WiP: Optimizing Real-Time Tag Identification in Dynamic Commodity RFID Systems: Harnessing Intrinsic Randomness of COTS Tags

Abstract

Dynamic RFID deployments must update the tags in a reader's range as objects stay, depart, or newly arrive. Existing solutions either recollect every Electronic Product Code (EPC), assume tag-side hashing unavailable on commercial off-the-shelf (COTS) tags, or prewrite management data. We present CETI, a C1G2-compatible approach that treats dynamic identification as backend-guided live-list repair. CETI observes that commodity tags already expose useful entropy in Bank 01: EPC bits and the stored CRC. The backend converts these fields into compact selection masks, verifies known tags, and limits subsequent inventories to candidate new arrivals using only standard Select and Query commands. Measurements across 700 COTS tags motivate the design. Across eight 50-cycle conveyor scenarios, CETI reduces identification time by approximately 42% versus three evaluated baselines and achieves complete new-EPC collection in the calibrated trials. These results show that write-free Bank 01 filtering is a practical path toward dynamic RFID identification on commodity tags.