that the average accuracy can reach 98.4% in five daily locomotion modes, including level-ground walking, stair ascending/descending, and upslope/downslope walking.
|