Classification of Finger Movements Based on Reflection Coefficient Variations of a Body-Worn Electrically Small Antenna

Abstract

We propose to classify finger movements based on the reflection coefficient variations of a wrist-worn electrically small antenna. Near field perturbed by finger movements results in unique variation of antenna input impedance, which can be used to identify finger-motion patterns. Two folded cylindrical helix (FCH) antennas, operating at 890 MHz and 2.43 GHz, are implemented in this study. First, four movements of the left-hand fingers are recorded by measuring S 11 of the antenna attached to the wrist at the left hand. Second, we measure four movements of the right-hand fingers above the antenna attached to the left wrist. To help classify the finger motions based on the S 11 variation with time, the dynamic time-warping technique was employed. We found that the average classification accuracies are higher than 97% for the first scenario and 100% for the second scenario using the FCH that operates at 890 MHz.