A human-machine interface based on tongue and jaw movements

Abstract

This paper describes a human-machine interface (HMI) that works on the basis of pressure changes, caused by tongue and jaw movements, in the closed, external auditory canal This work covers the definition of specific tongue movement sequences which are suitable for detection in the outer auditory canal. Furthermore an algorithmic concept for detection and differentiation of four movements based on pressure changes in the auditory canal is developed. Using the correlation coefficient of a recorded calibration and the current signal the problem of a deviating temporal course, caused by different speeds of movement and variation of the amplitude spectrums, is solved. As the calculation requires the complete capture of the movement signals, a real-time detection is not possible. However, for the recognition it is necessary to detect the entire motion signal. Due to the observed dependency between signal generation and jaw movement an additional examination of two jaw movements was conducted. After movement definition, a real-time capable algorithm for detection and differentiation was developed and subsequently implemented on an existing system for pressure value detection. Afterwards both resulting systems were tested and evaluated on study participants. Thereby it was demonstrated that the detection und differentiation of four tongue movements is possible. In total it could be shown that tongue and jaw movements are suitable input mechanisms for human machine interfaces and bear a great potential for future developments.