Negar Riahi - Center of Advance Systems and Technologies (CAST), School of Mechanical Engineering, College of Engineering, University of Tehran;
Aghil Yousefi-Koma - Center of Advance Systems and Technologies (CAST), School of Mechanical Engineering, College of Engineering, University of Tehran
Ali Abbaszadeh - Center of Advance Systems and Technologies (CAST), School of Mechanical Engineering, College of Engineering, University of Tehran
Reihaneh A. Hosseininejad - Center of Advance Systems and Technologies (CAST), School of Mechanical Engineering, College of Engineering, University of Tehran

Robotics and rehabilitation prosthesis are the types of robotic equipment which come to the aid of disabled people and retrieve some of their lost abilities. The main goal of this paper is to implement the control commands issued by a disabled person on a robotic hand online. By designing a user interface based on myoelectric signals of the forearm muscles, it became possible to successfully implement human’s issued control commands online on a robotic hand with classification error less than 9% for normal people with 5 gestureclasses and less than 6% for disabled people with 3 gesture classes. To study the myoelectric signal data, a commercialdevice named Myo Armband from Thalmic Labs Company was used which has a sampling frequency of 200 Hz. By classifying the signals and extracting theuseful characteristics, a pattern recognition based algorithm was implemented on the training data. The process was then transformed to an online one and the low level control commands for gesture classes’ imitation were sent to actuators using an Arduino

myoelectric signal, pattern recognition based algorithm, online implementation