Bionic arm: mapping of elbow and wrist flexion using neural network and fuzzy logic

Abstract

Cases pertaining to limb amputation necessitate the use of Transhumeral bionic for artificial limb rehabilitation which is controlled using Electromyographic (EMG) signals from the muscles. Before the implementation of EMG control, a mapping between the movements of an arm to the angle formed at the corresponding joints is essential to be made. In this article, 2 models have been discussed regarding the mapping of sensors’ readings to the angular displacements of the joints. First model  captures elbow and wrist flexion and maps them to their respective angular displacements of joints at bionic model using fuzzy logic. This model is about finding an angular displacement value corresponding to a sensor value which is determined from its relative membership in consecutive sensor readings. Model 2 involves data collection for different positions of elbow and wrist flexion (categories of classification), loading the recorded data into a neural network, optimizing the network’s parameters and real-time testing. This model was verified by comparing joint angles of a test person (measured using Goniometers) with the joint angles of Bionic models made (using a 360° protractor sheet).