A Programmable Robotic Phantom to Simulate the Dynamic Respiratory Motions of Humans for Continuous Identity Authentication

Abstract

Non-contact continuous identity authentication using microwave Doppler radar has shown great promises in security and surveillance applications and yet not mature. Reported results for unobtrusive identity authentication system are focused on dynamic segmentation which produces false classification when inhale/exhale area ratio becomes closer. For testing the efficacy of the proposed identity authentication system, many different respiratory patterns must be examined repetitively, and with consistency between experiments for reliable algorithm development. Since it is not practical to have humans repeat measurements in an exact, specified manner, a programmable respiratory phantom mover has been developed. We have tested the hypothesis that such a phantom mover system can be programmed to accurately simulate the distinct human respiratory patterns needed for research on respiratory pattern separation and identity authentication algorithms. The phantom incorporates a remotely programmable Arduino microcontroller driving a compact high-speed linear actuator. Comparison with real and/or simulated patterns have been successful for algorithm development. The impact of this research could result in new tools for medical diagnostics, security/surveillance and continuous identification.