Let's handshake and I'll know who you are: Gender and personality discrimination in human-human and human-robot handshaking interaction

In order to appear as natural as possible during social interaction, robots need either the ability to express or to measure emotions. Touch can be a very powerful communication modality but is very little exploited. Our paper aims to create a model of the tactile features activated during a handshake act, that can discriminate intrinsic characteristics of a person such as gender or extroversion. First, we construct a model of handshaking based on the human-human styles of handshaking. This model is thereafter compared with a human-robot handshaking interaction. The Meka robot is used in our experiments. A method is proposed to manage the modeling using feature selection with ANOVA and linear discriminant analysis (LDA). The first preliminary results show that it is possible to recognize gender and extroversion personality trait based on the firmness and movement of the style of handshaking. For instance, smaller pressure and frequency were found to describe female handshakes and higher speed amplitude describe introverted handshakes. Consistency was also found when comparing human-human handshaking with human-robot ones. These are encouraging results that will allow us to develop personalized interactions.