Localization performance of hum-noise-based touch sensor (HumTouch) with unknown participants

HumTouch is a touch sensing method that can turn object surfaces into touch sensors. The sensing method utilizes the current leak from the human body; therefore, the localization performance exhibits individual differences. In this paper, we research the localization performance of HumTouch involving multiple participants. A semi-conductive paper marked with 49 points was used for the experiment. Five participants participated in the experiments and the localization results were evaluated via leave-one-out cross validation, in which an estimation model was established by using the samples collected from four participants to localize the touch of the last participant. For four out of the five participants, there were mean estimation errors smaller than or equal to 1 cm, which is comparable to the size of a fingertip. This indicates that HumTouch has the potential to be applied to multiple participants without prior calibration conducted for a specific person.