Mind Controlled Wireless Robotic Arm Using Brain Computer Interface

Abstract

Brain Computer Interface (BCI) provides an individual to communicate using brain activity through Electroencephalogram (EEG) for controlling devices. Robotic arm is one such application which assists physically challenged people in day to day activities. The proposed method helps to interact and control the robotic arm wirelessly with user friendly interface. The wireless module and hardware interface used are cost effective. In this work, visual evoked potential based motor-imaginary hand movements like left, right, up and down hand movements are considered. Enobio-8 device is used for acquiring EEG signals. Visual evoked potential concept is adapted for acquiring the EEG signals from 14 healthy subjects of age group 20–23. These signals are pre-processed using a band pass filter of 2 to 40Hz to remove all the artifacts. Multilevel wavelet transform is used for extracting the features from specific band of interest. K-Nearest Neighbour (KNN), Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) classifiers are used for classifying motor-imaginary hand movements. Robotic arm is interfaced with the Arduino uno board and it is wirelessly controlled from EEG signals via HC-05 Bluetooth module. The results obtained from LDA for two scenarios: left/right and up/down movements is 87.5%. LDA showed better performance when compared to KNN and SVM. The accuracy of KNN is 56% (left/right) and 62% (up/down). The accuracy of SVM is 81% (left/right) and 68% (up/down). The result from LDA is promising for bringing wireless mind-controlled robots much closer to the human hand.