Neuro-Fuzzy Controller for Active Ankle Foot Orthosis

The objective of this paper is to design a neuro fuzzy controller for an AFO. Adaptive Neuro Fuzzy Inference System (ANFIS) is selected after a detailed study of existing neuro-fuzzy architectures. Data of gait pattern are collected with the help of analog gyro sensors. This data is fed to the ANFIS and a fuzzy rule base is created to complete the neuro-fuzzy system which is used to control the gait pattern. Angular velocity and angle of feet served as inputs to the controller and the output is actuation. The results obtained showed sigmoidal membership functions for the various inputs and outputs due to their close resemblance with the normal human gait. Output of the ANFIS showcased the initial data which is fed to the system; the modified data; changed membership functions and error after training.
Source: Perspectives in Science - Category: Science Source Type: research