Hand Gesture Based Speed and Direction Control of DC Motor Using Machine Learning Algorithm

The gesture is a foremost non-touching reciprocal action between humans and computers. The hand gesture can be utilized in many types of work such as home appliance control, physically handicapped person, smart car control, and many more. In this paper, a system is proposed to control the speed and direction of the DC motor by hand gesture. An apparel device called a data acquisition unit adheres to the hand so that it can record the gesticulation of the hand. This device converts the hand gesture into a set of discrete values and transmits the data via Bluetooth to another circuit that is connected to the computer. A data set is created by recording the values obtained for different gestures to train several machine learning models. The most efficient model to recognize the hand gesture is identified based on the performance of the models. The Cubic SVM model shows the maximum classification accuracy for our data set. The accuracy of Cubic SVM is 97.8%. This classification model is used in the real-time operation of the system. The system shows 95% accuracy in real-time gesture recognition in controlling the speed and direction of the DC motor.