Pengenalan Pola Gerakan Jari dengan Algoritma LDA Secara Waktu Sesungguhnya


  • Daniel S Pamungkas Politeknik Negeri Batam
  • Wahyu Pardede Politeknik Negeri Batam
  • Sumantri K Risandrya Politeknik Negeri Batam


Electromyograph, LDA, pattern recognition


Hands are one of the body parts that have an important role for living things, especially humans.
Most of the activities that humans do require the help of hands. But not all humans have perfect hands or
function like hands in general because of the impact of genetic disorders or the result of accidents. This is
certainly a very disturbing problem for people with disabilities to live their daily lives. In this study, the author
uses a fabricated device from Thalmic Labs named Myo Armband. This tool was created for gaming purposes,
computer control, and so on. However, this tool is also widely used for the benefit of technology development,
especially in the health sector. Myo Armband has eight Electromyograph (EMG) sensors that are able to record
and recognize every activity of arm muscle movement. In this study, the EMG signal is recorded and processed
which is intended to distinguish each hand movement based on the muscle signal read by the EMG sensor. After
the EMG signal is recorded, the EMG signal will be read and continued with the training phase. After obtaining
the training weights, the results will be used at the testing phase and classification will be carried out. In the
classification process, the writer chose the Linear Discriminant Analysis (LDA) method. This method was
chosen as a method for classifying finger movement pattern recognition. The percentage of success in the study
up to 76% Furthermore, after getting the conclusions from this study, the authors hope that this research can be a
reference for the development of making hand robots, especially for medical needs.