Evaluating the Handshake: A Study on Human-Human Interaction for the Identification of Social Cues

Incorporating authentic human behaviors into social robots is key to creating meaningful and engaging interactions. This research seeks to capture and model the distinct styles of human handshakes, allowing robots to perform these interactions with greater social awareness and expressiveness. By integrating sensor data from custom-designed gloves with self-reported social measures, in this paper, we preliminarily explore the interplay between physical interaction dynamics (extrinsic factors) and individual characteristics (intrinsic factors). Our research aims to develop a dataset that captures the style of human–human handshakes through a combination of motion data recorded by prototype gloves and subjective impressions collected through questionnaires. In this study, we recorded eighteen handshakes involving 36 participants. From these interactions, we extracted features such as applied forces and correlated them with participants’ personality traits, affective states, and levels of acquaintance. Using K-Means clustering, we analyzed four variable pairs and identified distinct handshake styles, each defined by unique motion dynamics that reflect both individual behavioral traits and interpersonal familiarity with the handshake partner. These results provide a foundation for further exploration into social signaling during physical interactions, with meaningful implications for understanding human behavior and advancing social robotics.