Contactless person identification using cardiac radar signals

Radar systems have been researched for the use of presence detection and contactless vital sign monitoring. However, there exists no established biometrics for remote and unique person identification during such monitoring. Conventional biometrics like fingerprint or iris scan yield the disadvantage that direct contact with the person is needed. This paper explores the possibility of using cardiac radar signals as new biometric parameter for unique person identification. Measurements on different persons are performed using a 24 GHz continuous wave radar system which utilizes the Six-Port technology. An advanced signal processing and classification routine is presented to perform automatic person identification. Among several classifiers, quadratic support vector machines achieve the best performance and reach an overall accuracy of up to 94.6 %.