Human Identification Using MIMO Array

This paper proposes a method of using a microwave multiple-input multiple-output (MIMO) array to recognize different people; it allows smart home applications to dispense with contact and wearable devices. This method captures and processes the time-variant channel generated by a human body present within the field of an MIMO antenna. The time-variant channel is measured, and the spatial and temporal correlation is used for identification by comparing the captured data to reference data. The proposed method identifies the subject as the registrant with maximum correlation. Experiments involving eight known subjects and four imposters demonstrate that the proposed method offers accurate identification with no false positives. The results show that an 8 × 8 array can achieve the equal error rate of 0% when the number of identification channels is five and above.