A novel non-contact hand motion classification technique for application to human machine interfaces

In this paper, we present an effective non-contact technique for the measurement of human hand motion with four square electrodes; this technique allows the detection of the subject´s hand movements in eight different directions. This new method involves the measurement of the current generated due to the difference in the capacitance between the subject´s hand and the measurement electrodes. Our classification approach yields satisfactory recognition results despite the use of a relatively primitive neural network model. This model allows us to recognize the direction of the hand movement with respect to the measurement electrodes under perfect non-contact conditions.