Development of Electromyography Signal Contraction Detection Using the Double Threshold Method for Controlling Wheelchair

Abstract

The consequence of spinal cord injury (SCI) is the malfunction of some parts of the body. This observation especially applies to trauma at the cervical level, because both patients' hands and legs are no longer functioning. The purpose of this study was to determine the mechanism of the comparator treshold system in electric wheelchairs for patients with paralysis. The research contribution is to get a treshold mechanism that can make a wheelchair run smoothly without being affected by the strength of the user's unstable muscle contractions. To set the muscle reference value, connect the transmitter to Android using USB Cable. On Android there are 2 settings modes, namely manual and auto settings. The method used in auto mode is the transmitter will take 100 contraction and 100 relaxation data and multiplied by certain constants to get the upper and lower threshold values. In setting the threshold manually, the value is set based on the strength of the patient's contractions that are displayed on Android manually. Where in the threshold setting manually obtained response 93.3% of forward movement, 100% of turn left movement, 100% of turn right movement and 93.3% of backward movement. Meanwhile, the threshold setting automatically gets 80% response for forward movements, 93.3% for turn left movements, 86.6% for turn right movements and 93.3% for backward movements. The results of the tapping into 4 parts of the muscle can be implemented to detect muscle contractions so that the wheelchair can move according to orders. The problem obtained in this study is that there are many tapping points on the face so that it interferes with the comfort and time of installing electrodes on the face, also the patient's muscle reference setting system on Android still needs the help of others. In the future, it is expected to reduce tapping points and improve the efficiency of muscle threshold settings.