EMG Hand movement recognition using Artificial Neural Network and Support Vector Machine

This paper contains detailed information about creating a system capable of classifying arm muscle responses (using EMG signal) into five hand gesture classes. Informations about collecting data, processing it and validating are described. Such system is especially useful for disabled people, but low number of electrodes used creates other possibilities of usage - for example a Human-Computer Interface (HCI). Although it still cannot be used in real life as a prosthesis (we know what to do, but still don’t know how to do it exactly - for example what force to apply) it shows how easily we can interact with signals coming from our body.

Paper is available here.

Arduino with HC-SR04 ranging sensor

How it works

To begin the measurement supply the TRIG pin with +5 V for 10 μs. The module begins to emit the ultrasonic wave (with frequency of 40 kHz).  ECHO pin receives the returning signal - it's high state time is proportional to the distance traveled by the wave.

distance [m] = (high level time [s] x 340 m/s) / 2
distance [cm] = (high level time [μs] x 34) / 1000 / 2

Arduino code

int getDistance(int trigPin, int echoPin) {
  unsigned long duration = 0;
  
  digitalWrite(trigPin, HIGH);
  delayMicroseconds(10);
  digitalWrite(trigPin, LOW);
  
  duration = pulseIn(echoPin, HIGH);

  return duration * 34 / 1000 / 2;
}