Behavioral Features for Different Haptic-based Biometric Tasks

The science of haptics or haptic technology has received enormous attention in the last decade for multiple applications. A new promising example is the use of haptic-based systems for individual authentication. A user's behavioral characteristics captured while interacting with a virtual scene are significantly more dfficult to compromise than traditional means (i.e. login ID and passwords). Moreover, another advantage is that this technology allows a user not only to gain access to the system but also to continuously verify individual authenticity during the whole session of haptic interaction. Our current haptic-based biometric system for authentication (the BioHaptic system) has integrated a variety of applications: solving a virtual maze, signing a virtual cheque and dialing phone codes on a virtual phone. In this paper, with the use of a 3D visual representation we study: (1) a subset of features with the greatest user-classificatory worth and (2) whether or not such features are dependent on the application used. We believe this study will enhance the success of individual authentication based on human-haptic interactions.