Location template matching on rigid surfaces for human-computer touch interface applications

In this chapter we discuss a category of signal processing techniques called location template matching (LTM), which is widely used to convert daily objects such as glass panels and tabletops into human-computer touch interfaces. In such applications, a single sensor is mounted on a rigid surface to capture vibrations induced by users' impacts. LTM methods compare location-dependent features of the vibration signal generated by a user with a pre-collected library to estimate the impact location. In the most basic form of LTM, the signal itself is utilized as a location-dependent feature where each library signal is correlated with the received signal to determine the best match. More advanced techniques utilize the mechanical model of vibration in plate to extract other useful location-dependent features. This chapter reviews different LTM techniques arising from the time-reversal theory and the classical model for flexural vibration on thin plates. Specifically, we focus our discussion on which features are utilized for location matching and the respective matching measures. The algorithms will also be compared against each other to highlight their advantages and disadvantages.