Design of a Neuromorphic Dexterous Prosthesis Hand

Motivation and Background

Hands play an important role in human life. We use our hands to do day-to-day tasks as well as to work. Unfortunately, there are those who lose their hands or arms due to accidents or diseases. It is estimated that at least 1.5 million people are living with an absence of upper limbs. The progress of Prosthetic arms and human-like robot hands research and development have been improved incessantly. There are many approaches to realize a prosthetic hand. The most recent approaches have been using electromyographic (EMG) signals as a source to control the prosthetic hands. With the help of machine learning, the numbers of controls and movements have been increasing compared to only close and open movements in the past.

With EMG, biological data taken from the human body, as the input source, Spiking Neural Network (SNN) has become a favorable solution. When a movement command is sent from the brain to the nerve of the arm, muscle contraction will occur. Using that very moment of EMG signal, the prosthetic arm will move according to the signal inputs.

Research Goal

To develop a neuroprosthetic hand control based on Neural Network.

Research Schedule

DateTask
☑️September 1~15 2022Online Testing of SNN model on Raspberry Pi
September 16~30, 2022Improve SNN model (currently there is a limitation in online testing)
October 1~31, 2022Hand rotation recognition using the IMU sensor; start writing thesis
November 1~30, 2022Fix and optimize the SNN/IMU model, thesis writing
December 1~31, 2022System integration, thesis writing
Schedule last Updated on: September/14/2022

Shared Google Folder

Reference

Links


トップ   新規 一覧 検索 最終更新   ヘルプ   最終更新のRSS