Optimizing the Torque of Knee Movements of a Rehabilitation Robot

The aim of this study is to employ the novel adaptive network-based fuzzy inference system to optimize the torque applied on the knee link of a rehabilitation robot. Given the special conditions of stroke or spinal cord injury patients, devices with minimum error are required for performing the rehabilitation exercises. After examining the anthropometric data tables of human body, some parameters such as the length of shins, weight, force, joint angle etc. were chosen as the input data. The torque is considered as the output of the system. Errors at any stage of the treatment can harm the patients or disrupt their recovery process. Therefore, after examining different numbers of various fuzzy inference system membership functions and their consequent error, cases with the lowest error were chosen to be the best possible conditions for the system. As a result, it can be said that a robot using the adaptive network-based fuzzy inference system offers negligible errors.