Design of Intent Recognition System in a Prosthetic Leg for Automatic Switching of Locomotion Modes

Prosthetic leg is an artificial device that replaces a missing leg, which may be lost through accident, disease, or due to a condition at birth. The development of prosthetic legs has provided a solution for the rehabilitation of the amputees so that their mobility and ability to manage daily activities can be improved. But this is only possible if a smart intent recognition system is added with it. The existing electromyography-based systems are too expensive to be afforded by the mass. In this paper, a low-cost pattern recognition-based method is proposed for locomotion mode recognition (LMR) by using a combination of angular data from multiple gyroscope sensors combined with a force sensor. Three types of locomotion activities including walking, ascending and descending staircase are identified in this work. A dataset with sufficient amount of gait cycle values for each activity is developed to create the classifier model, and train and test it. Finally, the performance is assessed through both offline and online locomotion mode detection.