Machine Learning with Wearable RFID Grid for Monitoring Fetal Movements

Fetal movements are crucial signs of the fetus’ well-being and are clinically monitored with self-reports or wearable detectors. However, wired detectors worsen mothers’ labour, and wireless ones are still not adopted due to their higher cost. Radiofrequency identification (RFID) can be exploited to develop low-cost solutions for detecting fetal movements based on tag response variations. In this contribution, a wearable RFID grid for fetal monitoring is manufactured and tested. Even though the backscattered EM wave unpredictably changes when tags are displaced by a probe emulating a foot of a 20-week-old fetus, a decision tree for classification detected motionless and moving tags with accuracy $\gt 91$% in the preliminary test.