Publication

Abstract : Current progress as a humanitarian challenge is the designing of a low cost neuro-prosthetic arm that could be controlled using EEG-based signal re-classification. Brain machine interfaces (BCI) are targeted at the people who are paralyzed and unable to perform motor actions. In future, a BCI could help these patients to have the robot to perform the motor actions providing the signals from the brain (Tan & Nijholt, 2010). From computational perspective, different algorithms have been proposed to solve the robotic kinematics; DH method (Iqbal et al., 2012), homogenous method (Mitra,2012) for forward kinematic model and analytical method (Iqbal et al., 2012) for inverse kinematics. Solving an inverse kinematics and trajectory planning are computationally cost effective. Brain computer interface (BCI) have been employed to control prosthetic arms (Wolpert & Flanagan 2010) in order to do specialized tasks, namely, reaching the target with an optimal feedback (Mitrovic et al. 2010). We are working on real time machine learning based kinematic algorithms with an alternative approach of using a spiking neuron network model (CIS-NN, unpublished data) substituting firing rate as a metric to measure trajectory patterns. We could optimally control a high-dimensional anthropomorphic robot without having to specify an explicit inverse kinematics (Mitrovic et al., 2010).